Hiring PhD students

There is at least one PhD position available for AI/ML/NLP research in my lab at UCSB each year. Before you apply, please read the official graduate student admission policy on the UCSB CS homepage for general departmental requirements. I have compiled the following FAQ for students who are interested in working with me. The applications are handled by the department, so I will not be able to evaluate your chance of getting into the program. You may choose to indicate your interests of working with me in the online application and your personal statement. If you would like to contact me, please read the following answers, and fill in this form. Due to time constraints, I might not be able to reply your email.

Q: What's your research interest?
A: My research interests are in the broad areas of Artificial Intelligence. I am particular interested in Machine Learning and Natural Language Processing. More specifically, I am passionate about statistical relational learning, learning to reason, information extraction, multimodality, social media, and spoken language processing. Currently, I'm interested in deep learning methods for natural language processing, multimodal computing, and knowledge graph reasoning. You may want to read my research statement to get a better idea (though it might be slightly dated).

Q: What are the basic requirements for working with you?
A: I have three basic requirements for my PhD students. First of all, students need to be Mathematically mature (i.e., proficient and comfortable with probability theory, statistics, linear algebra, optimization, and calculus). One way to test yourself is to carefully go through some of the best ML papers published at recent ICML/NIPS/AAAI/IJCAI conferences, and see if you can fully understand the detailed Mathematical presentations from the main equations in those papers. Secondly, I require my students to be proficient in programming: they should at least know one or more serious programming languages for relatively large-scale software engineering purposes (e.g., C, C++, Python, Java etc), and meanwhile, they have to master one of the command-line scripting languages such as Perl, AWK, Bash, Sed etc. The ability of prototyping ideas quickly in Matlab, Octave, R is often a big plus. Third, and probably most important: hard-working. No matter how smart you are and what fields you are in, if you don't spend enough time on your work, there is no way you can be an expert. An observation from the most successful PhD students is that they all work extremely hard. So, if you are a hard-working person who is passionate about advancing AI, and you are solid in Math and programming, please consider applying to UCSB.

Q: What makes me stand out from the application pool?
A: As you may know, applications to UCSB's graduate program is very competitive. To make you stand out, prior research experience is often a big plus: if you have prior publications at reputable locations, we know that you are on the right track. If you have published papers at leading conferences or journal, feel free to leave me a note and send me a copy. Alternatively, if you are a system builder / hacker, showing your contributions to major ML/NLP/AI projects on Github is also encouraged. Additionally, if you have very strong support letters from researchers that I know, it helps me to evaluate your potential. For undergraduate applicants without research experience, GPA and grades from standardized tests are often important.

Q: Why UCSB and UCSB CS?
A: UC Santa Barbara is one of America's Public Ivy universities and a member of the prestigious AAU. According to a recent report by PayScale, UCSB Computer Science ranked 1st in the 2015-16 college salary report. PhD graduates who chose to join academia have taken tenure-track positions at top schools such as UChicago, UIUC, Boston University, UT Austin, University College London, Dartmouth, Purdue, Virginia Tech, OSU, Northeastern, and Rochester. According to National Research Council, UCSB CS ranked in the top-3 to top-16 range nationwide (S-Ranking). Note that the CS ranking by US News is solely based on subjective human ratings, whereas the NRC ratings are based on data-driven statistics (as Ben Zhao pointed out on Quora, no rankings are perfect.). UMass has recently published an objective CS rankings based on top conference publications. It ranked UC Santa Barbara at the top-3 position for smaller departments that have <= 35 faculty members. The department has many leading researchers in the broad area of data science and systems: Xifeng Yan, Linda Petzold, Matthew Turk, Amr El Abbadi, Ambuj Singh, Tobias Höllerr, Chandra Krintz, Rich Wolski, Divy Agrawal... just to name a few. Location wise, Santa Barbara has the most beautiful campus in the world (see a video of the UCSB campus) and perfect weather year round.

Q: How long does it take to do a PhD?
A: It depends. I've seen some very productive PhD students finishing in only 4 years. So the assumption is, if you are productive and you have created enough impacts, it is definitely possible to finish a PhD in a relatively short amount of time.

Q: Do you have visiting student/scholar or intern positions?
A: I typically do not accept unsolicited applications for visiting students, interns, or scholars. The only exception is that you can demonstrate your expertise in my research area (typically via strong publications) and have sufficient funding to support yourself for a year, you may drop me a line.

Q: What computational resources are available?
A: We have recently built several deep learning machines with multiple NVIDIA Titan X Pascal GPUs in our lab, and we are in the process of acquiring more computational power as the lab grows. The CS department has shared multi-node computing infrastructure as well.

Last Updated: 03/16/2017.