Layer by layer, module by module: Choose both for optimal OOD probing of ViT
Paper • 2603.05280 • Published
Ambroise Odonnat
ambroiseodt
AI & ML interests
Transformers, distribution shifts, optimization
Recent Activity
submitted a paper about 1 month ago
Layer by layer, module by module: Choose both for optimal OOD probing of ViT authored a paper about 1 month ago
Layer by layer, module by module: Choose both for optimal OOD probing of ViT authored a paper about 1 month ago
Leveraging Gradients for Unsupervised Accuracy Estimation under Distribution ShiftOrganizations
submitted a
paper to Daily Papers about 1 month ago
authored 2
papers about 1 month ago
authored 4
papers 2 months ago
Provable Benefits of In-Tool Learning for Large Language Models
Paper • 2508.20755 • Published • 11
CauKer: classification time series foundation models can be pretrained on synthetic data only
Paper • 2508.02879 • Published
Vision Transformer Finetuning Benefits from Non-Smooth Components
Paper • 2602.06883 • Published • 4
Optimal Self-Consistency for Efficient Reasoning with Large Language Models
Paper • 2511.12309 • Published
submitted a
paper to Daily Papers 2 months ago
authored 3
papers about 1 year ago
Analysing Multi-Task Regression via Random Matrix Theory with Application to Time Series Forecasting
Paper • 2406.10327 • Published
Easing Optimization Paths: a Circuit Perspective
Paper • 2501.02362 • Published
SKADA-Bench: Benchmarking Unsupervised Domain Adaptation Methods with Realistic Validation On Diverse Modalities
Paper • 2407.11676 • Published
updated a
Space about 1 year ago
upvoted a paper over 1 year ago
authored a
paper over 1 year ago
upvoted a paper over 1 year ago
commented a
paper over 1 year ago
authored 4
papers over 1 year ago
Large Language Models as Markov Chains
Paper • 2410.02724 • Published • 33
MANO: Exploiting Matrix Norm for Unsupervised Accuracy Estimation Under Distribution Shifts
Paper • 2405.18979 • Published • 1
Leveraging Ensemble Diversity for Robust Self-Training in the Presence of Sample Selection Bias
Paper • 2310.14814 • Published • 1
SAMformer: Unlocking the Potential of Transformers in Time Series Forecasting with Sharpness-Aware Minimization and Channel-Wise Attention
Paper • 2402.10198 • Published • 1