For Current Members

Choosing Research Topics

A common problem among many students (in fact, most people) is that they cannot decide. Many people hesitate to choose because they are afraid of choosing something wrong and then regret later. Some studies suggest that people feel less happy when they have more options. In some retail experiments, customers leave without buying anything if too many items are on sale.

When students join research teams, they may face the options of choosing topics. Here are some common misunderstanding about choosing research topics:

  • Students decide research topics.

  • If a student chooses a topic, the student cannot change the topic for the next forty years.

The most important factor of deciding research topics is the needs of the research teams, not a student’s preference. A research team is often similar to a baseball team. If the team needs a left field now, you have to fill that position, or you are out of the team. The research topics are decided by professors according to multiple factors, for example, the existing expertise of the teams, the projection of the teams’ future needs, etc. It is rare that a professor allows a student to choose a topic without any boundary. Sometimes news reports major breakthroughs in science or technologies and says the people have been working on solving the same problems for decades. What is covered in news is rare; common situations are not covered in news. In other words, most researchers change topics over their careers.

Many students say, “I don’t know what I want to do. Thus, I am not going to choose.” This thinking is wrong in multiple ways:

  • First, you have already chosen a major.

  • Second, you have already chosen to study at Purdue University.

  • Third, you have already chosen to consider to join Dr. Lu’s research team.

  • Fourth, you have already chosen to read this page right now.

Everything has a cost. The highest cost is time because nothing can get time back. When you are reading this page right now, you have already been spending the most precious resource, time, of your life. You have already made the decision to spend this moment this way.

Many students think not choosing any topic gives them the freedom to choose something else later. This is also wrong for multiple reasons:

  • If you do not choose a research topic now, you may have no options later. If you want to attend a graduate school, research experience can be very helpful getting admissions. If you do not choose a topic, you may not receive the admission from the university you want to attend. As a result, you do not have the option of attending that university.

  • If you do not choose a research topic now, you may not stand out in job interviews. As a result, you do not have the option of joining the company you like.

  • Most important, if you do not choose a topic now, you may never know which topic you actually like. If you are an outsider, you do not know the good and bad things about a particular research topic. As a result, you have no information to make the best decision for you.

“How to choose a topic?”, you may ask. The best answer is “Jump in and do it.” The best way to know whether you like a research topic is to start doing it. Read research papers. Repeat the experiments. Derive the formulas. Talk to the people doing this research. Just do it. You will likely determine, within a few weeks, whether you really hate it.

If you do not do it, you will never know whether you like it or not.

Choice is a Package

Some people imagine that they could choose the best among different options. That is usually not possible. If you like to travel around the world, you have to tolerate flight delays, jetlags, etc. If you like to live in a city, you have to tolerate traffic. Every choice is a “package”. Everything in the package comes together. Each research team has specific culture and style of doing research. Some teams emphasize teamwork. Some other teams encourage students to work independently. Some professors usually write short papers (2-4 pages). Some professors usually write long papers (10 pages or longer). Some professors’ papers have many equations. Some professors’ papers have no equation. Some professors expect students to appear during daytime; some other professors do not care if students appear only between dinner and breakfast. Some research teams need to write software. Some research teams need field studies. When you join a research team, you embrace all these into your topic.

Understand Your Personality and Restrictions

Many students choose “hot” topics. They choose topics based on the popular terms in news. This has many problems. First, a hot topic means many people are working on it. When these students graduate, there may be over supply of talents and jobs may be hard to find. Second, a hot topic must be a hard topic. Easy things cannot be hot because everyone can do it and there is nothing to talk about. That means a hot topic takes a lot of effort to understand and to master.

Hot topics inevitably becomes cold.

When you choose a research topic, you need to know your personality. Do you like more conceptual problems, or you prefer something tangible? You also need to know the restrictions imposed on you. If you have two semesters before graduation, you should not intend to solve a problem at the scale of a doctoral thesis. In most cases, students do not have the knowledge or experience choosing the right scale of problems. They should discuss with professors.

Understand the Four Stages of Doing Research

It has been observed that most students go through four stages when they do research:

  • Excitement due to ignorance. Students are always excited when they encounter new research problems, They are ready to change the world by their endless energy and unlimited optimism.

  • Frustration and disappointment with knowledge. The first stage may last several days to several weeks. After reading research papers, the students lose excitement. They realize “Everything I want to do has been done.” The more they read, the more frustrated they become. They conclude that they are late about everything. Anything that is worth doing has already been published.

  • Experiments and failures. The second stage may last several weeks to many months. Most students give up during the second stage. Some students think deeply enough and move to the third stage. They identify things that have not been published. They think of ways to improve existing solutions. They try these improvements and compare these new methods with existing methods. The new methods do not work at all, or are worse than the existing methods.

  • Improvement and innovation. The third stage may last several months to forever. Reading more papers usually does not help getting out of the third stage. Staying in office or laboratory is often non productive. Improvements and innovations usually come from non-routine activities: attend a seminar, do exercise, watch a sci-fi movie, talk to strangers, listen to podcast, take a short vacation, attend a conference … Research has shown that people are more innovative when they break their daily routines.

How to get out of Stage 2 and Stage 3? There is no method that always works. Here are several suggestions that can help:

  • Talk to experts. If you find a paper that is truly inspiring, contact the first author and ask whether you can talk by vidoe call for 10 minutes. Of course, you have to read the paper very carefully and understand most of the context. Don’t ask trivial questions. Instead, ask questions about directions, such as “What direction would you suggest to take?” “What mistakes would you suggest to avoid?” Use you @purdue.edu email. DO NOT USE @gmail.com.

  • Pay attention to details. Maybe all existing studies use images taken indoors. If you use outdoor images, will anything change? Maybe a published study is conducted on a sunny day. If you do the study on a rainy day, will anything change?

  • Create a table comparing different methods. What are similar and different among the published methods?

  • Repeat published studies. You may get different results because you take slightly different approach and the details are not published. Discover the differences.

Read biographies of great inventors. Learn how they solve problems.

Please read the sample interview questions.

If you are unable to answer any questions, you are not ready to join this group yet. Please come back later. Joining a group before you are ready means you are going to waste time.

This is the grading guide for new members. One of the most important differences between “student thinking” and “real world” is that

Progress Reports

This is the template for progress reports.

Four Essential Elements in Progress Reports

  • What problem you are solving?

  • Why are you solving this problem? How is it relevant to the team’s purpose?

  • What have you done? What is the result? What is the evidence?

  • Why do you do it this way?

You need to provide details. Use figures, drawings, photos, equations, screenshots, source code … to explain.

Honesty, Integrity, and Trust

Honesty, integrity, and trust are the foundation of research. Never lie. Never fake data.

It is understandable that you encounter problems that are harder than expected. It is understandable that your other commitment may prevent you from making enough progress. It is understandable that unexpected things happen. Be honest. If you have not made progress, tell the truth. If something does not work, explain what happens.

Common Mistakes

The undergraduate program at Purdues does not require research. Hence, you should join a research team only if you want to solve research programs. You need to make progress in solving research problems.

One of the most difficult parts of being a beginning researcher is to understand how progress is measured. Let’s first review what students typically do in classrooms: attend lectures, take notes, submit homework assignments, answer exam questions. In many (probably most) cases, all students in the same classes have the same homework assignments and the same exam questions. This is the source of a lot of confusion for many students: they think everyone in the world is solving exactly the same problem described in the homework or the exams. The implication is that many students deeply believe that everyone in the world is solving exactly the same problem. Everyone knows the problem. Everyone has read the same textbooks. Everyone has seen the same exam questions. Also, professors are supposed to know the answers because professors write the homework assignments and the exam questions.

No organization will put 100 people sitting together solving the same problem and the answer is already known by the “instructor”.

In many cases, students can get good grades without speaking a word in class. In fact, some students think speaking (asking questions or answering questions) disrupt lectures and should be discouraged.

“Classroom environment is not real.” Please read it 10 times and understand the implications.

Many students never realize that classroom environment is not real. Many students would be completely surprised that other people solve different problems and that professors do not have answers ready for research problems.

If you are in a research team, you must become an expert in the problem you are solving and nobody else should know as much as you do. If someone else knows as much as you (or more than you), by definition, you are unnecessary and should solve a different problem.

Good Progress Presentations

  • Here is a list of suggestions about giving good presentations:

  • Know the audience.

  • Practice. Make sure all content can be explained clearly. The connections among different materials need to be logic and smooth.

  • Meaure time. A speech should uses approximatley 80% total time and leave 20% for QA. If the presentation is 15 minutes, speak for 12 minutes and leave 3 minutes for QA.

  • Start with a clear title and your name. The title should be the problem you are solving.

  • Explain details. You must be an expert in solving the specific problems. Provide details so that others can learn from you.

  • Use visual aids effectively. Use figures, drawings, photos, equations, screenshots, source code … to explain. DO NOT USE A LOT OF WORDS.

  • Remove all irrelevant decoration (anything that is irrelevant to the research problem or your contributions).

Poster Examples

example 1

Checklistof skills needed for new members.

Set up Development Environment

It is recommended that you create a Linux virtual machine for the development environment. A virtual machine allows you to experiment different settings (such as different versions of software packages) without affecting your real machine. You have several options for creating virtual machines. One of them isVirtualbox. It supports Windows, Mac, and Linux.

Learn git

A new member needs to learn many tools for communication and collaboration. One of the most important is git and the GitHub platform.

  1. Please create an account in github. Your account should include your first name and last name (no exception).

  2. Upload your photograph to github.

  3. Do not create any funny namethat is different from your real name. This is a large team and nobody has time connecting a funny account name with the real person.

See https://guides.github.com/.

Todo

George, add info on distributed workflows.

Please understand how to use branches and merge correctly. There are three types of branches

  1. Master branch. It is used to release software. It should be the most stable version.

  2. Development branch: It should contain everything in the master branch and additional features. It should be usually stable. This branch serves as the staging area for integration tests. This branch should not be too far ahead of the master branch. After a (or a few) feature is added and tested, this branch and the master branch should merge and the new feature (or features) should be released.

  3. Feature branches: These branches are created to adding new features. Each feature branch should have a short life-span: a branch is created for a feature, the feature is tested, and then the branch is merged into the development branch.

Please understand that the purposes of the branches are to stage changes into the master branch. Each branch should last only a few days. A common problem among students is that they do not merge quickly. As time passes, the differences among branches become greater and the chances of merge conflict increase. If a branch is not merged within two weeks, the branch may have too many conflicts and cannot be merged. As a result, the branch has to be abandoned and all efforts making that branch is lost.

Learn Python

You can find many tutorials online. This isan example. If you want practice problems, Consider to solvethese problems _using Python. If you want to understand objects, please watchmy lectures for ECE 30862(called ECE 462 earlier).

Learn OpenCV (for Image Team)

If you are in the image team, please learnOpenCV.

Teamwork

One of the most important difference between doing class homework and research is the need of “team thinking”. You are part of a team and your must contribute to the team. Many students make significant progress in their computers but they do not share what they have done with the team. As a result, whatever they have done is restricted to themselves.

What does it mean sharing work with the team? At the minimum, each member should document contributions

Document Your Work

You need to clearly document everything you want to do, you have done, and the results. One of the most common mistakes when students start doing research is that “student thinking”: as long as I have learned, I don’t need to document. This is wrong.

You are responsible explaining to the other group members that you are doing. Your document must provide enough details so thatother people can reproduce your work.

Leading a Research Team

I was writing a chapter of Beautiful Evidence on the subject of the sculptural pedestal, which led to my thinking about what’s up on the pedestal - the great leader. - Edward Tufte

One of the speacial parts of this research team is the comprehensive leadership development program. The experience of leading a research team can define your career for many years after you graduate from Purdue.

The CAM2 team has two types of leaders:

  • Project leader: A leader manages a specific project (for example, active learning, drone video, crowdsourcing …). A project leader needs to know the project’s goal (write a research paper, build software, create data …) and guide members toward the goal. A leader needs to know enough about different parts of the project but does not need to know all details of everyone’s work. A project leader is a technical leader.

  • CAM2 leader: The CAM2 team has multiple project and tne entire team needs a leader. The CAM2 leader needs to think about the entire team strategically, for example, how to recruit and screen new members and how to reorganize future projects. The CAM2 leader represents all project leaders and communicate with the advisers. It is typical that the CAM2 leader is also a project leader.

Benefits of Being a Leader

  • You get the opportunity to develop the skills that cannot be obtained in classrooms. In classrooms, most students are used to “mind their own business”. Their success depends only on their own abilities and efforts. A leader’s success depends on the team members. This is a completely different experience and requires new sets of skills.

  • Leading a research project or the entire research team gives you new prospects about how to interact with people (team members, other leaders, professors, external collaborators, sponsors …).

  • Leaders need to understand their members: their abilities, their long-term aspiration, their short-term constraints.

  • Leaders need to understand the projects’ and the team’s needs, such as paper deadlines.

  • A leader has to think about the project’s needs and plan ahead.

  • If the research project publishes a conference paper, the leader is the first choice to present the paper.

  • If you need recommendation letters from the advisers, the letters will be much, much, much stronger if you are an excellent leader.

How to Become a Leader

  • Talk to the advisers, often. The leaders are selected by the advisers. If you do not talk to the advisers, you will not be selected.

  • Talk to team members often and know their skills.

  • In most cases, new members are not leaders. A leader has to be in the project for at least one semester.

Letters of Recommendation

Helping students succeed is one of the missions of university professors. Dr. Lu writes recommendation letters for graduate schools, awards, jobs, etc. Before you ask Dr. Lu to write a letter for you, please ask yourself these question:

  • Does Dr. Lu know me well?

  • Do I have something special to be recommended?

  • Can Dr. Lu write a strong letter for me?

  • Does he have time writing a letter for me?

If you answer No to any question, STOP. Don’t waste your time. If you took his class but never talked to him, he knows nothing about you and has nothing to recommend. Your grade is already in the transcript. If the only thing Dr. Lu can write is about your grade, the letter does not help you. Thus, Dr. Lu will not write a letter for you.

This video explains how recommendation letters work for applications of graduate schools. Please watch.

Dr. Lu writes letters for a student only after talking to the student. If you want a letter from him, you have to talk to him. If you are not on Purdue campus, please schedule a video call.

Dr. Lu’s letters always focus on accomplishments. You need to provide evidence of your accomplishments. “I really enjoy your class” is not an accomplishment and Dr. Lu cannot write a letter because a student enjoys his class.

When you ask Dr. Lu for a letter, he will always ask you the following questions (because graduate schools ask them). Please bring your answers with evidence.

Your ability to speak and write. You can answer this question by giving an excellent technical presentation and a well-written technical document. Your ability to work in teams. You must give Dr. Lu at least three weeks to write your letter. When you ask him to write, you must give him a list of universities you want to apply. He will send letters to only these universities. You must send all applications within one week. You are busy. So is everyone.

If you need a letter, pleass fill this form.