Pascal Frossard

papers

984

total citations

papers (20)

Task Arithmetic in the Tangent Space: Improved Editing of Pre-Trained Models

NEURIPS 2023arXiv

185

citations

Building powerful and equivariant graph neural networks with structural message-passing

NEURIPS 2020arXiv

136

citations

IS-Fusion: Instance-Scene Collaborative Fusion for Multimodal 3D Object Detection

CVPR 2024arXiv

citations

PRIME: A Few Primitives Can Boost Robustness to Common Corruptions

ECCV 2022arXiv

citations

Hold me tight! Influence of discriminative features on deep network boundaries

NEURIPS 2020arXiv

citations

DeFoG: Discrete Flow Matching for Graph Generation

ICML 2025arXiv

citations

Joint Graph-Based Depth Refinement and Normal Estimation

CVPR 2020arXiv

citations

LiNeS: Post-training Layer Scaling Prevents Forgetting and Enhances Model Merging

ICLR 2025arXiv

citations

Neural Anisotropy Directions

NEURIPS 2020arXiv

citations

Pareto Low-Rank Adapters: Efficient Multi-Task Learning with Preferences

ICLR 2025arXiv

citations

MEMOIR: Lifelong Model Editing with Minimal Overwrite and Informed Retention for LLMs

NEURIPS 2025arXiv

citations

Pi-DUAL: Using privileged information to distinguish clean from noisy labels

ICML 2024arXiv

citations

Sequential Representation Learning via Static-Dynamic Conditional Disentanglement

ECCV 2024arXiv

citations

U-Boost NAS: Utilization-Boosted Differentiable Neural Architecture Search

ECCV 2022arXiv

citations

Inductive Domain Transfer In Misspecified Simulation-Based Inference

NEURIPS 2025arXiv

citations

What can linearized neural networks actually say about generalization?

NEURIPS 2021

citations

Flow based approach for Dynamic Temporal Causal models with non-Gaussian or Heteroscedastic Noises

NEURIPS 2025arXiv

citations

Pascal Frossard

papers (20)

Task Arithmetic in the Tangent Space: Improved Editing of Pre-Trained Models

Building powerful and equivariant graph neural networks with structural message-passing

GeoDA: A Geometric Framework for Black-Box Adversarial Attacks

A Structured Dictionary Perspective on Implicit Neural Representations

Localizing Task Information for Improved Model Merging and Compression

IS-Fusion: Instance-Scene Collaborative Fusion for Multimodal 3D Object Detection

PRIME: A Few Primitives Can Boost Robustness to Common Corruptions

Hold me tight! Influence of discriminative features on deep network boundaries

DeFoG: Discrete Flow Matching for Graph Generation

Joint Graph-Based Depth Refinement and Normal Estimation

LiNeS: Post-training Layer Scaling Prevents Forgetting and Enhances Model Merging

Neural Anisotropy Directions

Pareto Low-Rank Adapters: Efficient Multi-Task Learning with Preferences

MEMOIR: Lifelong Model Editing with Minimal Overwrite and Informed Retention for LLMs

Pi-DUAL: Using privileged information to distinguish clean from noisy labels

Sequential Representation Learning via Static-Dynamic Conditional Disentanglement

U-Boost NAS: Utilization-Boosted Differentiable Neural Architecture Search

Inductive Domain Transfer In Misspecified Simulation-Based Inference

What can linearized neural networks actually say about generalization?

Flow based approach for Dynamic Temporal Causal models with non-Gaussian or Heteroscedastic Noises

papers (20)

Task Arithmetic in the Tangent Space: Improved Editing of Pre-Trained Models

Building powerful and equivariant graph neural networks with structural message-passing

GeoDA: A Geometric Framework for Black-Box Adversarial Attacks

A Structured Dictionary Perspective on Implicit Neural Representations

Localizing Task Information for Improved Model Merging and Compression

IS-Fusion: Instance-Scene Collaborative Fusion for Multimodal 3D Object Detection

PRIME: A Few Primitives Can Boost Robustness to Common Corruptions

Hold me tight! Influence of discriminative features on deep network boundaries

DeFoG: Discrete Flow Matching for Graph Generation

Joint Graph-Based Depth Refinement and Normal Estimation

LiNeS: Post-training Layer Scaling Prevents Forgetting and Enhances Model Merging

Neural Anisotropy Directions

Pareto Low-Rank Adapters: Efficient Multi-Task Learning with Preferences

MEMOIR: Lifelong Model Editing with Minimal Overwrite and Informed Retention for LLMs

Pi-DUAL: Using privileged information to distinguish clean from noisy labels

Sequential Representation Learning via Static-Dynamic Conditional Disentanglement

U-Boost NAS: Utilization-Boosted Differentiable Neural Architecture Search

Inductive Domain Transfer In Misspecified Simulation-Based Inference

What can linearized neural networks actually say about generalization?

Flow based approach for Dynamic Temporal Causal models with non-Gaussian or Heteroscedastic Noises