Research Projects

We study test-time methods for self‑improving and self‑adapting agents which represent an emerging frontier of artificial intelligence in which autonomous systems can not only perform tasks but also learn from interactions, refine their own behaviors, and even modify their own code or learning processes over time.

Translating complex scientific content into accessible language for social media platforms.

We study the feature flow across layers in state-space models compared to Transformers.

We incoporate external knowledge and goals for better CRS.

Cross-lingual consistency should be considered to assess cross-lingual transferability, maintain the factuality of the model knowledge across languages, and preserve the parity of language model performance. Inspired by the classic linguistic theory that the reference of a sentence is its truth value, we evaluate cross-lingual consistency for factual knowledge by substituting an entity with an equivalent one in other languages that shares the same reference.

Fact-checking real-world claims often requires complex, multi-step reasoning due to the absence of direct evidence to support or refute them. M^3Check guides the model’s reasoning process by constructing a structured, supervised, and verified chain of clues for verifying a claim, fostering a transparent, explainable, and user-friendly fact-checking process for multiple languages.

We generate high-quality code-mixed sentences using language models.

A framework integrating LLMs with external tools to enhance explainability and accuracy in table-based reasoning tasks.

A challenging dataset for compositional reasoning and claim verification on scientific tables.

We find positive use cases, but also weaknesses that need much handholding from human scientists.