7__

What is Research?

Uday Khedker (www.cse.iitb.ac.in/Ëœuday) Department of Computer Science and Engineering, Indian Institute of Technology, Bombay

Mar 2011

IITB

1/88

Research: Outline

Outline

• What is research?

• The process of research • Richard Hamming on research (“You and Your Research”) • Specific questions related to Ph.D. • How to write a good research paper • Conclusions

March 2011

Uday Khedker, IIT Bombay

IITB

2/88

Research: Outline

Disclaimers

• General concepts presented in the CS/IT context • No “cook book” or “how to do it yourself”

March 2011

Uday Khedker, IIT Bombay

IITB

2/88

Research: Outline

Disclaimers

• General concepts presented in the CS/IT context • No “cook book” or “how to do it yourself” ◮ Views expressed through examples

March 2011

Uday Khedker, IIT Bombay

IITB

2/88

Research: Outline

Disclaimers

• General concepts presented in the CS/IT context • No “cook book” or “how to do it yourself” ◮ Views expressed through examples ◮

March 2011

Personal reflections and confessions

Uday Khedker, IIT Bombay

Part 2

Why Research?

Research: Why Research?

IITB

3/88

Alice’s Adventures

Through the Looking Glass Author: Lewis Carroll

March 2011

Uday Khedker, IIT Bombay

Research: Why Research?

IITB

3/88

Alice’s Adventures

Through the Looking Glass Author: Lewis Carroll

March 2011

Situation: Alice Running with the Red Queen

Uday Khedker, IIT Bombay

Research: Why Research?

IITB

3/88

Alice’s Adventures

Through the Looking Glass Author: Lewis Carroll

Situation: Alice Running with the Red Queen

‘Well, in our country,’ said Alice, still panting a little, ‘you’d generally get to somewhere else – if you ran very fast for a long time, as we’ve been doing.’

‘A slow sort of country!’ said the Queen. ‘Now, here, you see, it takes all the running you can do, to keep in the same place. If you want to get somewhere else, you must run at least twice as fast as that!’

March 2011

Uday Khedker, IIT Bombay

Research: Why Research?

IITB

4/88

Surviving in a World of Rapidly Developing Technologies

Need to run twice as fast to even remain in the same place . . . • Hard (Technical) Skills • Soft Skills

(Leadership, Motivation, Emotional Maturity, Communication etc.)

March 2011

Uday Khedker, IIT Bombay

Research: Why Research?

IITB

4/88

Surviving in a World of Rapidly Developing Technologies

Need to run twice as fast to even remain in the same place . . .

• Hard (Technical) Skills • Soft Skills

(Leadership, Motivation, Emotional Maturity, Communication etc.)

• Ability to acquire new skills

March 2011

Uday Khedker, IIT Bombay

Research: Why Research?

IITB

4/88

Surviving in a World of Rapidly Developing Technologies

Need to run twice as fast to even remain in the same place . . .

• Hard (Technical) Skills • Soft Skills

(Leadership, Motivation, Emotional Maturity, Communication etc.)

• Ability to acquire new skills

Quick self-learning is enhanced significantly by doing research

March 2011

Uday Khedker, IIT Bombay

Part 3

In Search of Research

IITB

5/88

Research: In Search of Research

What is Research?

• Carving Statues out of stones

March 2011

Uday Khedker, IIT Bombay

IITB

5/88

Research: In Search of Research

What is Research?

• Carving Statues out of stones • Methods and tools

March 2011

Uday Khedker, IIT Bombay

IITB

5/88

Research: In Search of Research

What is Research?

• Carving Statues out of stones • Methods and tools • Attempt to improve the methods and tools leads to

March 2011

◮

Better statues

◮

Better methods and tools

Uday Khedker, IIT Bombay

IITB

5/88

Research: In Search of Research

What is Research?

• Carving Statues out of stones • Methods and tools • Attempt to improve the methods and tools leads to

March 2011

◮

Better statues

◮

Better methods and tools

◮

Better sculptors

Uday Khedker, IIT Bombay

IITB

6/88

Research: In Search of Research

In Search of Research?

• Observed Phenomena with no explanations ◮

Puzzles and mysteries

• Lacunae in the known theory and/or practice ◮

The need of a better understanding/method

◮

Innovative ideas

• Innovative ideas waiting for new applications ◮

March 2011

Discovery of new puzzles, mysteries and/or lacunae

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research

• Is building a device, research?

March 2011

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research new • Is building a device, research?

March 2011

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research new • Is building a device, research?

• Is writing a software, research?

March 2011

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research new • Is building a device, research? new • Is writing a software, research?

March 2011

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research new • Is building a device, research? new • Is writing a software, research?

• Is repairing a device or debugging a software, research?

March 2011

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research new • Is building a device, research? new • Is writing a software, research?

• Is repairing a device or debugging a software, research?

• Is drawing a conclusion from a lot of data, research?

March 2011

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research new • Is building a device, research? new • Is writing a software, research?

• Is repairing a device or debugging a software, research?

• Is drawing a conclusion from a lot of data, research?

• Is proving a theorem, research?

March 2011

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research new • Is building a device, research? new • Is writing a software, research?

• Is repairing a device or debugging a software, research?

• Is drawing a conclusion from a lot of data, research?

• Is proving a theorem, research?

• Is formulating a theorem, research?

March 2011

Uday Khedker, IIT Bombay

IITB

7/88

Research: In Search of Research

The Essence of Research new • Is building a device, research? new • Is writing a software, research?

• Is repairing a device or debugging a software, research?

• Is drawing a conclusion from a lot of data, research?

• Is proving a theorem, research?

• Is formulating a theorem, research?

Research could involve any of the above, or none of the above

March 2011

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

March 2011

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

March 2011

Beauty Utility Innovation or Enhancement of knowledge

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

March 2011

Beauty Utility Innovation or Enhancement of knowledge

A Game of Significant Ideas

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

Beauty Utility Innovation or Enhancement of knowledge A Game of

Significant Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

A Game of Significant • Research is a game of significant ideas Ideas • The significance of ideas could lie in any of the following: ◮ ◮ ◮

March 2011

Beauty Utility Innovation or Enhancement of knowledge

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

Beauty Utility Innovation or Enhancement of knowledge

A Game of Significant Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research A Game of Significant Ideas • Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

March 2011

Beauty Utility Innovation or Enhancement of knowledge

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

Beauty Utility Innovation or Enhancement of knowledge

A Game of Significant Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

Beauty Utility Innovation or Enhancement of knowledge

A Game of Significant Ideas March 2011

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

• The significance A of Game ideasofcould lie in any of the following: ◮ ◮ ◮

March 2011

Significant Beauty Ideas Utility Innovation or Enhancement of knowledge

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research A Game of Significant Ideas • Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

March 2011

Beauty Utility Innovation or Enhancement of knowledge

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research

• Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

Beauty Utility Innovation or Enhancement of knowledge

A Game of Significant Ideas March 2011

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research A Game of Significant Ideas • Research is a game of significant ideas

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

March 2011

Beauty Utility Innovation or Enhancement of knowledge

Uday Khedker, IIT Bombay

IITB

8/88

Research: In Search of Research

The Essence of Research A Game of Significant Ideas

A Game of Significant Ideas

A Game of Significant A Game of Ideas Significant • Research is a game of significant ideas Ideas • The significance A of Game ideasofcould lie in any of the following: Significant A Game of ◮ Beauty Ideas Significant ◮ Utility ◮ Innovation or Enhancement Ideas of knowledge A Game of Significant A Game of Ideas A Game of Significant Significant A Game of Ideas Ideas Significant A Game of A Game of Ideas Significant Significant Ideas Ideas March 2011

Uday Khedker, IIT Bombay

IITB

9/88

Research: In Search of Research

Ingredients of Good Research

• Innovation

March 2011

Uday Khedker, IIT Bombay

IITB

9/88

Research: In Search of Research

Ingredients of Good Research

• Innovation • Aesthetics

March 2011

Uday Khedker, IIT Bombay

IITB

9/88

Research: In Search of Research

Ingredients of Good Research

• Innovation • Aesthetics • Other important aspects :

March 2011

◮

Completeness

◮

Rigour

◮

Empirical demonstration

◮

Effective communication

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The sphere of knowledge

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The sphere of knowledge • Initial general learning

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The sphere of knowledge • Initial general learning

• Learning increases with time

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The sphere of knowledge • Initial general learning

• Learning increases with time

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The sphere of knowledge • Initial general learning

• Learning increases with time • Begin focussing and

specializing

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The sphere of knowledge • Initial general learning

• Learning increases with time • Begin focussing and

specializing

• Specialize more and more

until you reach the unknown

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The sphere of knowledge • Initial general learning

• Learning increases with time • Begin focussing and

specializing

• Specialize more and more

until you reach the unknown

• Try to push the boundary

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The sphere of knowledge • Initial general learning

• Learning increases with time • Begin focussing and

specializing

• Specialize more and more

until you reach the unknown

• Try to push the boundary

• If you keep trying try hard

enough, you may succeed

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• Your view of knowledge

March 2011

Uday Khedker, IIT Bombay

IITB

10/88

Research: In Search of Research

Innovation

• The big picture

• Adapted from: The Illustrated Guide to a Ph.D. Matt Might.

http://matt.might.net/articles/phd-school-in-pictures/

March 2011

Uday Khedker, IIT Bombay

IITB

11/88

Research: In Search of Research

Aesthetics

• Total is greater than the sum of the parts

“Scientists study science not because it is useful, but because it is beautiful. Here I do not talk about the beauty of appearance or beauty of qualities . . . Here I talk about that profound beauty which comes from a harmonious order of parts . . . ” – Henry Poincare • Example : Painting

Proportion of colours Vs. their arrangements • Cassarole design

March 2011

Uday Khedker, IIT Bombay

IITB

12/88

Research: In Search of Research

Aesthetics and Beauty of an Idea

• What proportion of the box does this traingle occupy?

• Would the result hold for any traingle in a box?

March 2011

Uday Khedker, IIT Bombay

IITB

13/88

Research: In Search of Research

Aesthetics and Beauty of an Idea

• Idea: Draw a vertical line to divide the rectangle in two parts

March 2011

Uday Khedker, IIT Bombay

IITB

13/88

Research: In Search of Research

Aesthetics and Beauty of an Idea

• Idea: Draw a vertical line to divide the rectangle in two parts • The slanting lines now divide the two boxes in two equal parts

March 2011

Uday Khedker, IIT Bombay

IITB

13/88

Research: In Search of Research

Aesthetics and Beauty of an Idea

• Idea: Draw a vertical line to divide the rectangle in two parts • The slanting lines now divide the two boxes in two equal parts • Exactly as much area outside of the traingle as there is inside

March 2011

Uday Khedker, IIT Bombay

IITB

13/88

Research: In Search of Research

Aesthetics and Beauty of an Idea

h b • Idea: Draw a vertical line to divide the rectangle in two parts • The slanting lines now divide the two boxes in two equal parts • Exactly as much area outside of the traingle as there is inside • Area of a triangle =

March 2011

1 2

×b×h

Uday Khedker, IIT Bombay

IITB

14/88

Research: In Search of Research

Food for Thought

What about this triangle?

March 2011

Uday Khedker, IIT Bombay

IITB

15/88

Research: In Search of Research

Beauty of an Idea

• Ideas talk back to us

When we fix one idea, it fixes some other ideas

• Beauty lies in creating simple ideas that ◮ ◮ ◮

bring in unexpected implications relate the seemingly unrelated things illuminate and reveal much more than anticipated

• Total is greater than the sum of the parts

March 2011

Uday Khedker, IIT Bombay

IITB

16/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

March 2011

Uday Khedker, IIT Bombay

IITB

16/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

180◦

March 2011

Uday Khedker, IIT Bombay

IITB

16/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

180◦ 360◦

March 2011

Uday Khedker, IIT Bombay

IITB

16/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

180◦ 360◦ 540◦

March 2011

Uday Khedker, IIT Bombay

IITB

16/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

180◦ 360◦ 540◦

720◦

March 2011

Uday Khedker, IIT Bombay

IITB

17/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

◮

March 2011

Consider an n sided polygon

Uday Khedker, IIT Bombay

IITB

17/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

March 2011

◮

Consider an n sided polygon

◮

Choose an arbitrary inner point and connect it to all vertices

Uday Khedker, IIT Bombay

IITB

17/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

March 2011

◮

Consider an n sided polygon

◮

Choose an arbitrary inner point and connect it to all vertices

◮

We have n triangles

Uday Khedker, IIT Bombay

IITB

17/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

March 2011

◮

Consider an n sided polygon

◮

Choose an arbitrary inner point and connect it to all vertices

◮

We have n triangles

◮

Sum of all angles = n · 180◦

Uday Khedker, IIT Bombay

IITB

17/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

March 2011

◮

Consider an n sided polygon

◮

Choose an arbitrary inner point and connect it to all vertices

◮

We have n triangles

◮

Sum of all angles = n · 180◦

◮

Exclude the sum of the angles incident on the chosen point

Uday Khedker, IIT Bombay

IITB

17/88

Research: In Search of Research

Another Example of Beauty of an Idea • Question: What is the sum of all internal angles of a polygon?

◮

Consider an n sided polygon

◮

Choose an arbitrary inner point and connect it to all vertices

◮

We have n triangles

◮

Sum of all angles = n · 180◦

◮

Exclude the sum of the angles incident on the chosen point

• Answer: (n · 180 − 360)◦

March 2011

Uday Khedker, IIT Bombay

IITB

18/88

Research: In Search of Research

On the Role of Rigour in an Idea

• Rigour removes imprecision and adds concreteness • Makes an idea immune to personal interpretation

March 2011

Uday Khedker, IIT Bombay

IITB

18/88

Research: In Search of Research

On the Role of Rigour in an Idea

• Rigour removes imprecision and adds concreteness • Makes an idea immune to personal interpretation • Example: Divide 6 by 2

March 2011

Uday Khedker, IIT Bombay

IITB

18/88

Research: In Search of Research

On the Role of Rigour in an Idea

• Rigour removes imprecision and adds concreteness • Makes an idea immune to personal interpretation • Example: Divide 6 by 2

“Divide 6 into 2 parts and tell me the size of each part”

March 2011

Uday Khedker, IIT Bombay

IITB

18/88

Research: In Search of Research

On the Role of Rigour in an Idea

• Rigour removes imprecision and adds concreteness • Makes an idea immune to personal interpretation • Example: Divide 6 by 2

“Divide 6 into 2 parts and tell me the size of each part”

• Divide 6 by

1 2

“Divide 6 into half part and tell me the size”

March 2011

Uday Khedker, IIT Bombay

IITB

18/88

Research: In Search of Research

On the Role of Rigour in an Idea

• Rigour removes imprecision and adds concreteness • Makes an idea immune to personal interpretation • Example: Divide 6 by 2

“Divide 6 into 2 parts and tell me the size of each part”

• Divide 6 by

1 2

“Divide 6 into half part and tell me the size”

• More rigorous explanations: ◮ ◮

March 2011

“Divide 6 into parts of size 3 and tell me the number of parts” “Divide 6 into parts of size 21 and tell me the number of parts”

Uday Khedker, IIT Bombay

IITB

19/88

Research: In Search of Research

An Example of Research

• Area : Code optimization performed by compilers

Known Optimization : Common Subexpression Elimination (CSE)

1

0

if ()

a∗b

12

0 a∗b

2

3 a∗b 3

March 2011

Uday Khedker, IIT Bombay

IITB

19/88

Research: In Search of Research

An Example of Research

• Area : Code optimization performed by compilers

Known Optimization : Common Subexpression Elimination (CSE) 0 1

if ()

a∗b

12

0 a∗b

• a and b are not modified along

2

paths 1 → 3 and 2 → 3

3 a∗b 3

March 2011

Uday Khedker, IIT Bombay

IITB

19/88

Research: In Search of Research

An Example of Research

• Area : Code optimization performed by compilers

Known Optimization : Common Subexpression Elimination (CSE) 0 1

if ()

a∗b

12

0 a∗b

• a and b are not modified along

paths 1 → 3 and 2 → 3

2 • Computation of a ∗ b in 3 is redundant

3 a∗b 3

March 2011

Uday Khedker, IIT Bombay

IITB

19/88

Research: In Search of Research

An Example of Research

• Area : Code optimization performed by compilers

Known Optimization : Common Subexpression Elimination (CSE) 0

if ()

0

• a and b are not modified along

paths 1 → 3 and 2 → 3

1 t = a ∗ b 1 2 t = a ∗ b 2 • Computation of a ∗ b in 3 is redundant 3

March 2011

t

3

• Previous value can be used

Uday Khedker, IIT Bombay

IITB

20/88

Research: In Search of Research

An Example of Research

• Lacuna with CSE: Cannot optimize partial redundancy

0

if ()

0

1 a∗b 1 2 a =5 2 3 a∗b 3

March 2011

Uday Khedker, IIT Bombay

IITB

20/88

Research: In Search of Research

An Example of Research

• Lacuna with CSE: Cannot optimize partial redundancy

0

if ()

0

• Computation of a ∗ b in 3 is

1 a∗b 1 2 a =5 2 3 a∗b 3

March 2011

Uday Khedker, IIT Bombay

IITB

20/88

Research: In Search of Research

An Example of Research

• Lacuna with CSE: Cannot optimize partial redundancy

0

if ()

0

1 a∗b 1 2 a =5 2

• Computation of a ∗ b in 3 is ◮

redundant along path 1 → 3, but . . .

3 a∗b 3

March 2011

Uday Khedker, IIT Bombay

IITB

20/88

Research: In Search of Research

An Example of Research

• Lacuna with CSE: Cannot optimize partial redundancy

0

if ()

0

1 a∗b 1 2 a =5 2 3 a∗b 3

March 2011

• Computation of a ∗ b in 3 is ◮

redundant along path 1 → 3, but . . .

◮

not redundant along path 2 → 3

Uday Khedker, IIT Bombay

IITB

21/88

Research: In Search of Research

An Example of Research

• Lacuna with CSE: Partial Redundancy Elimination

0 if () 0 1 a∗b 1 2 a =5 2 3 a∗b 3

March 2011

Uday Khedker, IIT Bombay

IITB

21/88

Research: In Search of Research

An Example of Research

• Lacuna with CSE: Partial Redundancy Elimination

0 if () 0 a=5 1 a∗b 1 2 a∗b 2

• Computation of a ∗ b in 3

becomes totally redundant

• Can be deleted

3 a∗b 3

March 2011

Uday Khedker, IIT Bombay

IITB

22/88

Research: In Search of Research

The Beauty of PRE

Subsumes Loop Invariant Movement

1

a=b∗c

2

a=b∗c 1

3

a=b∗c

March 2011

1

3

Uday Khedker, IIT Bombay

IITB

22/88

Research: In Search of Research

The Beauty of PRE

Subsumes Loop Invariant Movement

1

a=b∗c

1

2

a=b∗c 1

3

a=b∗c

1

a=b∗c 2

2

March 2011

3

1 a=b∗c 3

a=b∗c 1

Uday Khedker, IIT Bombay

IITB

22/88

Research: In Search of Research

The Beauty of PRE

Subsumes Loop Invariant Movement

1

a=b∗c

1

2

a=b∗c 1

3

a=b∗c

1

a=b∗c 2

2

March 2011

3

1 a=b∗c 3

a=b∗c 1

1

2

3

t=b∗c 1 a=t

a=b∗c

1

3

Uday Khedker, IIT Bombay

IITB

23/88

Research: In Search of Research

The Beauty of PRE Can be used for Strength Reduction and Induction Variable Elimination Original Computation

i =0

a = A[i ] i =i +1

Compiled Program i =0 t0 = base(A) t1 = t0 + i ∗ 4

t1 = t1 + 4 a = A[t1] i =i +1

• ∗ and + in the loop have been replaced by + • i = i + 1 in the loop has been eliminated March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9

• Theoretically predicted number : 144

10 11 12

March 2011

Uday Khedker, IIT Bombay

IITB

24/88

Research: In Search of Research

Lacuna with PRE Theoretical Complexity Number of (fixed order) traversals in the worst case?

1 6 2 3

• Information could flow along arbitrary

paths

7 4

8

5

9 10

• Theoretically predicted number : 144 • Actual iterations : 5

11 12 March 2011

Uday Khedker, IIT Bombay

IITB

25/88

Research: In Search of Research

Lacuna with PRE: Complexity Measure

• Complexity O(n2 ) traversals

Practical graphs may have upto 50 nodes ◮

Predicted number of traversals : 2,500

◮

Practical number of traversals : ≤ 5

• No explanation for about 14 years despite dozens of efforts • Not much experimentation with performing advanced optimizations

involving bidirectional dependency

March 2011

Uday Khedker, IIT Bombay

IITB

26/88

Research: In Search of Research

The Story at IIT Bombay

• The challenge of the professor

• The mysterious picture and the moment of reckoning

March 2011

Uday Khedker, IIT Bombay

IITB

27/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved Chemist

Cloth Shop

Doctor

Task Buy OTC (Over-The-Counter) medicine

March 2011

Tailor

X-Ray Clinic

U-Turns

Trips

None

1

Uday Khedker, IIT Bombay

IITB

27/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved Chemist

Cloth Shop

Doctor

Tailor

Task

X-Ray Clinic

U-Turns

Trips

Buy OTC (Over-The-Counter) medicine

None

1

Buy cloth. Give it to the tailor for stitching

None

1

March 2011

Uday Khedker, IIT Bombay

IITB

27/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved Chemist

Cloth Shop

Doctor

Tailor

Task

X-Ray Clinic

U-Turns

Trips

Buy OTC (Over-The-Counter) medicine

None

1

Buy cloth. Give it to the tailor for stitching

None

1

1

2

Buy medicine with doctor’s prescription

March 2011

Uday Khedker, IIT Bombay

IITB

27/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved Chemist

Cloth Shop

Doctor

Tailor

Task

X-Ray Clinic

U-Turns

Trips

Buy OTC (Over-The-Counter) medicine

None

1

Buy cloth. Give it to the tailor for stitching

None

1

Buy medicine with doctor’s prescription

1

2

Buy medicine with doctor’s prescription The diagnosis requires X-Ray

2

3

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

• Every “incompatible” edge traversal

Graph Traversal

6 2 3

⇒ One additional graph traversal

7 4

8

5

9 10

11 12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

√

• Every “incompatible” edge traversal

Graph Traversal

6 2 3

⇒ One additional graph traversal

7 4

8

5

9 10

• Max. Incompatible edge traversals

= Width of the graph = 0?

• Maximum number of traversals =

1 + Max. incompatible edge traversals

11 12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

Graph Traversal

×

√

2 3

• Every “incompatible” edge traversal

6

⇒ One additional graph traversal

7 4

8

5

9 10

• Max. Incompatible edge traversals

= Width of the graph = 1?

• Maximum number of traversals =

1 + Max. incompatible edge traversals

11 12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

Graph Traversal

× ×

√

2 3

• Every “incompatible” edge traversal

6

⇒ One additional graph traversal

7 4

8

5

9 10

• Max. Incompatible edge traversals

= Width of the graph = 2?

• Maximum number of traversals =

1 + Max. incompatible edge traversals

11 12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

Graph Traversal

× ×

√

2 3

⇒ One additional graph traversal

7 4

8

5

×

• Every “incompatible” edge traversal

6

9 10

• Max. Incompatible edge traversals

= Width of the graph = 3?

• Maximum number of traversals =

1 + Max. incompatible edge traversals

11 12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

Graph Traversal

× ×

√

2 3

⇒ One additional graph traversal

7 4

8

5

×

• Every “incompatible” edge traversal

6

9 10

11

√

• Max. Incompatible edge traversals

= Width of the graph = 3?

• Maximum number of traversals =

1 + Max. incompatible edge traversals

12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

Graph Traversal

× ×

√

2 3

√

7 4

8

5

×

• Every “incompatible” edge traversal

6

9 10

11

√

⇒ One additional graph traversal

• Max. Incompatible edge traversals

= Width of the graph = 3?

• Maximum number of traversals =

1 + Max. incompatible edge traversals

12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

Graph Traversal

× ×

√ √ 7 √

2 3

4

8

5

×

• Every “incompatible” edge traversal

6

9 10

11

√

⇒ One additional graph traversal

• Max. Incompatible edge traversals

= Width of the graph = 3?

• Maximum number of traversals =

1 + Max. incompatible edge traversals

12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

Graph Traversal

× ×

×

4

√ 7 √

8

5

×

• Every “incompatible” edge traversal

6

2 3

√

9 10

11

√

⇒ One additional graph traversal

• Max. Incompatible edge traversals

= Width of the graph = 4

• Maximum number of traversals =

1 + Max. incompatible edge traversals

12

March 2011

Uday Khedker, IIT Bombay

IITB

28/88

Research: In Search of Research

The Story at IIT Bombay : Mystery Solved

1

Graph Traversal

× ×

×

4

√ 7 √

8

5

×

• Every “incompatible” edge traversal

6

2 3

√

9 10

11

√

⇒ One additional graph traversal

• Max. Incompatible edge traversals

= Width of the graph = 4

• Maximum number of traversals =

1+4 = 5

12

March 2011

Uday Khedker, IIT Bombay

IITB

29/88

Research: In Search of Research

Looking Back

Key Idea A partial redundancy can be converted to total redundancy The compatibility or a lack or it, between the order of dependence and the order of traversal determines the number of traversals

March 2011

Its Significance Subsumes loop invariant movement without having to detect loops Bridges the gap between theoretical prediction and practical observation

Uday Khedker, IIT Bombay

Part 4

The Process of Research

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

March 2011

Uday Khedker, IIT Bombay

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

March 2011

Uday Khedker, IIT Bombay

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

March 2011

Uday Khedker, IIT Bombay

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

March 2011

• Relative stupidity Vs.

Productive stupidity

Uday Khedker, IIT Bombay

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

• Relative stupidity Vs.

Productive stupidity

• We are taught to feel bad

about relative stupidity

March 2011

Uday Khedker, IIT Bombay

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

• Relative stupidity Vs.

Productive stupidity

• We are taught to feel bad

about relative stupidity

• Productive stupidity:

March 2011

Uday Khedker, IIT Bombay

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

• Relative stupidity Vs.

Productive stupidity

• We are taught to feel bad

about relative stupidity

• Productive stupidity: ◮

March 2011

No research is possible unless we are willing to feel vulnerable and stupid

Uday Khedker, IIT Bombay

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

• Relative stupidity Vs.

Productive stupidity

• We are taught to feel bad

about relative stupidity

• Productive stupidity: ◮

◮

March 2011

No research is possible unless we are willing to feel vulnerable and stupid If we don’t feel stupid, we are not trying enough!

Uday Khedker, IIT Bombay

IITB

30/88

Research: The Process of Research

The Pursuit of Knowledge Very few of us know, how much we have to know, in order to know, how little we know

• Relative stupidity Vs.

Productive stupidity

• We are taught to feel bad

about relative stupidity

• Productive stupidity: ◮

◮

No research is possible unless we are willing to feel vulnerable and stupid If we don’t feel stupid, we are not trying enough!

It’s important to know what we know and what we don’t and be comfortable with it

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

March 2011

B

C

D

E

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

D

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

A

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

A

Breadth or Depth?

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

D

E

C

Achieve depth in one area and overall breadth

March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

31/88

Research: The Process of Research

Breadth Vs. Depth

A

B

A

C

C

D

E

D

Achieve depth in one area and overall breadth Other areas can be understood on need basis relatively quickly March 2011

Uday Khedker, IIT Bombay

IITB

32/88

Research: The Process of Research

Ability to Abstract and Modularize

• Different levels of abstraction and different granularities of

modularization

Example : Describing a car to ◮

a person who wants to travel in a car

◮

a person who wants to drive a car

◮

a person who wants to repair a car

◮

a person who wants to design a car

Each of the above views is correct and complete w.r.t to chosen level of abstraction

March 2011

Uday Khedker, IIT Bombay

IITB

33/88

Research: The Process of Research

From Confusion to Conviction

Start

March 2011

End

Uday Khedker, IIT Bombay

IITB

33/88

Research: The Process of Research

From Confusion to Conviction

Start

March 2011

Growing Confusion

End

Uday Khedker, IIT Bombay

IITB

33/88

Research: The Process of Research

From Confusion to Conviction

Start

March 2011

Growing Confusion

End

Uday Khedker, IIT Bombay

IITB

33/88

Research: The Process of Research

From Confusion to Conviction

Start

March 2011

Growing Confusion

Growing Conviction

End

Uday Khedker, IIT Bombay

IITB

33/88

Research: The Process of Research

From Confusion to Conviction

Start

March 2011

Growing Confusion

Growing Conviction

End

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Observation, Absorption

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Observation, Absorption

Analysis, Evaluation

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Observation, Absorption

Analysis, Evaluation

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Observation, Absorption

Analysis, Evaluation

Innovative Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Observation, Absorption

Analysis, Evaluation Could be a clean formulation of a problem or its solution Innovative Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Observation, Absorption

Analysis, Evaluation Could be a clean formulation of a problem or its solution

Ideation (Conscious Thinking) Innovative Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Observation, Absorption

Incubation (Subconscious Thinking)

Analysis, Evaluation Could be a clean formulation of a problem or its solution

Ideation (Conscious Thinking) Innovative Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Observation, Absorption

Incubation (Subconscious Thinking)

Analysis, Evaluation Could be a clean formulation of a problem or its solution

Ideation (Conscious Thinking) Innovative Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Synthesis (Building, Extending)

Observation, Absorption

Incubation (Subconscious Thinking)

Analysis, Evaluation Could be a clean formulation of a problem or its solution

Ideation (Conscious Thinking) Innovative Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Synthesis (Building, Extending)

Observation, Absorption

Incubation (Subconscious Thinking)

Analysis, Evaluation Could be a clean formulation of a problem or its solution

Ideation (Conscious Thinking) Innovative Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Synthesis (Building, Extending)

Observation, Absorption

Incubation (Subconscious Thinking)

Analysis, Evaluation Could be a clean formulation of a problem or its solution

Ideation (Conscious Thinking) Innovative Ideas

March 2011

Uday Khedker, IIT Bombay

IITB

34/88

Research: The Process of Research

The Birth of Innovative Ideas Synthesis (Building, Extending)

Observation, Absorption

Incubation (Subconscious Thinking)

Analysis, Evaluation Could be a clean formulation of a problem or its solution

Ideation (Conscious Thinking) Innovative Ideas

March 2011

End of the Beginning

Uday Khedker, IIT Bombay

IITB

35/88

Research: The Process of Research

The “S” Curve of Research Life Cycle

Quality or Maturity

Time/Effort

March 2011

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven

March 2011

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven • A product is typically based on a large number of ideas

March 2011

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven • A product is typically based on a large number of ideas • Industry ◮ Needs to combine results of many idea driven research efforts

March 2011

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven • A product is typically based on a large number of ideas • Industry ◮ Needs to combine results of many idea driven research efforts ◮ Faces the challenge of deployment much more significantly

March 2011

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven • A product is typically based on a large number of ideas • Industry ◮ Needs to combine results of many idea driven research efforts ◮ Faces the challenge of deployment much more significantly ◮ Focus on an independent idea may be an unaffordable luxury

March 2011

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven • A product is typically based on a large number of ideas • Industry ◮ Needs to combine results of many idea driven research efforts ◮ Faces the challenge of deployment much more significantly ◮ Focus on an independent idea may be an unaffordable luxury • Research in academia ◮

March 2011

Typically not bound by a particular product

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven • A product is typically based on a large number of ideas • Industry ◮ Needs to combine results of many idea driven research efforts ◮ Faces the challenge of deployment much more significantly ◮ Focus on an independent idea may be an unaffordable luxury • Research in academia ◮ ◮

March 2011

Typically not bound by a particular product Can afford to focus on ideas exclusively

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven • A product is typically based on a large number of ideas • Industry ◮ Needs to combine results of many idea driven research efforts ◮ Faces the challenge of deployment much more significantly ◮ Focus on an independent idea may be an unaffordable luxury • Research in academia ◮ ◮ ◮

March 2011

Typically not bound by a particular product Can afford to focus on ideas exclusively Can have longer gestation periods

Uday Khedker, IIT Bombay

IITB

36/88

Research: The Process of Research

Research in Industry Vs. Research in Academia Research in industry is product driven whereas research in academia is idea driven • A product is typically based on a large number of ideas • Industry ◮ Needs to combine results of many idea driven research efforts ◮ Faces the challenge of deployment much more significantly ◮ Focus on an independent idea may be an unaffordable luxury • Research in academia ◮ ◮ ◮

Typically not bound by a particular product Can afford to focus on ideas exclusively Can have longer gestation periods

Deliverables exist but of a different nature Deadlines are usually as strict

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask 4. Seek generality by removing specificities

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask 4. Seek generality by removing specificities 5. Seek symmetry by testing for duality

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask 4. Seek generality by removing specificities 5. Seek symmetry by testing for duality 6. Observe patterns in ideas

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask 4. Seek generality by removing specificities 5. Seek symmetry by testing for duality 6. Observe patterns in ideas 7. Distill the essence, refine your ideas

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask 4. Seek generality by removing specificities 5. Seek symmetry by testing for duality 6. Observe patterns in ideas 7. Distill the essence, refine your ideas 8. Distinguish the relevant from the irrelevant

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask 4. Seek generality by removing specificities 5. Seek symmetry by testing for duality 6. Observe patterns in ideas 7. Distill the essence, refine your ideas 8. Distinguish the relevant from the irrelevant 9. Build levels of abstractions and migrate between them

March 2011

Uday Khedker, IIT Bombay

IITB

37/88

Research: The Process of Research

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask 4. Seek generality by removing specificities 5. Seek symmetry by testing for duality 6. Observe patterns in ideas 7. Distill the essence, refine your ideas 8. Distinguish the relevant from the irrelevant 9. Build levels of abstractions and migrate between them 10. Mix deep thinking with routine mechanical work

March 2011

Uday Khedker, IIT Bombay

Part 5

Richard Hamming on Research

IITB

Research: Richard Hamming on Research

On the Role of Luck in Research Pasteur: “Luck favours a prepared mind.”

Hamming:“Our society frowns on people who set out to do really good work. You’re not supposed to; luck is supposed to descend on you and you do great things by chance. Well, that’s a kind of dumb thing to say.

I spoke earlier about planting acorns so that oaks will grow.

You can’t always know exactly where to be, but you can keep active in places where something might happen.

A prepared mind sooner or later finds something important and does it. So yes, it is luck. The particular thing you do is luck, but that you do something is not. ” Luck is when preparation meets opportunity March 2011

Uday Khedker, IIT Bombay

38/88

IITB

Research: Richard Hamming on Research

39/88

Turning Difficulties into Opportunities in Research

Knowing right questions is the first step in knowing the answers Hamming: “Often the great scientists, by turning the problem around a bit, changed a defect to an asset.

Hamming, you think the machines can do practically everything. Why can’t you make them write programs?” What appeared at first to me as a defect forced me into automatic programming very early. What appears to be a fault, often, by a change of viewpoint, turns out to be one of the greatest assets you can have. ”

March 2011

Uday Khedker, IIT Bombay

Research: Richard Hamming on Research

IITB

The Role of Hard Work in Research

• Research is 1% inspiration and 99% perspiration

Newton: “If others would think as hard as I did, they would get similar results.”

March 2011

Uday Khedker, IIT Bombay

40/88

Research: Richard Hamming on Research

IITB

40/88

The Role of Hard Work in Research

• Research is 1% inspiration and 99% perspiration

Newton: “If others would think as hard as I did, they would get similar results.”

• Continuous consolidation works like compound interest

Hamming: “Given two people of approximately the same ability and one person who works ten percent more than the other, the latter will more than twice outproduce the former.

The more you know, the more you learn; the more you learn, the more you can do; the more you can do, the more the opportunity . . . ”

March 2011

Uday Khedker, IIT Bombay

Research: Richard Hamming on Research

IITB

41/88

On the Role of Intelligence and Courage in Research Hamming: “How about having lots of ‘brains?’ It sounds good. . . . But great work is something else than mere brains.

One of the characteristics you see, and many people have it including great scientists, is that usually when they were young they had independent thoughts and had the courage to pursue them.

Once you get your courage up and believe that you can do important problems, then you can. If you think you can’t, almost surely you are not going to. . . . That is the characteristic of great scientists; they have courage. They will go forward under incredible circumstances; they think and continue to think. ”

March 2011

Uday Khedker, IIT Bombay

Research: Richard Hamming on Research

IITB

42/88

On the Role of Drive and Commitment in Research

Hamming: “Well, one of the reasons is drive and commitment. The people who do great work with less ability but who are committed to it, get more done that those who have great skill and dabble in it, who work during the day and go home and do other things and come back and work the next day. They don’t have the deep commitment that is apparently necessary for really first-class work. ”

March 2011

Uday Khedker, IIT Bombay

Research: Richard Hamming on Research

IITB

43/88

On the Role of Commitment and Creativity in Research Hamming: “If you are deeply immersed and committed to a topic, day after day after day, your subconscious has nothing to do but work on your problem. And so you wake up one morning, or on some afternoon, and there’s the answer. For those who don’t get committed to their current problem, the subconscious goofs off on other things and doesn’t produce the big result.

So . . . you don’t let anything else get the center of your attention - you keep your thoughts on the problem. Keep your subconscious starved so it has to work on your problem, so you can sleep peacefully and get the answer in the morning, free. ”

March 2011

Uday Khedker, IIT Bombay

Research: Richard Hamming on Research

IITB

Committment, Hard Work, and Preparation Enhance Creativity

• Keep your antennas tuned

• You never know when and where the signals come from

March 2011

Uday Khedker, IIT Bombay

44/88

IITB

Research: Richard Hamming on Research

45/88

On the Role of Uncertainty in Research

Hamming: “Most people like to believe something is or is not true. Great scientists tolerate ambiguity very well. They believe the theory enough to go ahead; they doubt it enough to notice the errors and faults so they can step forward and create the new replacement theory. If you believe too much you’ll never notice the flaws; if you doubt too much you won’t get started.

It requires a lovely balance. But most great scientists are well aware of why their theories are true and they are also well aware of some slight misfits which don’t quite fit and they don’t forget it. ”

March 2011

Uday Khedker, IIT Bombay

IITB

Research: Richard Hamming on Research

46/88

Incremental Research Vs. Fundamental Research

Hamming: “Most great scientists know many important problems. They have something between 10 and 20 important problems for which they are looking for an attack. And when they see a new idea come up, one hears them say ‘Well that bears on this problem.’

The great scientists, when an opportunity opens up, get after it and they pursue it. They drop all other things. They get rid of other things and they get after an idea because they had already thought the thing through. Their minds are prepared; they see the opportunity and they go after it. ”

March 2011

Uday Khedker, IIT Bombay

Part 6

Specific Questions Related to Ph.D.

Research: Specific Questions Related to Ph.D.

IITB

On the Role of Ph.D. in Teaching

Q Is Ph.D. necessary for teaching?

March 2011

Uday Khedker, IIT Bombay

47/88

Research: Specific Questions Related to Ph.D.

IITB

47/88

On the Role of Ph.D. in Teaching

Q Is Ph.D. necessary for teaching?

A Ph.D. is not a degree, Ph.D. is a state of mind A state in which ◮ ◮ ◮

one does not feel threatened by the unknown one derives immense satisfaction in a journey from known to unknown one may depend on others for questions, but not necessarily for answers

A Ph.D. is attainment of a certain level of maturity of ones mental faculties A Ph.D. enhances quick self learning abilities A Ph.D. makes a person a much better learner

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

On Deciding a Research Area

Q Should I do research in the knowledge domain that I am familiar with or should I work in a “hot” area?

March 2011

Uday Khedker, IIT Bombay

48/88

Research: Specific Questions Related to Ph.D.

IITB

On Deciding a Research Area

Q Should I do research in the knowledge domain that I am familiar with or should I work in a “hot” area?

A In general it is easier to work in an area in which you have ◮ ◮

a good background and good insights personal interest

A What matters is the group with which you work. Having people with whom you can discuss your work is a very big advantage.

A Expertise of your guide is an important consideration.

March 2011

Uday Khedker, IIT Bombay

48/88

Research: Specific Questions Related to Ph.D.

IITB

49/88

On Deciding Ph.D. Topic

Q Once the research area is finalized, how should one identify specific research problems?

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

49/88

On Deciding Ph.D. Topic

Q Once the research area is finalized, how should one identify specific research problems?

A Read research papers in well known journals. Be selective about conference proceedings. Try to find scope for improvement in the papers you read.

A Attend good conferences. Talk to people.

A Interact with known researchers working in that area. Visit their web pages, contact them over email. Ask them what they are working on and why.

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

On Deciding Ph.D. Topic Q How should one identify and select National/International conferences? Is there any ranking of conferences?

March 2011

Uday Khedker, IIT Bombay

50/88

Research: Specific Questions Related to Ph.D.

IITB

On Deciding Ph.D. Topic Q How should one identify and select National/International conferences? Is there any ranking of conferences?

A CORE ranking (www.core.edu.au) is a good starting point. Ask the people who are active in that area.

A Read past proceedings. If proceedings are not published, forget about the conference.

A Find out if the conference has a reasonable review process.

March 2011

Uday Khedker, IIT Bombay

50/88

Research: Specific Questions Related to Ph.D.

IITB

On Deciding Ph.D. Topic Q How should one identify and select National/International conferences? Is there any ranking of conferences?

A CORE ranking (www.core.edu.au) is a good starting point. Ask the people who are active in that area.

A Read past proceedings. If proceedings are not published, forget about the conference.

A Find out if the conference has a reasonable review process.

Q How to do literature survey on the finalized area of interest? Are there any thumb rules or guidelines for selecting number of papers for literature survey?

March 2011

Uday Khedker, IIT Bombay

50/88

Research: Specific Questions Related to Ph.D.

IITB

On Deciding Ph.D. Topic Q How should one identify and select National/International conferences? Is there any ranking of conferences?

A CORE ranking (www.core.edu.au) is a good starting point. Ask the people who are active in that area.

A Read past proceedings. If proceedings are not published, forget about the conference.

A Find out if the conference has a reasonable review process.

Q How to do literature survey on the finalized area of interest? Are there any thumb rules or guidelines for selecting number of papers for literature survey?

A Read research papers in relevant journals/conferences of repute.

A There is no thumb rule for number of papers.

March 2011

Uday Khedker, IIT Bombay

50/88

Research: Specific Questions Related to Ph.D.

IITB

On Deciding Ph.D. Topic

Q How does one define a Ph.D. problem?

March 2011

Uday Khedker, IIT Bombay

51/88

Research: Specific Questions Related to Ph.D.

IITB

51/88

On Deciding Ph.D. Topic

Q How does one define a Ph.D. problem?

A In general, it is preferable to let the problem definition evolve over time. Some times the problem may turn out to be much harder than anticipated while some times it might be too easy.

A Begin with a tentative problem. It is best to let the scope and the depth emerge naturally as one’s study matures.

A Interacting with an active group, asking for sub-problems that could contribute to the research goals of the group is a big help.

March 2011

Uday Khedker, IIT Bombay

IITB

Research: Specific Questions Related to Ph.D.

On Deciding Ph.D. Topic

A common mistake: Fixing a problem and then seeking guidance for Ph.D. on that problem.

March 2011

Uday Khedker, IIT Bombay

52/88

Research: Specific Questions Related to Ph.D.

IITB

52/88

On Deciding Ph.D. Topic

A common mistake: Fixing a problem and then seeking guidance for Ph.D. on that problem.

Fixing a problem only shows your initiative and abilities. However: • your perception of relevance or significance of the problem may be

flawed.

• If your problem fits into a broader research agenda, there are better

prospects.

• Prospective guides have their own research agendas and are busy.

Your problem must be very very attractive for them to deviate from their existing commitments.

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

On Deciding Ph.D. Topic

Q How do I identify the right guide for my Ph.D.?

March 2011

Uday Khedker, IIT Bombay

53/88

Research: Specific Questions Related to Ph.D.

IITB

On Deciding Ph.D. Topic

Q How do I identify the right guide for my Ph.D.?

A Find out someone who is actively working in that area. Find out from others how well is she/he known in the area.

A Find out someone who can give a lot of time to you.

A Find out someone with whom you are very comfortable.

March 2011

Uday Khedker, IIT Bombay

53/88

Research: Specific Questions Related to Ph.D.

IITB

54/88

On the Process of Ph.D.

Q Does research (Ph.D.) starts with proof of concept? How does one define proof of concept? How does one create it?

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

54/88

On the Process of Ph.D.

Q Does research (Ph.D.) starts with proof of concept? How does one define proof of concept? How does one create it?

A Proof of concept refers to a demonstration of the validity, meaningfulness, or usefulness of an idea. It could be in terms of an implementation, a conceptual reasoning, or a design of a solution.

A Ph.D. begins with questions, proof of concept tries to validate possible answers. So it comes somewhere in the middle (or even at the end) of a Ph.D.

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

On the Process of Ph.D.

Q How to identify the requirements needed? Does it have relation with problem defined? How to extract requirements such as tools, databases, methods and technology from the problem definition?

March 2011

Uday Khedker, IIT Bombay

55/88

Research: Specific Questions Related to Ph.D.

IITB

55/88

On the Process of Ph.D.

Q How to identify the requirements needed? Does it have relation with problem defined? How to extract requirements such as tools, databases, methods and technology from the problem definition?

A Requirements spring up in the course of discovery. It is very hard to specify requirements a priori. Some broad and general requirements may be easy to visualize at the start.

A Ph.D. is a game of ideas. It is not a game of technology. Even an experimental Ph.D. begins with a speculation.

What is common to all Ph.D.’s is a cycle of: Speculate, design, apply/perform experiment, observe, interpret, infer and repeat.

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

On the Process of Ph.D.

Q What is state-of-art Technology? How I should identify that the selected topic has it?

March 2011

Uday Khedker, IIT Bombay

56/88

Research: Specific Questions Related to Ph.D.

IITB

56/88

On the Process of Ph.D.

Q What is state-of-art Technology? How I should identify that the selected topic has it?

A State-of-art refers to the frontiers of technology. Remain in touch with the people who are working in the area. Keep reading. Attend conferences.

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

56/88

On the Process of Ph.D.

Q What is state-of-art Technology? How I should identify that the selected topic has it?

A State-of-art refers to the frontiers of technology. Remain in touch with the people who are working in the area. Keep reading. Attend conferences.

Q When can I conclude that my Ph.D. is over?

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

56/88

On the Process of Ph.D.

Q What is state-of-art Technology? How I should identify that the selected topic has it?

A State-of-art refers to the frontiers of technology. Remain in touch with the people who are working in the area. Keep reading. Attend conferences.

Q When can I conclude that my Ph.D. is over?

A When you find that after you give talks on your work, people feel you have done something interesting.

March 2011

Uday Khedker, IIT Bombay

Research: Specific Questions Related to Ph.D.

IITB

On the Process of Ph.D.

Q What are the common problems all Ph.D. students face?

March 2011

Uday Khedker, IIT Bombay

57/88

Research: Specific Questions Related to Ph.D.

IITB

On the Process of Ph.D.

Q What are the common problems all Ph.D. students face?

A Difficulty in dealing with uncertainty.

A Keeping up motivation.

A Handling frustration.

A Realizing the importance (or unimportance) of particular ideas/developments.

Common problems across most human pursuits!

March 2011

Uday Khedker, IIT Bombay

57/88

Part 7

Writing A Good Research Paper

Research: Writing A Good Research Paper

IITB

So You Want to Write A Good Paper?

• Writing good papers is important for your research ◮ ◮ ◮

For your own understanding For communicating your ideas to others For getting feedback from others

A paper is a mirror of your ideas • A research paper is a public document • Many good papers have had their initial versions rejected • Which paper qualifies as a good paper?

A paper you would be proud of even after a decade or so!

March 2011

Uday Khedker, IIT Bombay

58/88

Research: Writing A Good Research Paper

IITB

Why Should You Care About Good Writing?

• Writing helps you understand your ideas better • Writing helps you refine your ideas:

Write, read, revise. Repeat as long as you can

• Writing is a creative process and a big source of satisfaction

March 2011

Uday Khedker, IIT Bombay

59/88

Research: Writing A Good Research Paper

IITB

Our Focus

• Significant distinction between ◮ How to write a good paper? ◮

How does a good paper look like?

• Process of writing Vs. Product of writing

March 2011

Uday Khedker, IIT Bombay

60/88

Research: Writing A Good Research Paper

IITB

Our Focus

• Significant distinction between ◮ How to write a good paper? ◮

How does a good paper look like?

• Process of writing Vs. Product of writing • We focus on the product rather than the process

• Assumptions: ◮ Process is goal oriented. Once goals are clear, process is clear ◮

March 2011

Repeat: Write, read, review, refine, revise . . .

Uday Khedker, IIT Bombay

60/88

Research: Writing A Good Research Paper

IITB

What is a Research Paper?

Does a research paper describe • A software?

• Design of a software?

• A piece of hardware?

• A theorem?

• A proof?

• Empirical measurements?

• ...

March 2011

Uday Khedker, IIT Bombay

61/88

Research: Writing A Good Research Paper

IITB

What is a Research Paper?

Does a research paper describe • A software?

• Design of a software?

• A piece of hardware?

• A theorem?

• A proof?

• Empirical measurements?

• ...

A research paper describes an idea!

March 2011

Uday Khedker, IIT Bombay

61/88

Research: Writing A Good Research Paper

IITB

What is a Research Paper?

• Cannot be a complete description of work

(process/product/modelling/reasoning/postulate/evidence) Must omit many details

• Should describe a logically complete idea

Or a collection of logically complete related ideas

• Must embody a running theme that forms the essence of an idea • This theme must be explicated in the paper

March 2011

Uday Khedker, IIT Bombay

62/88

Research: Writing A Good Research Paper

IITB

What is a Research Paper?

• Cannot be a complete description of work

(process/product/modelling/reasoning/postulate/evidence) Must omit many details

• Should describe a logically complete idea

Or a collection of logically complete related ideas

• Must embody a running theme that forms the essence of an idea • This theme must be explicated in the paper

Right choice of form and content is important

March 2011

Uday Khedker, IIT Bombay

62/88

IITB

Research: Writing A Good Research Paper

What is a Research Paper?

Net information content of your paper ∆ = Information in your paper − Relevant Information in references

March 2011

Uday Khedker, IIT Bombay

63/88

Research: Writing A Good Research Paper

IITB

What is a Research Paper?

Net information content of your paper ∆ = Information in your paper − Relevant Information in references ∆>0 ∆=0 ∆<0

March 2011

Uday Khedker, IIT Bombay

63/88

Research: Writing A Good Research Paper

IITB

What is a Research Paper?

Net information content of your paper ∆ = Information in your paper − Relevant Information in references ∆>0

Your paper makes research contributions

∆=0 ∆<0

March 2011

Uday Khedker, IIT Bombay

63/88

Research: Writing A Good Research Paper

IITB

What is a Research Paper?

Net information content of your paper ∆ = Information in your paper − Relevant Information in references ∆>0

Your paper makes research contributions

∆=0

Your paper is a survey paper

∆<0

March 2011

Uday Khedker, IIT Bombay

63/88

Research: Writing A Good Research Paper

IITB

What is a Research Paper?

Net information content of your paper ∆ = Information in your paper − Relevant Information in references

March 2011

∆>0

Your paper makes research contributions

∆=0

Your paper is a survey paper

∆<0

Why do you want to write a paper?

Uday Khedker, IIT Bombay

63/88

Research: Writing A Good Research Paper

IITB

Ingredients of Good Research

â&#x20AC;˘ Innovation

March 2011

Uday Khedker, IIT Bombay

64/88

Research: Writing A Good Research Paper

IITB

Ingredients of Good Research

• Innovation • Aesthetics

March 2011

Uday Khedker, IIT Bombay

64/88

Research: Writing A Good Research Paper

IITB

Ingredients of Good Research

• Innovation • Aesthetics • Other important aspects :

March 2011

◮

Completeness

◮

Rigour

◮

Empirical demonstration

◮

Effective communication

Uday Khedker, IIT Bombay

64/88

Research: Writing A Good Research Paper

IITB

Ingredients of Good Survey

â&#x20AC;˘ Have you identified all key ideas?

March 2011

Uday Khedker, IIT Bombay

65/88

Research: Writing A Good Research Paper

IITB

Ingredients of Good Survey

• Have you identified all key ideas?

• Have you distilled the essence of key ideas?

March 2011

◮

Why are they important?

◮

Why should they be considered key ideas?

Uday Khedker, IIT Bombay

65/88

Research: Writing A Good Research Paper

IITB

Ingredients of Good Survey

• Have you identified all key ideas?

• Have you distilled the essence of key ideas? ◮

Why are they important?

◮

Why should they be considered key ideas?

• Have you illustrated key ideas?

March 2011

◮

New examples

◮

New pictures

◮

New applications

◮

Better explanations

Uday Khedker, IIT Bombay

65/88

Research: Writing A Good Research Paper

IITB

66/88

Intuition and Rigour in a Paper

• Ideas should be presented at two levels: ◮ Intuition The key idea − Distilling the idea to its essence − Emphasizing the most important aspects (ignoring some details) − Using representative examples (even if all aspects are not illustrated) ◮

Rigour Ensuring that there is no possibility of a different interpretation − Plugging all the holes through formalism or reasoning − Presenting convincing empirical evidence

• Description of ideas at these two levels should be interleaved

March 2011

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

Presenting Evolution of Ideas • Evolution in discovery is different from evolution in presentation • Evolution in presentation

March 2011

◮

Perspective

◮

Problem Definition

◮

Key Ideas

◮

Key contributions

◮

Development of the key ideas

◮

Experiments/Measurements/Results

◮

Conclusions

Uday Khedker, IIT Bombay

67/88

Research: Writing A Good Research Paper

IITB

Presenting Evolution of Ideas • Evolution in discovery is different from evolution in presentation • Evolution in presentation ◮

◮

March 2011

Perspective Motivation, Background, Challenges Problem Definition

◮

Key Ideas

◮

Key contributions

◮

Development of the key ideas

◮

Experiments/Measurements/Results

◮

Conclusions

Uday Khedker, IIT Bombay

67/88

Research: Writing A Good Research Paper

IITB

Presenting Evolution of Ideas • Evolution in discovery is different from evolution in presentation • Evolution in presentation ◮

◮

◮

March 2011

Perspective Motivation, Background, Challenges Problem Definition Why is this an important problem?

Key Ideas

◮

Key contributions

◮

Development of the key ideas

◮

Experiments/Measurements/Results

◮

Conclusions

Uday Khedker, IIT Bombay

67/88

Research: Writing A Good Research Paper

IITB

Presenting Evolution of Ideas • Evolution in discovery is different from evolution in presentation • Evolution in presentation ◮

◮

◮

◮

March 2011

Perspective Motivation, Background, Challenges Problem Definition Why is this an important problem?

Key Ideas Basis of the hope of a solution, Approach of the solution, Key contributions

◮

Development of the key ideas

◮

Experiments/Measurements/Results

◮

Conclusions

Uday Khedker, IIT Bombay

67/88

Research: Writing A Good Research Paper

IITB

Presenting Evolution of Ideas • Evolution in discovery is different from evolution in presentation • Evolution in presentation ◮

◮

◮

◮

◮

March 2011

Perspective Motivation, Background, Challenges Problem Definition Why is this an important problem?

Key Ideas Basis of the hope of a solution, Approach of the solution, Key contributions Your claim to fame Development of the key ideas

◮

Experiments/Measurements/Results

◮

Conclusions

Uday Khedker, IIT Bombay

67/88

Research: Writing A Good Research Paper

IITB

Presenting Evolution of Ideas • Evolution in discovery is different from evolution in presentation • Evolution in presentation ◮

◮

◮

◮

◮

◮

◮

March 2011

Perspective Motivation, Background, Challenges Problem Definition Why is this an important problem?

Key Ideas Basis of the hope of a solution, Approach of the solution, Key contributions Your claim to fame Development of the key ideas You have succeeded if readers reach here!

Experiments/Measurements/Results Conclusions

Uday Khedker, IIT Bombay

67/88

Research: Writing A Good Research Paper

IITB

Presenting Evolution of Ideas • Evolution in discovery is different from evolution in presentation • Evolution in presentation ◮

◮

◮

◮

◮

◮

◮

March 2011

Perspective Motivation, Background, Challenges Problem Definition Why is this an important problem?

Key Ideas Basis of the hope of a solution, Approach of the solution, Key contributions Your claim to fame Development of the key ideas You have succeeded if readers reach here!

Experiments/Measurements/Results Support and elaborate your key contributions Conclusions

Uday Khedker, IIT Bombay

67/88

Research: Writing A Good Research Paper

IITB

Presenting Evolution of Ideas • Evolution in discovery is different from evolution in presentation • Evolution in presentation ◮

◮

◮

◮

◮

◮

◮

March 2011

Perspective Motivation, Background, Challenges Problem Definition Why is this an important problem?

Key Ideas Basis of the hope of a solution, Approach of the solution, Key contributions Your claim to fame Development of the key ideas You have succeeded if readers reach here!

Experiments/Measurements/Results Support and elaborate your key contributions Conclusions Honest analysis of work done, The Moral of the story Uday Khedker, IIT Bombay

67/88

Research: Writing A Good Research Paper

IITB

The Structure of a Paper

• Title

• Abstract

• Introduction, Motivation

• Background, Related work • Main Contents

• Results, Conclusions • Future Work • References

March 2011

Uday Khedker, IIT Bombay

68/88

Research: Writing A Good Research Paper

IITB

How to Write a Good Title?

â&#x20AC;˘ Should be precise and inviting

March 2011

Uday Khedker, IIT Bombay

69/88

Research: Writing A Good Research Paper

IITB

How to Write a Good Title?

• Should be precise and inviting • Examples of bad titles ◮ A Software ◮ A theorem and its proof ◮ Wireless networks ◮ Partial redundancy elimination in presence of critical edges for practical imperative programs with recursion and large number of functions and calls through function pointers

March 2011

Uday Khedker, IIT Bombay

69/88

Research: Writing A Good Research Paper

IITB

How to Write a Good Title?

• Should be precise and inviting • Examples of bad titles ◮ A Software ◮ A theorem and its proof ◮ Wireless networks ◮ Partial redundancy elimination in presence of critical edges for practical imperative programs with recursion and large number of functions and calls through function pointers • Possible good titles ◮ A software based control system for . . . ◮ On XYZ theorem ◮ Reducing congestion in wireless networks ◮ Partial redundancy elimination in extreme situations

March 2011

Uday Khedker, IIT Bombay

69/88

Research: Writing A Good Research Paper

IITB

70/88

How to Write a Good Abstract

• Should be a succinct and stand alone description ◮ ◮

March 2011

Reading the rest of the paper should not be necessary to get the gist Self-containment only at a high level of description

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

70/88

How to Write a Good Abstract

• Should be a succinct and stand alone description ◮ ◮

Reading the rest of the paper should not be necessary to get the gist Self-containment only at a high level of description

• An abstract is neither a summary nor an outline of the paper

March 2011

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

70/88

How to Write a Good Abstract

• Should be a succinct and stand alone description ◮ ◮

Reading the rest of the paper should not be necessary to get the gist Self-containment only at a high level of description

• An abstract is neither a summary nor an outline of the paper

• A checklist: Motivation, problem statement, approach, results,

conclusions

March 2011

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

70/88

How to Write a Good Abstract

• Should be a succinct and stand alone description ◮ ◮

Reading the rest of the paper should not be necessary to get the gist Self-containment only at a high level of description

• An abstract is neither a summary nor an outline of the paper

• A checklist: Motivation, problem statement, approach, results,

conclusions

• Common mistakes ◮ ◮ ◮ ◮

March 2011

Too verbose Too long Too short Omitting essential details

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

How About this Abstract?

“We worked in Computer Science. We proved some theorems.

Some were big, some were small. Big theorems had big proof, small theorems had small proofs. We tried to connect the proofs to the theorems. Sometimes we succeeded, sometimes we didn’t. By then, the time for submission had arrived, so we submitted the paper . . . ” M. Leunen and R. Lipton. “How to Have Your Abstract Rejected”.

March 2011

Uday Khedker, IIT Bombay

71/88

IITB

Research: Writing A Good Research Paper

How to Write a Good Introduction?

What does a good introduction contain?

March 2011

Uday Khedker, IIT Bombay

72/88

Research: Writing A Good Research Paper

IITB

How to Write a Good Introduction?

What does a good introduction contain?

• More details of motivation and problem statement ◮ ◮ ◮

March 2011

From general to specific Supported by concrete examples, puzzles, mysteries, Contextualizing the problem

Uday Khedker, IIT Bombay

72/88

Research: Writing A Good Research Paper

IITB

How to Write a Good Introduction?

What does a good introduction contain?

• More details of motivation and problem statement ◮ ◮ ◮

From general to specific Supported by concrete examples, puzzles, mysteries, Contextualizing the problem

• Importance of the problem ◮ Significance of the outcome, ◮ Intellectual challenge, other difficulties Why would simple approaches not work?

March 2011

Uday Khedker, IIT Bombay

72/88

Research: Writing A Good Research Paper

IITB

How to Write a Good Introduction?

What does a good introduction contain?

• More details of motivation and problem statement ◮ ◮ ◮

From general to specific Supported by concrete examples, puzzles, mysteries, Contextualizing the problem

• Importance of the problem ◮ Significance of the outcome, ◮ Intellectual challenge, other difficulties Why would simple approaches not work?

• Overview of proposed approach ◮ ◮

March 2011

A sketch of main proof, algorithm, key idea Novelty of the proposed approach

Uday Khedker, IIT Bombay

72/88

Research: Writing A Good Research Paper

IITB

How to Write a Good Introduction?

What does a good introduction contain?

• More details of motivation and problem statement ◮ ◮ ◮

From general to specific Supported by concrete examples, puzzles, mysteries, Contextualizing the problem

• Importance of the problem ◮ Significance of the outcome, ◮ Intellectual challenge, other difficulties Why would simple approaches not work?

• Overview of proposed approach ◮ ◮

A sketch of main proof, algorithm, key idea Novelty of the proposed approach

• Outline of the paper March 2011

Uday Khedker, IIT Bombay

72/88

Research: Writing A Good Research Paper

IITB

How to Write a Good Introduction?

Why is a good introduction important? â&#x20AC;˘ First chance of making first impression

Writing style, overall quality, analytical skills, confidence in ideas

March 2011

Uday Khedker, IIT Bombay

73/88

Research: Writing A Good Research Paper

IITB

73/88

How to Write a Good Introduction?

Why is a good introduction important? â&#x20AC;˘ First chance of making first impression

Writing style, overall quality, analytical skills, confidence in ideas

â&#x20AC;˘ Last chance of making the reader want to read your paper in details

March 2011

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

How to Describe Background and Related Work?

• Not just what? but also why?, why not?, how?, how else? etc. ◮ Not just a list of summaries ◮ Analysis, comparison, strengths, limitations • Organized by ideas rather than by references

March 2011

Uday Khedker, IIT Bombay

74/88

Research: Writing A Good Research Paper

IITB

Main Contents

• Answer important questions ◮ Questions: Problem, Theorem, . . . ◮ Answers: Algorithm/Protocol/Formulation/Proof/Counter example • Elaborate and justify the type of contribution:

Beauty, utility, or both

• Use formalism but don’t hide ideas behind notation

Introduce notation, terminologies only where required

• Describe experiments, measurements

March 2011

Uday Khedker, IIT Bombay

75/88

Research: Writing A Good Research Paper

IITB

How to Write Conclusions?

â&#x20AC;˘ Not a just summary

March 2011

Uday Khedker, IIT Bombay

76/88

Research: Writing A Good Research Paper

IITB

How to Write Conclusions?

• Not a just summary

• Different from abstract and introduction

March 2011

Uday Khedker, IIT Bombay

76/88

Research: Writing A Good Research Paper

IITB

76/88

How to Write Conclusions?

• Not a just summary

• Different from abstract and introduction • Represents the moral of the story ◮ A high level description of the significance of ideas, what they could further lead to Last chance to highlight importance ◮ Analysis of strength and limitations ◮ What the readers should remember after they have forgotten the details

March 2011

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

76/88

How to Write Conclusions?

• Not a just summary

• Different from abstract and introduction • Represents the moral of the story ◮ A high level description of the significance of ideas, what they could further lead to Last chance to highlight importance ◮ Analysis of strength and limitations ◮ What the readers should remember after they have forgotten the details • Brings back the general level ◮ ◮

March 2011

Abstract → Introduction → Details: Journey from general to specific Details → Conclusions: Return journey from specific to general

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

77/88

How to Write References?

• Reference or Bibliography? ◮ References: List of sources that you actually cite in your paper ◮ Bibliography: List of all related publications • Each item in the list must have at least the following fields:

Title, Author(s), Journal/Proceedings, Publisher, Year

• URLs don’t have a publication date, hence say when accessed it last • Follow the style specified by the publisher

March 2011

Uday Khedker, IIT Bombay

Research: Writing A Good Research Paper

IITB

Elements of Writing

• Be careful about the language: Grammar, sentence formations,

spellings, punctuation etc.

• Each paragraph should represent a specific idea • Smooth transition from ◮ One paragraph to the next ◮ One sentence to the next

Plenty of help available on Internet

March 2011

Uday Khedker, IIT Bombay

78/88

Research: Writing A Good Research Paper

IITB

Writing a Good Sentence â&#x20AC;˘ Write short sentences

March 2011

Uday Khedker, IIT Bombay

79/88

Research: Writing A Good Research Paper

IITB

Writing a Good Sentence • Write short sentences • Be precise ◮

March 2011

“The problem stated above is difficult” Difficult for whom? NP-complete? Believed by you? Believed by others? Proved by someone?

Uday Khedker, IIT Bombay

79/88

Research: Writing A Good Research Paper

IITB

Writing a Good Sentence • Write short sentences • Be precise ◮

“The problem stated above is difficult” Difficult for whom? NP-complete? Believed by you? Believed by others? Proved by someone?

• Avoid unnecessary words in a sentence ◮

◮

March 2011

“This is a subject which is liked by people” “This subject is liked by people” “The logger program notes the fact that the event has happened” “The logger notes the event”

Uday Khedker, IIT Bombay

79/88

Research: Writing A Good Research Paper

IITB

Writing a Good Sentence • Write short sentences • Be precise ◮

“The problem stated above is difficult” Difficult for whom? NP-complete? Believed by you? Believed by others? Proved by someone?

• Avoid unnecessary words in a sentence ◮

◮

“This is a subject which is liked by people” “This subject is liked by people” “The logger program notes the fact that the event has happened” “The logger notes the event”

• Convert nouns to verbs ◮ “His verticality changed to horizontality” “He fell down” ◮ “This paragraph provides a conclusion of the description” “This paragraph concludes the description” March 2011

Uday Khedker, IIT Bombay

79/88

Research: Writing A Good Research Paper

IITB

Good Writing is an Art

• Rules are not absolute

• Break rules if you must, but understand why it is necessary • How to be a good writer? Read a lot

March 2011

◮

Best resource: “Elements of Style” by Strunk and White

◮

Plenty of help available on Internet

Uday Khedker, IIT Bombay

80/88

Part 8

Conclusions

IITB

81/88

Research: Conclusions

The Role of Research in Long Satisfying Technical Career

• In a rapidly changing field like CS/IT, quick self learning is the most

important ability

• Quick self learning is enhanced significantly by exposure to research • Research experience is likely to become more and more important in

future

March 2011

Uday Khedker, IIT Bombay

IITB

82/88

Research: Conclusions

The Essence of Research

• Research is a game of significant ideas

Even an experimental research begins with an observation and speculation

• The significance of ideas could lie in any of the following: ◮ ◮ ◮

Beauty Utility Innovation or Enhancement of knowledge

• Research is often a cycle of:

Speculate, design, apply/perform experiment, observe, interpret, infer and repeat

March 2011

Uday Khedker, IIT Bombay

IITB

83/88

Research: Conclusions

Ingredients of Good Research

• Innovation • Aesthetics • Other important aspects :

March 2011

◮

Completeness

◮

Rigour

◮

Empirical demonstration

◮

Effective communication

Uday Khedker, IIT Bombay

IITB

84/88

Research: Conclusions

What Does it Take to Be a Researcher?

â&#x20AC;˘ What we all may already possess

March 2011

Uday Khedker, IIT Bombay

IITB

84/88

Research: Conclusions

What Does it Take to Be a Researcher?

â&#x20AC;˘ What we all may already possess

Motivation, Curiosity, Creativity, Perseverance, Good Grasp

March 2011

Uday Khedker, IIT Bombay

IITB

84/88

Research: Conclusions

What Does it Take to Be a Researcher?

• What we all may already possess

Motivation, Curiosity, Creativity, Perseverance, Good Grasp • What we may have to acquire with effort

March 2011

Uday Khedker, IIT Bombay

IITB

84/88

Research: Conclusions

What Does it Take to Be a Researcher?

â&#x20AC;˘ What we all may already possess

Motivation, Curiosity, Creativity, Perseverance, Good Grasp â&#x20AC;˘ What we may have to acquire with effort

Strong Background Enhances the effectiveness of all the above traits, particularly when time is a crucial factor

March 2011

Uday Khedker, IIT Bombay

IITB

85/88

Research: Conclusions

To Be or Not To Be? That is the Question!

• Research is fun!

• Research makes a researcher a much better learner • Research enables better consolidation of skills ◮

Depth of skills

◮

The “skill” of applying various skills!

One gets involved with all aspects of solving a particular problem

Research takes us to a different state of mind!

March 2011

Uday Khedker, IIT Bombay

IITB

86/88

Research: Conclusions

The Ten Commandments of Creativity in Research 1. Seek extension of an earlier known solution 2. If you have a solution, find a problem 3. Find out the right questions to ask 4. Seek generality by removing specificities 5. Seek symmetry by testing for duality 6. Distill the essence, refine your ideas 7. Distinguish the necessary from the unnecessary 8. Distinguish the relevant from the irrelevant 9. Build levels of abstractions and migrate between them 10. Mix deep thinking with routine mechanical work

March 2011

Uday Khedker, IIT Bombay

IITB

87/88

Research: Conclusions

The Ten Commandments of Success in Research 1. Work on important problems 2. Work on multiple problems 3. For each problem, identify where you are on the S curve 4. Seek beauty in everything you do 5. Get emotionally involved 6. Remain committed 7. Work hard, work continuously 8. Consolidate your understanding 9. Donâ&#x20AC;&#x2122;t depend on luck 10. Handle uncertainty well

March 2011

Uday Khedker, IIT Bombay

IITB

87/88

Research: Conclusions

The Ten Commandments of Success in Research 1. Work on important problems 2. Work on multiple problems 3. For each problem, identify where you are on the S curve 4. Seek beauty in everything you do 5. Get emotionally involved 6. Remain committed 7. Work hard, work continuously

Also applicable to • A practising engineer • A student • A teacher

and almost all walks of life!

8. Consolidate your understanding 9. Don’t depend on luck 10. Handle uncertainty well

March 2011

Uday Khedker, IIT Bombay

IITB

88/88

Research: Conclusions

Last But Not the Least

Thank You!

March 2011

Uday Khedker, IIT Bombay