AIED
Simulating Validity: Modal Decoupling in MLLM Generated Feedback on Science Drawings
Wed Jul 1, 2:30 PM–2:45 PM · North 206
Part of Controllable and Safe AIED
Automated Feedback Generative AI & Large Language Models Science & STEM Inquiry Explainable & Trustworthy AI
★ Notable speakers
Xiaoming Zhai
★★
— AI and machine learning for science education assessment; automated scoring; GenAI in STEM
Investigates modal decoupling in multimodal LLM feedback on science drawings, probing how faithfully feedback reflects visual content.