3 papers indexed
Susmit Agrawal, Matthias Bethge, Matthias Kümmerer
Understanding human visual attention on a scene over time has applications in domains such as interface design and inferring cognitive states. Modeling visual scanpaths has historically relied on specialized architectures with hand-crafted priors. While these architectures can mo…
Sergio Hernández-Gutiérrez, Matteo Merler, Ilze Amanda Auzina, Joschka Strüber, Ameya Prabhu, Matthias Bethge
LLM agents increasingly act over long horizons, where a single trajectory can contain hundreds or thousands of actions. In these settings, outcome-only rewards provide too sparse guidance, failing to inform the model about the goodness of intermediate actions. Dense supervision m…
Matteo Farina, Vishaal Udandarao, Thao Nguyen, Selim Kuzucu, Maximilian Böther, Andreas Hochlehnert, et al.
Building performant Vision-Language Models (VLMs) requires carefully curating large-scale training datasets, yet the community lacks systematic benchmarks for evaluating such curation strategies. We introduce DataComp for VLMs (DCVLM), a benchmark for controlled data-centric expe…