Overview

A talk on Natural Language Processing (NLP), with a specific emphasis on creating learning environments to model multi-turn agent interactions for goal-oriented tasks. Also with emphasis on CoLa: Learning to Interactively Collaborate with Large Language Models (COLM 2025), proposes a novel learning procedure that simulates a "Guide" agent to assist a "Reasoner" LLM in achieving step-by-step goals. This framework trains a 3B parameter language model that outperforms GPT-4o and other competitive multi-agent frameworks across four distinct language tasks.

2026-01-16_Abhishek_Sharma_Multi-turn_LLM_Agent_Interaction.pdf

Objectives of Event

• To understand and generate human language by learning patterns from massive datasets
• To strengthen faculty research by promoting interdisciplinary collaboration between various fields.

Convener Details

• Prof. Ashok Kumar, Head, Center for AI in Medicine, Imaging & Forensics, Sharda University, ashok.kumar6@sharda.ac.in

Co-ordinators:

• Dr.Sunita Sharma, 9884429021, sunita.sharma@sharda.ac.in

Speaker Details:

• Abhishek Sharma Dept of Computer science Purdue University USA