Johan W. Verjans

Australian Institute for Machine Learning, Adelaide University

Papers on Lattice

Total citations

Topics

Research focus

Multimodal Models (2)Computer Vision (1)Constitutional AI & AI Ethics (1)Eval Frameworks & Benchmarks (1)Interpretability & Mechanistic Interp (1)

Frequent co-authors

Vu Minh Hieu Phan (2)Ta Duc Huy (1)Trang Nguyen (1)Townim Chowdhury (1)

Papers (2)

Jun 30, 2026

Australian Institute for MachineJun 30, 2026·also Adelaide University, Corresponding author, Flinders University

Visual Semantic Entropy: Do Vision Language Models Recognize Visual Ambiguity?

Overconfident visual embeddings can mask true uncertainty, but Visual Semantic Entropy reveals the hidden complexity of visual ambiguity in vision-language models.

Ta Duc Huy, Trang Nguyen, Townim Chowdhury +5

Computer Vision Constitutional AI & AI Ethics Multimodal Models

Apr 6, 2026

Hanoi University of Science and TechnologyApr 6, 2026·also Adelaide, Adelaide University, Australian Institute for Machine, Hanoi-Amsterdam High +1

Beyond the Global Scores: Fine-Grained Token Grounding as a Robust Detector of LVLM Hallucinations

Hallucinated tokens in LVLMs betray themselves through diffuse attention patterns and a failure to semantically align with any specific image region, enabling highly accurate detection.

Tuan Dung Nguyen, Yutong Xie, Nguyen Cam-Tu +5

Eval Frameworks & Benchmarks Interpretability & Mechanistic Interp Multimodal Models

Search

Johan W. Verjans

Research focus

Frequent co-authors

Papers (2)