Slides

NLKI Presented Poster

Presented at EMNLP 2025, Suzhou, China

Mixture of Experts Low Rank Decomposition

Lightweight Natural Language Knowledge Integration Framework for Improving Small VLMs in Commonsense VQA Tasks Poster.

Commonsense Reasoning Under Resource Constrained Environment

Presented at IIT-Kharagpur

AI Compression

Lightweight Knowledge-Enhanced Reasoning For “Small” Vision-Language Models

A Lightweight Natural Language Knowledge Integration Framework for Improving Small VLMs in Commonsense VQA Tasks

Presented at EMNLP 2025, Suzhou, China

Reasoning Vision Language Models

Small Models Lack Common Sense

MOE-fying VLM’s Following the path of LLMs with LORA

Presented at IIT-Kharagpur

Mixture of Experts Low Rank Decomposition

A parameter-efficient mixture-of-experts method that constructs multiple LoRA based experts and a frozen shared FFN block from a pre-trained dense model.