How I Study AI - Learn AI Papers & Lectures the Easy Way

From Blind Spots to Gains: Diagnostic-Driven Iterative Training for Large Multimodal Models

Intermediate

Hongrui Jia, Chaoya Jiang et al.Feb 26arXiv

Large multimodal models (LMMs) can look at pictures and read text, but they still miss tricky cases, like tiny chart labels or multi-step math.

#Large Multimodal Models#Diagnostic-driven Progressive Evolution#Reinforcement Learning

DeepVision-103K: A Visually Diverse, Broad-Coverage, and Verifiable Mathematical Dataset for Multimodal Reasoning

Intermediate

Haoxiang Sun, Lizhen Xu et al.Feb 18arXiv

DeepVision-103K is a new 103,000-example picture-and-text math dataset designed to help AI think better using rewards that can be checked automatically.

#DeepVision-103K#multimodal reasoning#RLVR

Latent Implicit Visual Reasoning

Intermediate

Kelvin Li, Chuyi Shang et al.Dec 24arXiv

Large Multimodal Models (LMMs) are great at reading text and looking at pictures, but they usually do most of their thinking in words, which limits deep visual reasoning.

#Latent Implicit Visual Reasoning#latent tokens#bottleneck attention masking

Papers3

From Blind Spots to Gains: Diagnostic-Driven Iterative Training for Large Multimodal Models

DeepVision-103K: A Visually Diverse, Broad-Coverage, and Verifiable Mathematical Dataset for Multimodal Reasoning

Latent Implicit Visual Reasoning