Tag
This paper proposes a modality transfer task for Large Multimodal Models (LMMs) in GIS workflows and finds that current OpenAI LMMs struggle to transfer spatial information between image and text modalities, highlighting a critical bottleneck for autonomous GIS agents.
Presents ReGraph, a large-scale recipe graph dataset with explicit procedural structure and Recipe Reasoning Chain-of-Thought traces, plus a two-stage Recipe Graph Learning framework that enables LMMs to generate fine-grained cooking workflows from food images.
This paper introduces MM-CreativityBench, a benchmark for evaluating creative tool use in large multimodal models under physically constrained environments, and proposes affordance-grounded alignment using Direct Preference Optimization to reduce hallucination and improve grounded reasoning.