Undergraduate and Masters Research: Current JHU Students
I work with a small number of undergraduate and masters students on research each semester. I collect names of interested students and select candidates from this list at the start of each semester based on available projects.
If you are interested in a project in my lab and are a current JHU undergraduate or masters student, please complete this Google form:
This is a rolling process. I contact students only when a project becomes available; I don’t accept/reject submissions. If I don’t have a project for you then I will keep your name on file in case something comes up.
I’m glad you are considering JHU. I have no input into the masters student admission process, so I cannot assist you in your application. Additionally, I do not provide funding to accepted masters students. Do not contact me about masters admissions.
I do not generally take summer in interns. In previous years I have hosted interns as part of internship programs run by various groups at Johns Hopkins. Please do not contact me directly about summer internships. I am unlikely to reply.
I often have postdoc positions in my group. You should check to see if I have any projects listed here. I expect postdocs to have multiple publications in research areas relevant to my interests, e.g. papers at mainstream NLP and ML conferences, or other related conferences. I rarely consider applicants without this experience.
PhD Students: Prospective
Are you interested in joining my research group as a PhD student? Great!
I receive a large number of requests asking if my research group has openings, and if I can accept a PhD student. Emailing me directly will not help your application. If you have a specific question, I may be able to answer.
The Computer Science department in general, and CLSP and Malone in particular, accepts PhD students in Machine Learning, Natural Language Processing and Health every year. Whether my group specifically has a position or not, there are many other faculty with similar interests. In addition, I take students almost every year. There is no need to email me ahead of time asking if I have open positions.
The only way to become a PhD student is to apply. You can find more information about applying to the Computer Science department here: https://www.cs.jhu.edu/graduate-studies/phd-program/
Be sure to checkout our advice on how to apply, as well as common questions.
I applied to the PhD program. What’s next?
The admissions committee reviews all PhD applications and makes admission decisions. You do not need to contact me directly, and I won’t be able to provide any updates on the admissions process.
PhD Students: Common Questions
I’m glad you are considering joining my research group! This page provides answers to the most common questions I receive from prospective PhD students about my research group.
Why should I come to JHU?
Many students come because we are one of the leading research Universities in Natural Language Processing, engineering in healthcare and AI. Beyond that, I love JHU because you can have a meaningful influence on important problems.
The key to doing transformative work is collaboration. Johns Hopkins has interdisciplinary work in its foundation. Every university says “we encourage interdisciplinary work” but Johns Hopkins puts in place the policies and practices to support it.
CLSP and the Malone Center are two of many examples of interdisciplinary work. Consider Bloomberg Professors, faculty who must be jointly appointed between schools. It’s hard to appreciate the logistical hurdles this presents, but we do it because it’s important. The CS department currently has four Bloomberg professors.
I am particularly interested in health and medicine, an area where JHU is an international leader. In public health especially, Johns Hopkins is unmatched. I’ve oriented my research to benefit from this expertise. I work with the Institute for Clinical and Translational Research to develop clinical NLP that can be integrated into production environments. I work with numerous centers: Center for Population Health Information Technology, Institute for Global Tobacco Control, and Center for Gun Policy and Research. We have active projects in vaccination, HIV, tobacco, and infectious disease. I also have extensive collaborations with the JHU Applied Physics Lab, home to thousands or researchers.
These connections provide an environment at JHU that affords amazing opportunities for my research to have a meaningful effect on the world.
What is your advising style?
The most important aspect of a PhD program is your adviser. Your interests may drift, projects change, and the people who surround you will evolve. You adviser will be a constant, someone you will interact with weekly, and the person who will set you on your career path. Your relationship with your adviser is critical to your success.
I understand the importance of an adviser, so I take my job seriously. My job is to ensure that my advisees are, to the best of my ability, successful, well educated researchers.
Let me briefly describe how I run my group. Each one of my advisees has a weekly one hour meeting with me, usually one on one, with additional group project meetings depending on the project. I want to speak to each student at least once a week. If there is nothing to discuss we may cancel, but the meeting is on my calendar as weekly.
Group meetings occur every other week. We begin each meeting with a short summary of recent work from each student and plans for the next week. While this updates me on progress, the goal is to ensure research connections between students in the lab. After updates, one student presents their work in detail (usually for 45 minutes).
We have an active Slack channel where we discuss relevant topics all the time. Students use Slack to send me updates on experiments, share related work, and ask non-research related questions. We also use Slack to send weekly snippets; each student sends the group a short bullet list of their tasks for the past week, and goals for the next.
These channels are meant to ensure communication within the group.
How do you foster a cohesive research group?
We do a number of activities to strengthen group cohesion. In addition to regular group meetings, we usually do a group lunch once a semester. We also do a group outing; check out some of our previous group outings.
What is your goal in advising students?
My goal is to help you develop a research career over which you have ownership. I prefer not to hand students already formed projects, but instead let them develop their own interests and ideas. This doesn’t happen overnight. Most first year students need hands on guidance and direction, but that slowly lessens. In year 3 (approximately), I discuss with each student their longer term goals, both in terms of research (thesis) and career. We continue this conversation until graduation. I provide opportunities for the student tailored to career goals: teaching, mentorship of other students (graduate, undergraduate, highschool), research talks, internships, and grant writing. My goal is to help you find a job that you want, not one that I want (I already have a job!).
What will I work on if I become your PhD student?
Read recent publications from your prospective adviser. In my case, I publish on a lot of different topics, but these don’t necessarily form the central effort of my students. Instead, I recommend you look at the thesis topics of my PhD students (see the Alumni on the Group page). Your thesis will looks like these documents.
Everyone in my group works on machine learning methods and models for language data. Students differ in types of models and applications; half of my students work exclusively on core NLP topics, while the other half write NLP papers that consider health applications.
What type of job will I get after graduation?
When considering an adviser, look at the jobs his/her/their former students accepted after graduation. I list the first job of each alum here: Group
Many of my students pursue academic careers, with the majority pursuing industry research careers. Students decide their own career goals; my job is to help students achieve those goals. I help students understand their options, and then support them in achieving their goals.
How do you teach students about academia?
Many students consider academic careers, so I try to teach students about what academic jobs entail. As much as possible, I share my own activities with students, including during group meetings and weekly snippets. In particular, I make sure to share my (constant) research and career failures, including grant and paper rejections. I want students to have a realistic sense of the academic world, which includes both successes and failures.
How do I know if I am making progress towards graduation?
Unlike an undergraduate or masters degree, progress in a PhD program is not measured by counting credits. Therefore, establishing measures of progress is critical.
My goal is for students to graduate in five years, and I have a five year plan.
- Year 1: Coursework. Figure out the basics. Take part in a research project.
- Year 2: Finish coursework. Bring a research project to completion. Develop a clear sense of at least one research area.
- Year 3: Identify the general area in which you will write your thesis. Thesis proposal/Graduate board oral exam (GBO)
- Year 4: Select a specific dissertation topic.
- Year 5: Wrap up thesis work, job market, graduate.
Not everyone finishes in five years; many students stay in “Year 3” for a while. But this plan describes most student’s path through the program.
As a means to measure progress, each student sets concrete goals at the start of a semester by completing a spreadsheet listing every task and estimated level of effort. At the end of the semester, we review the list and fill in accomplishments and actual level of effort.
This process has a few goals. First, I want students to have a clear understanding of their priorities, without ambiguity as to what needs to be achieved. Second, I want students to have a realistic expectation of their work. Explicitly listing effort forces decisions about priorities and sets realistic goals. Third, goals enable measures of progress. We can concretely determine if the semester was productive in achieving goals or not. Finally, it creates opportunities to make improvements as unmet goals lead to discussions about improvements for the future.
What do you ask of students?
Work hard. A PhD is what you make of it. I work hard, and expect my students to work hard. This is your career; I expect you to take it seriously.
Balance. While hard work is critical, you must balance work responsibilities with other aspects of your life. The undergraduate experience can be all consuming but graduate school is better viewed as a job. Make sure you take time away from work to care for yourself, pursue hobbies, have a social life, etc. You’ll be more productive, happier and healthier in the long run.
Honesty. I rely on my students to be open and honest about their research, coursework and frustrations in the program. If you don’t tell me what is going on, I can’t help you. My role is a mixture of boss and adviser. You work for me and have responsibilities, but you also have your own research agenda and career goals, where my role is to advise you.
What do you do when students are struggling?
PhDs programs are challenging. Nearly every student faces periods of struggle in the program. During my PhD, I spent nearly a year working on several research ideas that all fell apart. Frustration and slow progress are natural parts of research. Since this happens to everyone, advisers develop strategies to help students through difficult periods.
The first step is for a student to identify that they are struggling. Difficult weeks happen to everyone, and this often comes up in weekly meetings. The most effective solution is to take a break, relieve stress, and take a fresh approach to the problem. It is vital that students have hobbies and interests to offer distractions and stress relief. Pushing harder won’t get you through a wall when you need a break and time to find another way around.
Occasionally, frustration lasts longer than a week or two, which could indicate a serious issue. We need to determine the cause: slow research progress? Lack of enjoyment in research topic? Unclear feedback from your adviser? Whatever the case, it’s important to discuss as soon as you are aware there is an issue.
While you know best when you are struggling, I keep watch as well. I read your snippets and listen during meetings to understand how you are doing. I sometimes recognize issues before students, and will start a discussion on solutions. Feedback and dialogue are key to identification and improvement. This is part of the reason we have yearly PhD student reviews and PhD progress letters.
It is also important for students to provide feedback to faculty. Students complete an annual self-review that includes adviser feedback. While it’s difficult to tell your adviser they need to do better, it’s sometimes necessary. I receive this feedback, and it’s usually accurate! I adjust accordingly.
We also have ways for students to submit feedback anonymously. CLSP has a student faculty liaison committee, where students can anonymously bring issues to the faculty. Students approach other faculty in confidence to share concerns; I have relayed student feedback anonymously to colleagues, and received it as well. It’s important to have a culture of openness, so we can all improve.
Do students attend conferences?
My expectation is that each student will attend at least one conference a year, regardless of papers published. Beyond that, students typically attend conferences where they have papers. All travel costs and fees associated with conference travel are paid by Johns Hopkins.
Do your students do industry internships?
Yes. Nearly every one of my students has done at least one industry internship. My students also occasionally work on a JSALT workshop team.
What are your expectations regarding student vacations?
There isn’t a fixed allotment of vacation days. I trust students to be reasonable about vacations. Vacations mean no work! No email! No running experiments! Weekends are NOT vacations, they are weekends. If you are home visiting family but working during the days, that’s not a vacation day. My general assumption is that students take two weeks of vacation in the winter, and two weeks in the summer. If they need more time, we have a discussion. Conference travel does NOT count as a vacation, but I often encourage students to take advantage of travel by adding on vacation days. When students aren’t traveling or on vacation, I generally expect they are on campus working (working from home occasionally is fine.)