In Feb 2010, I was invited to give a talk at my alma mater Punjab Engineering College to the ECE and CS department students. The goal was to get them to think a bit more about pursuing a PhD.
1. Tales & Lessons from a Nomadic Academic Life NitinMadnani Research Associate & Ph.D. Student Institute for Advanced Computer Studies University of Maryland, College Park (UMCP) http://www.umiacs.umd.edu/~nmadnani/
2. Why am I giving this talk? I made many mistakes during my academic journey [as you will see] I want you to learn the same lessons I learned from my mistakes without making them again I like the sound of my own voice
3. What is this talk about? My nomadic academic life Where I am How I (finally) got there Your (hopefully non-nomadic) academic life You should think about a PhD You SHOULD THINK about a PhD! You thought about it. Now what? Things to do BEFORE applying for a PhD Things NOT TO DO when applying for a PhD TALES LESSONS
4. Where I Am PhD [expected May 2010] Top 10 CS program @ UMCP 4.0/4.0 GPA Graduate Research Assistant for entire duration [$23000/yr] Publications 15 papers in peer-reviewed conferences and workshops 5 papers in refereed journals Working on a book chapter Interned with IBM Research, New York Adjunct lecturer in the Department of Computer Science and College of Information Studies @ UMCP. Taught/teaching courses on Statistical Natural Language Processing and Cloud Computing to first year graduate students Collaborating with researchers in Germany, India, Scotland and China Primary/Core Developer for 4 open source projects including the very popular Natural Language Toolkit (http://www.nltk.org)
5. How I Got Here Returned to India for personal reasons (Oct 2000) Graduated from PEC (2000) Joined UMCP as MSEE student (Aug 2000) Recd. GRA offer from Computer Engg. @ UMCP(May 2001) Wrote MBA exams in India (Nov ‘00 – Feb’01) Cleared NITIE & XLRI (Apr 2001) Re-joined UMCP as MSCE student and GRA (Aug 2001) Took elective “Intro to Computational Linguistics” (Jan 2003) Offered Internship with Computational Linguistics Lab (Jun-Dec 2003) Offered a fully-funded PhD position in Department of Computer Science by Prof. Bonnie Dorr (my current advisor) (Jan 2004) Graduated with MS in Computer Engg. (Dec 2003) Got my first CS journal paper published as a co-author (Dec 2003)
6. How I Got Here Returned to India for personal reasons (Oct 2000) Graduated from PEC (2000) Joined UMCP as MSEE student (Aug 2000) I think the term ‘Nomadic’ describes this picture pretty well! Recd. GRA offer from Computer Engg. @ UMCP(May 2001) Wrote MBA exams in India (Nov ‘00 – Feb’01) Cleared NITIE & XLRI (Apr 2001) Re-joined UMCP as MSCE student and GRA (Aug 2001) Took elective “Intro to Computational Linguistics” (Jan 2003) Offered Internship with Computational Linguistics Lab (Jun-Dec 2003) Offered a fully-funded PhD position in Department of Computer Science by Computational Linguistics Professor (my current advisor) (Jan. 2004) Graduated with MS in Computer Engg. (Dec. 2003) Got my first CS journal paper published as a co-author (Dec. 2003)
9. Why do a PhD? You contribute something absolutely new to the body of world knowledge You get to shape the next generation of engineers and scientists You will actually understand the scientific process and can use it to evaluate the scientific merit of any argument in any discipline You can affect someone’s life in a positive way as a teacher and mentor Your family and friends will give you the respect you deserve People will regularly google your name ‘Dr. NitinMadnani’ instead of ‘Mr. NitinMadnani’ everywhere You can be really, really famous if you do good research You get to give cool talks like this one The politically correct reasons The reasons you really care about
11. Why NOT do a PhD? Reason #1: Time Reason #2: Employment Reason #3: Salaries Reason #4: Research
12. Why NOT do a PhD? Reason #1: TIME Source: National Science Foundation Survey of Earned Doctorates for 2003 http://www.nsf.gov/statistics/infbrief/nsf06312/
13. Why NOT do a PhD? Reason #1: TIME Median CS PhD awardee age is 31.4 and Engg. PhD awardee age is 30.5 Source: National Science Foundation Survey of Earned Doctorates for 2008 http://www.nsf.gov/statistics/nsf10309/
14. Why NOT do a PhD? Reason #2: Employment Source: PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=1056
15. Why NOT do a PhD? Reason #2: Employment Source: PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=679
16. Why NOT do a PhD? Reason #3: Salaries Source: National Science Foundation Survey of Earned Doctorates for 2008 http://www.nsf.gov/statistics/nsf10309/
17. Why NOT do a PhD? Reason #3: Salaries +6 years +6 years +9 years (also starting salary cap for most industry jobs) +7 years +8 years Sources: College and University Professional Association for Human Resources Survey, American Association of University Professors Survey, Chronicle of Higher Education Surveyof Graduate Assistants
18. Why NOT do a PhD? Reason #3: Salaries Do an MBA if you just want to earn money! +6 years +9 years +7 years +8 years PhD = Comfort [+Contentment + Happiness] Sources: College and University Professional Association for Human Resources Survey, American Association of University Professors Survey, Chronicle of Higher Education Surveyof Graduate Assistants
19. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
20. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
21. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
22. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
23. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
24. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
25. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
26. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
27. Why NOT do a PhD? Reason #4: Research while (funding == true) { thesis_topic = find_original_idea(); GOOD_ENOUGH = query(advisor); if (thesis_topic > GOOD_ENOUGH) { graduate(); break; } else { years_in_gradschool += 1; continue; } } “THE BIG UNKNOWN” Adapted from PhD Comics, http://www.phdcomics.com/comics/archive.php?comicid=626
28. On a more serious note … The point of the last few slides was to present the pros and cons of choosing to do a PhD They should inspire thought, not fear Consider all aspects before you decide whether to embark on a PhD All of the reasons not to do a PhD can be overcome (Exhibit A: ME!) Conclusion: A PhD from the US is one of the best learning and rewarding experiences in the world if you make an informed choice to pursue it!
30. Prepare Yourself Develop a research-driven mindset Understand the problem/topic Explore existing solutions/trends Use all available resources (esp. the library) Knowledge also lives beyond the textbook and the class notes “A well-understood wrong answer is probably more useful than a misunderstood right answer”
31. Prepare Yourself Develop technical paper comprehension/writing skills A well presented argument is inherently more convincing Develop a habit of reading published journal/conference/workshop papers esp. in the 3rd and 4th years [Use: Google Scholar/Arxiv/CiteSeer] Example: when learning about “superscalar processors” in CS 308, it will be instructive to read (a) the 1967 paper† on Tomasulo’s algorithm (a strong precursor of superscalar) and (b) Smith and Sohi’s 1995 paper‡ on superscalar microarchitecture [both freely available] Start with papers on well-established research ideas OK if you don’t understand everything. Important to understand how scientific arguments are constructed and presented. †“The Microarchitecture of Superscalar Processors”. James E Smith and GurinderSohi. Proceedings of the IEEE. (1995) ‡“An Efficient Algorithm for Exploiting Multiple Arithmetic Units”. R. M. Tomasulo. IBM Journal of Research and Development. (1967).
32. Prepare Yourself Develop technical paper comprehension/writing skills Learn LaTeX†. DO NOT use Microsoft Word for technical writing unless absolutely necessary! Try to submit any original research to good peer-reviewed workshops and conferences It’s okay to get rejected. Learn from the reviews and improve both your writing and the research idea Treat class assignment write-ups as example workshop papers, given the professor’s permission †http://www.latex-project.org/
33. Prepare Yourself Learn to value originality and intellectual property Original ideas are the cornerstones of scientific progress and must be respected! Make a conscious effort to look for prior work similar to yours You MUST learn to cite and attribute everything that you borrow from other sources [See my slides] Wheels should not be reinvented just for an assignment or a project
34. Prepare Yourself My strong suggestion: a less glamorous but much better long-term option! Consider doing a Masters degree from India before the PhD [What I should have done] Should go a long way towards inculcating the three points I described Will drastically increase the odds of getting admitted and---more importantly---getting funded Only institutions with well-known, research-oriented Masters. Examples: IITD, IITK, IITB, IIIT-Hyd, IIIT-Blore & IISc
35. Prepare Your Application GRE/Academics You guys are already smart enough to handleboth of these aspects! For some of my thoughts and suggestions ontaking/preparing for the GRE, please see: http://www.umiacs.umd.edu/~nmadnani/grefaq.pdf
36. Prepare Your Application Choose your PhD area wisely I thought I liked Control Systems enough to do research in it. I was wrong. Make sure you really like the field you think you like I have already told you two ways to do this: Read papers from your chosen area to get a taste of what research in that area entails A pre-PhD Masters from India is a good (and cheap) way to validate your choice
37. Prepare Your Application Put lots of thought into the Statement of Research/Statement of Purpose (SoP) I polled 10 CS professors in 3 universities on SoPs they receive from India (for both MS and PhD) Reactions almost universally negative Reads like bad fiction (“I knew I was destined to become an Electrical Engineer ever since I touched a live wire and lived …”) It’s not supposed to be your life story! Why research? Why this area? Why this university? See my point about a pre-PhD Masters
38. Prepare Your Application Build a web presence Create and maintain an academic web page No flash! No weird music! No animation! Feel free to steal the design of my webpage Always have the latest PDF version of your resume linked to from that page Use official web space; otherwise Google Sites Put up all published papers
39. Prepare Your Application Leverage community resources Participate in community forums dedicated to higher education in the US Best example: Edulix.com. Well organized and very popular forums with lots of good advice Find other avenues on the web Share the knowledge!
40. Prepare Your Application Ask for help Do not hesitate to ask PEC alumni [incl. me] for advice and suggestions You can also write to existing students at a particular university It’s okay to write to professors to express interest in doing research with them but look at their webpage first. They may ask for specific format or to be left alone! Phrase the email properly. Include all relevant details depending on the recipient. Do not expect replies from everyone … esp. professors! Do not spam anyone ... esp. professors!
41. Summary This talk is meant to be both informative and transformative GOAL: Think hard before you apply for graduate school! Feel free to contact me†. I try to help as many people as I can esp. Pecobians. Consider getting a Masters degree from India if there are some good options available A PhD is not for everyone but it is for anyone! †nmadnani@gmail.com
42. Thank you!Questions?Happy to chat individually with students as well! Slides from this talk are available at: http://www.umiacs.umd.edu/~nmadnani/pectalk.pdf
Editor's Notes
I apologize in advance that this might sound like blowing my own horn but this is the “happy ending” part of the talk … that you can end up successful despite making the mistakes I did and so I wanted to share that right at the beginning of the talk