From the 1 of 5 linked papers with an AI index.
5 papers
Privileged, but Biased: How PI-Conditioned Teachers Break Self-Distillation
Sarthak Harne, Chinmay Karkar, Yash Pandya +2
Self-distillation (SD) has emerged as a compute-efficient alternative to reinforcement learning with verifiable rewards: a self-teacher, conditioned on privileged information (PI)…
Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at Scale
Yash Pandya, Sahil Gupta, Sarthak Harne +10
Echoverse introduces a pipeline that compiles specifications into deep, stateful synthetic applications for training computer-use agents, using a co‑evolution loop that repairs env…
Dialogue to Discovery: Attribute-Aware Preference Elicitation for Conversational Product Search Assistants
Sarthak Harne, Natwar Modani, Debabrata Mahapatra +1
Conversational product search assistants offer a more expressive, natural, and interactive alternative to traditional keyword-based product search. With limited screen space, showi…
LLM driven Text-to-Table Generation through Sub-Tasks Guidance and Iterative Refinement
Rajmohan C, Sarthak Harne, Arvind Agarwal
Transforming unstructured text into structured data is a complex task, requiring semantic understanding, reasoning, and structural comprehension. While Large Language Models (LLMs)…
CASE: Efficient Curricular Data Pre-training for Building Assistive Psychology Expert Models
Sarthak Harne, Monjoy Narayan Choudhury, Madhav Rao +5
The limited availability of psychologists necessitates efficient identification of individuals requiring urgent mental healthcare. This study explores the use of Natural Language P…