Prospective Students
Thank you for your interest in working with me! This page outlines my research interests in more detail as well as my advising style when working with students. Hopefully, this gives you a better idea of what you can expect should we work together on a research project. If you have any further questions, please feel free to reach out with “[Prospective Student]” in the email header.
Future Research Directions
I am primarily interested in research topics within the field of natural language processing (NLP). The goal of my research is to bridge the gap between machines and humans in their abilities to process language. As a result, I am also interested in research questions in computational linguistics and computational social science that inform us about how humans use language. I am particularly interested in how knowledge about human language from related fields can translate into performance gains in practical NLP systems.
My prior research has focused on enhancing NLP systems in their abilities to process informal language such as slang. While doing so, I have also made contributions toward the understanding of how humans use slang. Going forward, I envision my research to focus on improving language AI in the following aspects:
Contextualizing AI: We as humans have the ability to process language with respect to the appropriate communicative context. For example, while it may be permissible, or even preferred, to use informal language such as “what’s up” between friends and families, such language use is likely inappropriate in formal settings (e.g., greeting a supervisor, addressing a judge in a hearing). In addition, humans not only have the ability to select contextually appropriate language, but we also have the ability to diversify our language, resulting in language variation. The goal of this line of research examines the plausibility for AI systems to 1) be able to use language flexibly so it conforms to the communicative context, and 2) generate and interpret context-specific use of language across a rich set of communicative scenarios.
Creative AI: Humans can use language in very creative ways. For example, making a joke, creating a meme, using metaphors, and writing poetry. The creativity in human language use does not restrict to the use of such creative expressions. They can also be embedded in the evolution of our languages. For instance, one may extend a sense (i.e., meaning) of a word using metaphorical extension (e.g., extending the ‘a wild animal’ sense of the word wolf to mean ‘a predatory person’). This line of work tries to understand how humans process creative language use and how we can make AI systems to also use language in creative ways.
Responsible AI: The aforementioned goals try to make AI agents more human-like. Inevitably, we are making the assumption that such advancements will be applied with goodwill. In a world where bad actors are prevalent (e.g., impersonation, deep-fakes), AI researchers have the responsibility to create systems that conform to human values. Here, the AI systems not only have to act appropriately with respect to the communicative context, but they also need to have an awareness of the consequences of their words. In other words, taking responsibility of one’s words just like humans do. This line of work investigates methods to define and evaluate responsible behavior and aims to develop precautionary measures toward potential AI misuse.
These research directions serve as a broad roadmap for my research program. Aside from these core interests, I am open to discussing potential projects that you may want to work on. In some cases, we may be able to figure out a brand new project together! If you’re interested in discussing potential research collaborations with me, feel free to reach out!
Advising Philosophy
Every advisor is a bit different and finding the right fit is usually the most meaningful step towards your academic goals. While describing one’s own advising style is surprisingly difficult due to subjective bias, I will try my best to offer such a description based on my core philosophy. Namely, being an advisor who is available, flexible, and responsible. These descriptions also serve as my testament on the kinds of relationships I want to establish with my students.
Availability: I believe that a good advisor needs to be available for their students in times of need. This can be as simple as a quick question on a technical concept or something more nuanced like a reference. I expect to meet with you at least once per week during the semesters. The meetings do not need to be long, but it should provide you with an opportunity to discuss any needs or concerns. If you prefer, I am also happy to sign up for a messaging app of your choice for more immediate feedback when it is feasible. In the past, some of my students find this very helpful as they can address technical questions immediately instead of waiting until the next day/formal meeting. That being said, this is certainly not an obligation nor expectation. My availability will likely be more constrained as I become more senior but I am committed to maintain a research group size so that every student can receive the attention they deserve.
Flexibility: What you are interested in may not perfectly align with what your advisor is interested in. It may be inevitable that some gap exists, but I would like to ensure that my students are working on projects that they are genuinely interested in. I am open to discussing your interest and work with you to find a project of joint interest, even if it does not contribute toward my immediate research agenda. If you are interested in a research topic that is tangential to my area of expertise but can convince me why I would be a good fit to advise the project, I am open to discussion as well. As I accrue more seniority, I would like to offer more academic freedom to my students to work on topics that they are the most passionate about.
Responsibility: I understand that students have their own academic goals they would like to achieve and such goals may change over the course of a project/degree. Before we begin working on a project, I would like to learn more about your goals so that I can set expectations in a responsible manner. For example, an exploratory project is valuable for someone to gain hands-on research experience but less so if your goal is to have a publication on your resume for a grad application in 6 months. Given the highly uncertain nature of research, I would like to set reasonable expectations so there are no surprises down the road. I would like to have a thorough discussion with my students at least twice per year to understand your goals and be able to best support you as an advisor.
Hopefully, my philosophy provides you a glimpse into my personality/style and what you can expect when working with me. I look forward to hearing from you and discuss potential avenues of collaboration!