Let's make both models and humans think critically.

I am a third year PhD candidate working in natural language processing at University of Illinois at Urbana-Champaign, advised by Prof. Jiawei Han. I am supported by the NSF Graduate Research Fellowship. Prior to this, I was working on 3D scene representations with Prof. Ren Ng at the University of California, Berkeley, funded by the Intel SRC Fellowship.

My research aims to integrate structured knowledge and reasoning in order to make both models and their human users critically think.

My work explores this in both the educational (LLMs + Education) and scientific (AI for Research) domains. Specifically, how do we structure both the interactions between models and users (e.g., an Instructor model guiding a student through a problem), as well as the data critical to guiding such interactions (e.g., a student's knowledge state or external knowledge required to solve the problem)?

Email  /  CV  /  Scholar  /  Twitter  /  Github  /  Linkedin

Highlights

• I am attending EMNLP to present TreeInstruct. Happy to chat, feel free to reach out!
• I will be presenting my poster at the NeurIPS'24 workshop for Large Foundation Models for Educational Assessment.
• TreeInstruct is accepted to EMNLP'24 Findings!
• I received the 2024 NSF Graduate Research Fellowship.

Research

I'm currently interested in exploring how we can model the dynamics between different agents and/or users in different settings in a structured manner (e.g., students and LLM-TAs; researcher-personas debating their works). I am also interested in devising new ways to mine and structure textual data, so that we can extract interesting insights from them (e.g., identifying a set of hierarchical topics commonly considered for developing a nuanced scientific perspective).

An Instructor agent guided by a novel state space-based planning algorithm. TreeInstruct asks probing questions to help students independently identify and resolve errors. It estimates a student's conceptual and syntactical knowledge to dynamically construct a question tree based on their responses and current knowledge state, effectively addressing both independent and dependent mistakes concurrently in a multi-turn interaction setting.

Introduces a novel task, episode detection, aimed at identifying episodes from a news corpus containing key event articles. EpiMine is an unsupervised episode detection framework that (1) automatically identifies the most salient, key-event-relevant terms and segments, (2) determines candidate episodes in an article based on natural episodic partitions estimated through shifts in discriminative term combinations, and (3) refines and forms final episode clusters using large language model-based reasoning on the candidate episodes. We construct three diverse, real-world event datasets annotated at the episode level.

An extremely weakly-supervised text classification method that leverages Mutually-Enhancing Text Granularities. MEGClass utilizes coarse- and fine-grained context signals obtained by jointly considering a document's most class-indicative words and sentences. This approach enables the learning of a contextualized document representation that captures the most discriminative class indicators.

A system which interacts with raw scientific literature, delivering precise and more informative chemical reactions. Going beyond mere extraction, Reaction Miner integrates a holistic workflow: it accepts PDF files as input, bypassing the need for pre-processing and bolstering user accessibility. Subsequently, a text segmentation module ensures that the refined text encapsulates complete chemical reactions, augmenting the accuracy of extraction. Moreover, Reaction Miner broadens the scope of existing pre-defined reaction roles, including vital attributes previously neglected, thereby offering a more comprehensive depiction of chemical reactions. Evaluations conducted by chemistry domain users highlight the efficacy of each module in our system, demonstrating Reaction Miner as a powerful tool in this field.