Publications and Pre-prints

Reasoning over very long inputs remains difficult for LLMs. We introduce COSMIR, a chain-style framework that replaces ad hoc messages with a structured memory. A Planner agent turns a query into sub-questions, and worker agents process chunks via a fixed micro-cycle (Extract, Infer, Refine). This yields higher faithfulness and better long-range aggregation on datasets like HELMET.

We introduce a novel copilot for automating the generation of P&IDs from natural language descriptions. Leveraging a multi-step agentic workflow, our copilot provides a structured and iterative approach to diagram creation directly from Natural Language prompts.

We introduce a novel framework for evaluating epistemic fairness in preference learning models inspired by economic theories of inequality and Rawlsian justice. We propose metrics adapted from the Gini Coefficient and Atkinson Index to quantify fairness in these models.

We propose a methodology involving a co-design model that actively engages stakeholders, integrating principles of Equity, Diversity, and Inclusion (EDI). We apply this to develop a dataset and AI model for evaluating public space quality using street view images, demonstrating effectiveness in capturing diverse perspectives.

Over the past decade, analogies have played a significant role as an intrinsic measure of evaluating word embeddings. We present ANALOGICAL, a new benchmark to intrinsically evaluate LLMs across a taxonomy of analogies of long text. Using thirteen datasets, we evaluate the abilities of eight LLMs in identifying analogical pairs in the semantic vector space.