Abhishek Panigrahi

PhD student@Princeton University·美国·OpenReview

研究方向

Theory of Deep Learning. Statistical Machine Learning · Topic Models · RNNs · Contextualized models · Machine learning · deep learning · Statistical Machine learning

5.3

AdaptMI: Adaptive Skill-based In-context Math Instructions for Small Language Models

COLM 2025Poster

二作

6.8

Efficient stagewise pretraining via progressive subnetworks

ICLR 2025Poster

一作

4.9

Generalizing from SIMPLE to HARD Visual Reasoning: Can We Mitigate Modality Imbalance in VLMs?

合作者 (20)

Abhishek Panigrahi

STAT: Skill-Targeted Adaptive Training

In Good GRACES: Principled Teacher Selection for Knowledge Distillation

Progressive distillation induces an implicit curriculum

On the Power of Context-Enhanced Learning in LLMs

AdaptMI: Adaptive Skill-based In-context Math Instructions for Small Language Models

Efficient stagewise pretraining via progressive subnetworks

Generalizing from SIMPLE to HARD Visual Reasoning: Can We Mitigate Modality Imbalance in VLMs?