Upload
satyajeet-raje
View
103
Download
0
Embed Size (px)
DESCRIPTION
This talk was given to students in the third year of their Computer Engineering degree program at the University of Pune, India. It was intended to instill ideas of basic research methodology and that research is not something other-worldly. The message was that they can perform "research" in what they do everyday. The talk was very interactive, encouraging students to clarify their misconceptions about research.
Citation preview
RESEARCH AND THE ART OF DOING WHAT YOU WANT SATYAJEET RAJE
• BE 2009 SKN
• Final year project “Automa;c Text Summariza;on” • Natural Language Processing, Intelligent Search
• MS 2012 The Ohio State University
• Thesis: Sema;c Search Engine for Resource Discovery • Domain: Biomedical and Healthcare
• Currently pursuing PhD • Research in Seman;c Web, Ontologies and their
applica;ons in Knowledge Management and Data Integra;on
INTRODUCTIONS
TODAY’S TALK • Is “Research” what it really seems to be?
• Can anybody do research? • Is it really necessary?
• How will it help • for my project • in my job • for higher studies
• What research can be done right now?
• Steps in research • Easy research prac;ces • Transla;on to project
RESEARCH?
• Helps ORGANIZE your work
• The best way to SHARE YOUR THOUGHTS AND IDEAS to the world
• A way to GET RECOGNIZED for your work
A NEW WAY TO LOOK AT RESEARCH? • Not restricted to ACADEMIC
circles
• Not all research needs to be GROUND BREAKING
• “Research” does NOT need to be tedious and cumbersome!
“Research” is a WAY OF DOING THINGS !
TYPES OF RESEARCH
COMPUTER SCIENCE AND ENGINEERING
COMPUTER SCIENCE AND ENGINEERING
Applied Research
Theore;cal Research
COMPUTER SCIENCE AND ENGINEERING • More of an applied science
• Making the connec;ons is important
• Every so\ware or hardware is meant to aid someone in a different domain
RESEARCH AND COMMON SENSE • Commonly used “big” words
• Hypothesis – The UNANSWERED!! research ques;on
• Research Methodology – Your research PLAN • Literature Review – Your READING on the subject • Cita;on / Reference – An ar;cle relevant to your work
THE RULES OF THE RESEARCH GAME
• KNOW your resources
• USE resources well • DO NOT GIVE UP the game
• BE FLEXIBLE about the outcome
• FAILED EXPERIMENT does not mean failed research
• EVALUATE your solu;on
THE RESEARCH GAME
Illustra;on: Ernest Harburg. Source: The Science Game, N. Agnew and S. Pyke, Pren;ce-‐Hall (1978)
WHAT CAN BE DONE RIGHT NOW BE Project • Choose something that YOU like
• Use it to test the waters • Experiment now rather than during the Job, MS, ME ..
Tips and Tricks
• USE WIKIPEDIA but properly
• READ RESEARCH ARTICLES related to your project • PARAPHRASE from what you have read
• Have a RESEARCH PLAN and more importantly an EVALUATION PLAN for your project
BE PROJECT • Stay Focused on what you want to do • A “plan” needs well defined reasons, methods and objec;ves
• Have not only a PLAN A but PLAN B and C as well • Be cau;ous of the companies and the projects that they offer
• An in-‐house project is as good if done properly • Take note of everything you do and your contribu;ons to the
project
• Know your ul;mate professional goal
BE PROJECT • Stay Focused on what you want to do • A “plan” needs well defined reasons, methods and objec;ves
• Have not only a PLAN A but PLAN B and C as well • Be cau;ous of the companies and the projects that they offer
• An in-‐house project is as good if done properly • Take note of everything you do and your contribu;ons to the
project
• Know your ul;mate professional goal
USING WIKIPEDIA • Great staring point for new research and researchers • Go one stage deeper
• DO NOT refer to Wikipedia or Blogs directly
• Compare Wikipedia to Big Bazaar
The most important part of a Wikipedia ar;cle
READING RESEARCH PAPERS • Where to find them
• Wikipedia references • Top conferences and journals for your domain (Microso\
Academic Search) • Google Scholar
• What to read in them
Style Language Content
• “How to read a research paper” -‐ Michael Mitzenmacher www.eecs.harvard.edu/~michaelm/postscripts/ReadPaper.pdf
• Make a note or summary of the paper and how it relates to your work
• A collec;on of these notes is your literature survey
PARAPHRASE AND NOT PLAGIARIZE • Plagiarism – copy pas;ng directly
• Professional and academic suicide
• Copy the thought paste your ideas • Learn to paraphrase “although the storage capaci;es of hard disks have increased massively over the years, access speeds – the rate at which data can be read from drives – have not kept up” [1] White, Tom (10 May 2012). Hadoop: The Defini<ve Guide. O'Reilly Media. p. 3. ISBN 978-‐1-‐4493-‐3877-‐0
Over the years there has been an exponen<al increase in the storage capacity of hard disks. However, according to [1] the access speeds for these disks have not been able to scale at the same rate.
HAVE A RESEARCH PLAN • Know the scope of your project • Define a realis;c ;meline for your own benefit
• Does not need to be the same as the one in progress report
• You must have an idea of how you are going to evaluate your work apart from a working demo / technology
User Interface / Human Computer
Interac;on
Survey based evalua;on
AI / Machine Learning / Task
based
Experimental / Task based evalua;on
Architectural / So\ware Engg / Technology
Comparison based
evalua;on
PLAN TO WRITE A PAPER • Not as hard and ;me consuming as it seems
• Most of the work is already done.. It is just needed to be expressed in the right way
• Concentrate on the idea (WHY?) and not the implementa;on (HOW?)
• Every BE project has the poten;al to be converted into a decent published paper
• Need not be at the best conference or journal • Worst that can happen is that it will get rejected
• Learn from the reviews
ANATOMY OF A GOOD PAPER • Clear and Concise • Every word counts • As short as possible • Good relevant references • Explana;on of what is done and planned to be done • Evalua;on method and results if possible
ANATOMY OF A GOOD PAPER • Introduc;on
• Brief overview of your idea or project. What sets you apart. • Literature Review
• Notes on what you have read • Methods
• Explain what is done • Discussion
• What you think about it and future work • Evalua;on
• How have you evaluated the work and its results • Conclusion
TO SUMMARIZE • Research is not difficult
• Cri;cal thinking is the most important outcome
• Ask the right ques;ons to the right source
• Choose your project well • Do consider wri;ng a paper right now
• ALWAYS DO WHAT YOU WANT
CONCLUSION Ques;ons? Project, Research Process, Mentality a\er BE, Studies in the US, Start-‐ups, …