Shyamgopal Karthik

Researcher@Genmo·美国·OpenReview

研究方向

Text-to-Image Generation · Generative Models · RLHF · Vision-Language Models · Compositional Generalization · Computer Vision · Machine Learning · Optimization Methods · Multi-Object Tracking · Visual Object Tracking · Re-Identification

6.0

合作者 (20)

Shyamgopal Karthik

Post-hoc Probabilistic Vision-Language Models

Noise Hypernetworks: Amortizing Test-Time Compute in Diffusion Models

Concept-Guided Interpretability via Neural Chunking

Sparse Autoencoders Learn Monosemantic Features in Vision-Language Models

Scalable Ranked Preference Optimization for Text-to-Image Generation