September1999 October 1999 Research: Topics and Advisors Marie desJardins CMSC 601 February 19 and 26, 2009

September1999 October /19 and 2/26/09 2 Sources u Robert L. Peters, Getting What You Came For: The Smart Student’s Guide to Earning a Master’s or Ph.D. (Revised Edition). NY: Farrar, Straus, and Giroux, u Richard Hamming, “You and your research.” Transcription of the Bell Communications Research Colloquium Seminar, March 7, u Robert L. Hampel, “In search of new frontiers: How scholars generate ideas,” Chronicle of Higher Education 55(17): 72, December 19, 2008.

September1999 October /19 and 2/26/09 3 Outline u Advisors u Doing Research u Words from the Wise u Research Topics

September1999 October 1999 Advisors

September1999 October /19 and 2/26/09 5 Advisors u Temporary advisors u Research advisors u Approaching a potential advisor u Secondary/informal “advisors” u Changing advisors

September1999 October /19 and 2/26/09 6 Questions to Ask Faculty u Are you taking on new (M.S./Ph.D.) students? u Do you have RA funding? For how long into the future? u What research areas are you working in? u Do you have specific open problems you are looking for students to work on? u Do you generally suggest research topics to your students, or do you expect them to find their own topics? u Are you willing to advise a thesis/dissertation in an area not directly related to your current research projects? u Are you willing to advise an interdisciplinary thesis/dissertation, or to co-advise? u Have you (recently) graduated any (M.S./Ph.D.) students?

September1999 October /19 and 2/26/09 7 Questions to Ask Students u Is Prof. X accessible?  How much time does Prof. X spend with you?  In what contexts (individual meetings, lab meetings, etc.)? u Do Prof. X’s students finish quickly? u Do Prof. X’s students publish in top conferences and journals? u Does Prof. X give credit to students for their work? u Is Prof. X consistent in expectations and directions? u Is Prof. X reasonable in the amount of work expected? u Do students respect Prof. X intellectually? (From Peters, p )

September1999 October /19 and 2/26/09 8 Expectations u You can reasonably expect your advisor to:  Be available on a somewhat regular schedule  Suggest courses and schedules  Help you to select and solve research topics  Suggest committee members  Provide feedback on written work and work in progress  Suggest possible solutions to research problems  Encourage you to publish  Write letters of reference u Your advisor may also:  Provide financial support (stipends and travel money)  Provide career advice  Help you find a job

September1999 October /19 and 2/26/09 9 Expectations u Your advisor can reasonably expect you to:  Develop ideas independently  Do what you say you will do, in a reasonable timeframe  Make (reasonably) continuous progress  Go beyond the minimum amount of work  Be pro-active in pursuing ideas and looking for resources  Ask for help when you need it  Meet relevant deadlines, even if heroic short-term effort is required u Your advisor may also expect you to:  Provide written progress reports  Review papers (theirs and others’)  Work with other students in the lab  Publish  Contribute to grant proposals

September1999 October /19 and 2/26/09 10 Not-so-Great Expectations u Your advisor should not expect you to:  Perform excessive administrative tasks or paperwork  Contribute to research without authorship  Consistently work unreasonably long hours  Have no life outside of the lab u You should not expect your advisor to:  Constantly remind you what you need to be doing  Solve every problem you encounter  Be familiar with every aspect of your research problem  Provide unlimited resources (time, money, equipment...)

September1999 October /19 and 2/26/09 11 In the Unlikely Event... u What if your advisor is seriously abusing or neglecting you?  Talk to the GPD  Talk to another faculty member you trust  Change advisors  Talk to the department chair  Talk to the Associate Dean  File a formal complaint

September1999 October 1999 Doing Research

September1999 October /19 and 2/26/09 13 What Is Research?? u Asking “why” and “how” u Creating innovative solutions to novel problems u Also:  Understanding previous work  Testing hypotheses  Analyzing data  Publishing results u Not:  Applying existing techniques to a new problem  Developing a one-shot solution to a problem

September1999 October /19 and 2/26/09 14 A Good Topic u...is unsolved u...is important u...is interesting to you u...is interesting to your advisor u...is interesting to the research community u...has useful applications u...applies to more than one problem

September1999 October /19 and 2/26/09 15 Scope u Too broad is bad u Too narrow is bad u Too constrained is bad u Too unconstrained is bad u “Telescoping” is best

September1999 October /19 and 2/26/09 16 Getting Jumpstarted u Read! u Write  Annotated bibliographies  Literature surveys (including open challenges) u Replicate previous work  Re-implement  Re-derive  Re-experiment u Start varying parameters, assumptions, environments

September1999 October /19 and 2/26/09 17 Read, Read, Read! u You have to read a lot of research papers to become an expert u You have to become an expert before you can produce high-quality results u You have to produce high-quality results before you can complete your Ph.D. (or M.S.) u  you have to read a lot of research papers (and other people’s theses/dissertations) u  you might as well get started now!

September1999 October /19 and 2/26/09 18 Just Do It u “People have an amazing ability to become interested in almost anything once they are working on it.” (Peters, p. 181)

September1999 October /19 and 2/26/09 19 Write Early! u Write an annotated bibliography u Write a proposal outline u Write a literature survey u Write an outline of a conference paper u Write an outline of the dissertation u Show your writing to your advisor, other graduate students, colleagues,...

September1999 October /19 and 2/26/09 20 Articulating Your Topic u What is the question to be answered? u What is an approach you might try to get started? u What is the claim you’d like to make? u What is the evidence you could gather?

September1999 October /19 and 2/26/09 21 CS Units of Study u Each area chooses different units of study:  Algorithms: Algorithms (duh)  AI/Graphics: Methods, techniques, algorithms  Languages: Language components or features  Architecture: Instruction sets, memory hierarchies, architectures  Theory: Models, theorems, proof techniques  Systems: Systems, components, architectures Used with permission from Tom Dietterich’s CS 519 (Oregon State University) course slides

September1999 October /19 and 2/26/09 22 Unit Diversity u Size:  Small units: algorithms, language features, architectural components  Medium-size units: instruction sets, proof techniques  Large units: languages, architectures, machines u Potential variety:  Low variety: sorting algorithms, cache design  High variety: AI algorithms, languages, architectures u Cost of evaluation:  Low cost: algorithms, AI methods  Medium cost: theorems, components of architectures  High cost: languages, architectures, ideal models u Smaller units generally exhibit less variety and lower evaluation cost, so they are easier to do research on Used with permission from Tom Dietterich’s CS 519 (Oregon State University) course slides

September1999 October /19 and 2/26/09 23 Research Life Cycle u Definition. Exploratory research defines a new problem, constraints, opportunities, and/or approaches u Initial Solutions. Initial algorithms, designs, theorems, and/or programs are developed. u Evaluation of Initial Solutions. Initial solutions are evaluated and refined in isolation. u Comparison of Solutions. Solutions are compared to one another, to baselines, and to ideal solutions. u Space of Possible Solutions. Theorems are proved about the limits on any solutions. Existing solutions are placed in a common framework to determine whether all possible solutions have been found. u Technology Transfer. Best approaches are transferred to users and to new problems. Used with permission from Tom Dietterich’s CS 519 (Oregon State University) course slides

September1999 October /19 and 2/26/09 24 The Research Process u Research is not linear u Balance your time among  reading  writing  thinking  doing u and between  narrow focus  broad focus

September1999 October /19 and 2/26/09 25 Research Results u a definition of a problem or task u a unit for solving a problem or performing a task u identification of factors influencing the cost, effectiveness, or applicability of a unit (perhaps with some idea of the relative importance of factors) u development of an ideal model u a finished unit that can be distributed to users u measurement of some properties of a unit: run time, chip area, representation requirements, reliability, usability, etc. u identification of problems and shortcomings in a unit u a demonstration that one unit is better than another u a definition and demonstration of a tradeoff u analysis of a tradeoff showing how different points on the curve can be obtained and selected u a generative (explanatory) theory for some set of units Used with permission from Tom Dietterich’s CS 519 (Oregon State University) course slides

September1999 October /19 and 2/26/09 26 Research Methods u writing programs u writing systems u developing architectures u developing content architectures (ontologies, knowledge bases, class libraries, graphics toolboxes, etc.) u measuring properties of units u finding and proving theorems u analyzing and consolidating previous research u interviewing experts and customers u performing psychological experiments, surveys, observations Used with permission from Tom Dietterich’s CS 519 (Oregon State University) course slides

September1999 October /19 and 2/26/09 27 Research Methods (cont.) u building hardware u reading literature u importing techniques and results from other fields u measuring and predicting constraints on future units (e.g., VLSI technology, government regulation, user expectation and requirements) u writing papers, monographs, and textbooks Used with permission from Tom Dietterich’s CS 519 (Oregon State University) course slides

September1999 October /19 and 2/26/09 28 Research Project Phases u An individual research project (such as a Ph.D. dissertation) follows a lifecycle related to the research life cycle:  Choose research question/problem/tradeoff  Determine current state of knowledge  Apply appropriate methods to produce research results  Write up research results Used with permission from Tom Dietterich’s CS 519 (Oregon State University) course slides

September1999 October /19 and 2/26/09 29 Supporting Skills for Research u Programming and hardware design u Organization u Mathematics u Algorithm analysis and proof methods u Psychological techniques: Protocol analysis, experimental manipulations, survey methods u Statistics u Writing proposals u Writing papers u Critiquing papers u Designing experiments u Giving talks Used with permission from Tom Dietterich’s CS 519 (Oregon State University) course slides

September1999 October 1999 Words from the Wise

September1999 October /19 and 2/26/09 31 Peters’ Criteria for Research Topics u Financial support u Interest to you u Extendability after completion u Controversy u Time to complete u “Hotness” u Advisor’s enthusiasm u Closeness to advisor’s research u Depth of existing research u Duplication or uniqueness u Narrow focus u Tractability Peters p. 189

September1999 October /19 and 2/26/09 32 A Comment on Peters u “Once you have identified some topics you are interested in, you can research them rapidly by spending a few hours on the telephone calling up experts in the field and pumping them for information.”  Proceed with caution!!  Don’t do this without your advisor’s blessing  is better than telephone  Be specific

September1999 October /19 and 2/26/09 33 Hamming’s Advice u Work on important problems u Commit yourself emotionally to your work u Work hard u Tolerate uncertainty u Generalize u Don’t make excuses u Sell yourself and your work u Don’t fight the system u Be collegial u Look for the positive u Know your strengths and weaknesses

September1999 October /19 and 2/26/09 34 Research Advice from CRA-W u Start with problems, not with solutions  I have a hammer! Where’s the nearest nail?  Ideally, focus on a general problem or class of applications u Question assumptions ...of your work and previous work u Break your research into manageable pieces u Know how you will evaluate your method  Understand the standard methodologies for your field  Identify evaluation metrics  Develop baseline methods and benchmark problems u Have long-term and short-term goals u Sell yourself and your work

September1999 October /19 and 2/26/09 35 Good Research Practices from CRA-W u Be a good colleague  Help your advisor, other students, other faculty, colleagues  Collaborate! u Use other activities to benefit your research u Announce your accomplishments u Seek out supportive environments