8 papers
PRiSM: Prototype Regularization for Few-Shot VLMs
Ghassen Baklouti, Omprakash Chakraborty, Jose Dolz +1
Training-free few-shot adaptation methods have gained significant attention recently in the context of Vision-language Models (VLMs). Yet, current benchmarks rely on strong assumpt…
Quantile Adaptive Temperature Scaling for Confidence Calibration
Omprakash Chakraborty, Leo Fillioux, Ismail Ben Ayed +1
Deep neural networks often produce poorly calibrated confidence estimates, overstating their certainty even when predictions are incorrect. Temperature Scaling remains the most wid…
Are Online Skill and Memory Modules Always Worth Their Tokens? A Budget-Constrained Study of Web Agents
Sina Hajimiri, Masih Aminbeidokhti, Jose Dolz +4
Online web agents often augment a base actor with memory, workflow, or skill modules. These modules can improve performance, but they also consume test-time tokens, a cost rarely r…
MolRGen: A Training and Evaluation Setting for De Novo Molecular Generation with Reasonning Models
Philippe Formont, Maxime Darrin, Ismail Ben Ayed +1
Recent reasoning-based large language models have shown strong performance on tasks with verifiable outcomes, but their use in de novo molecular generation remains limited by the l…
ORION: ORthonormal Text Encoding for Universal VLM AdaptatION
Omprakash Chakraborty, Jose Dolz, Ismail Ben Ayed
Vision language models (VLMs) have demonstrated remarkable generalization across diverse tasks, yet their performance remains constrained by the quality and geometry of the textual…
Information Maximization for Long-Tailed Semi-Supervised Domain Generalization
Leo Fillioux, Omprakash Chakraborty, Quentin Gopée +6
Semi-supervised domain generalization (SSDG) has recently emerged as an appealing alternative to tackle domain generalization when labeled data is scarce but unlabeled samples acro…