This document provides tips and strategies for effectively reading academic papers. It discusses deciding what papers to read based on relevance and credibility. It recommends making best use of academic resources like preprint sites, blogs, and mailing lists to stay updated. It explains the importance of reading for breadth to understand the big picture and reading for depth to critically examine assumptions, methods, statistics and conclusions. The document concludes by discussing how to take notes and think creatively after reading papers to develop new research ideas.
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
How to Read Academic Papers Effectively
1. How to Read Academic Papers?
Jia-Bin Huang
jbhuang0604@gmail.com
http://jbhuang0604.blogspot.com/
January, 2011
Taiwan
2. What this talk is about?
Efficient and effective paper reading
Useful websites, tools, tips you should know
A common sense talk
3. Paper reading and paper writing
"What’s the most resilient parasite?"
"An Idea."
"A single idea from the human mind can build cities. An idea
can transform the world and rewrite all the rules."
Papers communicate ideas
Reading :: Writing = Extraction :: Inception
5. Outline
1 Deciding What to Read
2 Make the Best Use of Academic Resources
3 Reading for Breadth and Reading for Depth
4 Summary, Review, and Creative Thinking
6. Outline
1 Deciding What to Read
2 Make the Best Use of Academic Resources
3 Reading for Breadth and Reading for Depth
4 Summary, Review, and Creative Thinking
7. Deciding what to read
Why?
Information explosion → too many papers
Only very few of them are helpful for your own research
8. Deciding what to read
How?
Evaluate papers by their credibility
Select papers by their relevance
10. Select papers by relevance
The key question
Why you want to read this paper? (What do you expect from
reading this paper?)
11. Select papers by relevance
Reasons to read papers
Get to know a new problem
Describe current research
Understand a well-known algorithm
Follow conventional experiment setup
Replicate/extend the results
Learn how to write
12. Get to know a new problem
Questions to ask
Why is this problem important/hard?
What is the problem setting? (input, output)
Search : keywords + tutorial/lecture/course/video/introduction/wiki
18. Replicate/extend the results
How?
Papers may provide useful data (or state-of-the-art performance)
and serve as building blocks in your research
Search : keywords + suvery/review/benchmark/qualitative/study
19. Learn how to write
Resources
The Science of Scientific Writing by George Gopen, Judith Swan
Notes on writing by Fredo Durand
Writing Research Papers by Aaron Hertzman
Advice on Research and Writing at CMU
How to Get Rejected by Fabrice Neyret
20. Outline
1 Deciding What to Read
2 Make the Best Use of Academic Resources
3 Reading for Breadth and Reading for Depth
4 Summary, Review, and Creative Thinking
21. Make the best use of academic resources
Why?
Well, I am a lazy graduate student...
22. Make the best use of academic resources
How?
Seek other forms of research product
Stay updated
23. Seek other forms of research product
Publication may not be the only product of a research work.
What else?
Abstract Supplemental material
Presentation Demo video
Author webpage Project page
Code Dataset
24. Stay updated
How?
Websites (research blog/preprint sites/author personal page)
Mailing list subscription
25. Blog: a new research platform
Examples
What’s new by Terence Tao, see also his google buzz
Godel’s Lost Letter and P=NP by Dick Lipton
Machine learning (Theory) by John Langford
Nuit Blanche - compressive sensing news
ScienceBlogs
26. Preprint: get access to the most up-to-date papers
Examples
arXiv.org: e-prints in physics, mathematics, computer science,
quantitative biology, quantitative finance and statistics
Sciweavers: academic bookmarking network
Resource for Computer Graphics/Vision by Ke-Sen Huang
Author personal page
27. Mailing list subscription
Examples
Imageworld: announce worldwide events and academic vacancies
MIT CSAIL Seminar
UC Berkely computer vision mailing list
28. Outline
1 Deciding What to Read
2 Make the Best Use of Academic Resources
3 Reading for Breadth and Reading for Depth
4 Summary, Review, and Creative Thinking
29. Reading for breadth
Build a framework
What did they do? (by skimming abstract, introduction, headings,
graphics, definitions, conclusions and bibliography)
Decide whether to go on
32. Reading for depth
Scientific skepticism
Examine the (implicit) assumptions
Examine the methods
Examine the statistics
Examine the conclusions
33. Examine the (implicit) assumptions
Questions to ask
Do their results rely on any assumptions about trends or
environments?
Are these assumptions reasonable?
34. Examine the methods
Questions to ask
Did they measure what they claim?
Can they explain what they observed?
Did they have adequate controls?
Were tests carried out in a standard way?
35. Examine the statistics
Questions to ask
Were appropriate statistical tests applied properly?
Did they do proper error analysis?
Are the results statistically significant?
36. Examine the conclusions
Questions to ask
Do the conclusions follow logically from the observations?
What other explanations are there for the observed effects?
What other conclusions or correlations are there in the data that
they did not point out?
37. Outline
1 Deciding What to Read
2 Make the Best Use of Academic Resources
3 Reading for Breadth and Reading for Depth
4 Summary, Review, and Creative Thinking
38. After reading the paper...
How?
React to what you read
Creative thinking
39. React to what you read
Taking Notes
Highlight major points
React to the points in the paper
Construct your own example
Summarize what you read
40. React to what you read
Analogy: Gram-Schmidt process
Papers :: Reading = Vectors :: Orthogonalization
Extract the “innovation" of the paper.
41. Creative thinking
Resources
How to come up with new research ideas by Jia-Bin Huang
How to invent? Raskar idea hexagon by Ramesh Raskar
42. Creative thinking
Five ways to come up with new ideas
Seek different dimension
Combine two or more topics
Re-think the research directions
Use powerful tools, find suitable problems
Add an appropriate adjective
43. One example - Content-aware image resizing
[Avidan and Shamir SIGGRAPH 2007]
Idea
Resize (reduce/expand) images while preserving the image
content.
The dimension: space
44. Video retargeting
[Shamir et al. SIGGRAPH 2008]
Idea
Extend dimensions from 2D image to 3D video: image resizing →
video resizing
The dimension: space
45. Nonchronological video synopsis and indexing
[Pritch et al. PAMI 2008]
Idea
Resizing (reduce) the temporal dimension.
The dimension: time
46. Data-driven enhancement of facial attractiveness
[Leyvand et al. SIGGRAPH 2008]
Idea
Reshape the face to enhance attractiveness
The dimensions: distances between facial feature points
47. Parametric reshaping of human bodies in images
[Zhou et al. SIGGRAPH 2010]
Idea
Reshape the human bodies in image
The dimensions: human shape
50. A Comparative Study of Image Retargeting
[Rubinstein et al. SIGGRAPH 2010]
Idea
Provide the comprehensive perceptual study and analysis of
image retargeting
51. PatchMatch
[Barnes et al. SIGGRAPH 2009]
Idea
Add constraint into the resizing process
Adjective: Constrained
52. Motion-aware video resizing
[Wang et al. SIGGRAPH 2010] [Wang et al. SIGGRAPH Asia 2010]
Idea
Exploit motion information for better video resizing quality
Adjective: Motion-aware
53. References
Paper reading
How to read a paper by S. Kesha
How to Read a Scientific Paper by John W. Little and Roy Parker
Efficient Reading of Papers in Science and Technology by
Michael J. Hanson
How to read a research paper by Michael Mitzenmacher
Tools
Publish or Perish (a program that analyzes academic citations)
Mendeley (Academic reference management software)
VideoLectures.NET (Free on-demand educational video lectures)
54. For more complete materials and explanations, please visit my blog
Redefining Open Mind: http://jbhuang0604.blogspot.com/