Peer-reviewer for VLM4RWD (NeurIPS 2026)
Reviewer for Grounded and Faithful Vision-Language Models for Real-World Deployment (VLM4RWD) at NeurIPS 2026 (Sydney), reviewing three submitted papers.
Reviewing for Grounded and Faithful Vision-Language Models for Real-World Deployment (VLM4RWD), an official workshop at the 2026 Conference on Neural Information Processing Systems (NeurIPS 2026), held on 11 December 2026 in Sydney, Australia. NeurIPS, the parent conference, is rated A*, the top tier, by the CORE conference ranking.
The workshop is organised around three themes: grounded understanding and faithful reasoning, covering visual grounding, evidence alignment, and hallucination mitigation; reliable interaction and decision-making, covering embodied systems and causal decision-making; and evaluation and deployment reliability, covering benchmarks, uncertainty quantification, and robustness under distribution shift.
I am reviewing three submitted papers, assessing their methods, evidence, and contribution. The workshop's questions about grounding, evidence, and faithfulness are the same ones I write about in the RAG series, applied to vision and language rather than text retrieval alone.
