← Search

Vijay Kumar BG

2 accepted papers

2024

Self-Training Large Language Models for Improved Visual Program Synthesis With Visual Reinforcement

CVPR 2024poster

Visual program synthesis is a promising approach to exploit the reasoning abilities of large language models for compositional computer vision tasks. Previous work has used few-shot prompting with frozen LLMs to synthesize visual programs. Training an LLM to write better visual programs is an attrac…

Cited by 8SourcePDFScholar
2023

Q: How To Specialize Large Vision-Language Models to Data-Scarce VQA Tasks? A: Self-Train on Unlabeled Images!

CVPR 2023poster

Finetuning a large vision language model (VLM) on a target dataset after large scale pretraining is a dominant paradigm in visual question answering (VQA). Datasets for specialized tasks such as knowledge-based VQA or VQA in non natural-image domains are orders of magnitude smaller than those for ge…