Workshop | CRAG-MM: Comprehensive RAG Benchmark for Multi-modal, Multi-turn Challenge

The goal of this workshop is to engage in a detailed discussion on multimodal retrieval-augmented generation (RAG), with a special focus on the standout solutions from the KDD Cup 2025 on Comprehensive RAG for Multi-Modal, Multi-Turn Challenge. The workshop is an in-person event co-located with KDD 2025 in Toronto, Canada on Monday, August 04, 2025. See paper submission, registration, hotel, and venue details for attending.

The event will run from 08:00 to 12:00, with a coffee break (and poster session) scheduled from 09:30 to 10:00.

Location: Toronto Convention Centre, 712.

08:00–09:30	Session 1 [08:00-08:30] Opening remarks and competition summary [08:30-09:00] Team talk 1-2 (15 min each) Team Team_NVIDIA Team db3 [09:00-9:30] Keynote talk 1 Jianwei Yang: Toward General-Purpose Multimodal Agents: From GUIs to Robots and Beyond
09:30-10:00	Coffee break & poster session
10:00-11:55	Session 2 [10:00-10:35] Keynote talk 2 Jingrui He: Towards Multimodal Understanding on Rich Data IID vs. Non-IID [10:35–11:05] Team talk 3-4 (15 min each) Team AcroYAMALEX Team Dianping-Trust-Safety [11:05–11:25] Team spotlight talk (5 min each) Team otonadake Team NEC_AI_ROCKETS Team blackpearl Team cruise [11:25-11:55] Keynote talk 3 Dhruv Batra: Scouts: multi-modal agentic search for monitoring the web
11:55–12:00	Closing remarks (speaker: Luna Dong)