Buchumschlag

Better Representations via Adversarial Training in Pre-Training: A Theoretical Perspective

Fuente: arXiv

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	Xing, Yue, Lin, Xiaofeng, Song, Qifan, Xu, Yi, Zeng, Belinda, Cheng, Guang
Format:	Preprint
Veröffentlicht:	2024
Schlagworte:	Machine Learning
Online-Zugang:	Acceder al recurso
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Ähnliche Einträge

Theoretical Understanding of In-Context Learning in Shallow Transformers with Unstructured Data
von: Xing, Yue, et al.
Veröffentlicht: (2024)

Adversarial Vulnerability as a Consequence of On-Manifold Inseparibility
von: Haldar, Rajdeep, et al.
Veröffentlicht: (2024)

Task-tailored Pre-processing: Fair Downstream Supervised Learning
von: Sohn, Jinwon, et al.
Veröffentlicht: (2026)

Effect of Ambient-Intrinsic Dimension Gap on Adversarial Vulnerability
von: Haldar, Rajdeep, et al.
Veröffentlicht: (2024)

Lower Difficulty and Better Robustness: A Bregman Divergence Perspective for Adversarial Training
von: Wu, Zihui, et al.
Veröffentlicht: (2022)

f-GRPO and Beyond: Divergence-Based Reinforcement Learning Algorithms for General LLM Alignment
von: Haldar, Rajdeep, et al.
Veröffentlicht: (2026)

How to Enhance Downstream Adversarial Robustness (almost) without Touching the Pre-Trained Foundation Model?
von: Liu, Meiqi, et al.
Veröffentlicht: (2025)

On the Clean Generalization and Robust Overfitting in Adversarial Training from Two Theoretical Views: Representation Complexity and Training Dynamics
von: Li, Binghui, et al.
Veröffentlicht: (2023)

Fair Supervised Learning with A Simple Random Sampler of Sensitive Attributes
von: Sohn, Jinwon, et al.
Veröffentlicht: (2023)

Revisiting the Relationship between Adversarial and Clean Training: Why Clean Training Can Make Adversarial Training Better
von: Zhou, MingWei, et al.
Veröffentlicht: (2025)

LLM Safety Alignment is Divergence Estimation in Disguise
von: Haldar, Rajdeep, et al.
Veröffentlicht: (2025)

Zero-Sacrifice Persistent-Robustness Adversarial Defense for Pre-Trained Encoders
von: Lei, Zhuxin, et al.
Veröffentlicht: (2026)

Theoretical Perspectives on Data Quality and Synergistic Effects in Pre- and Post-Training Reasoning Models
von: Javanmard, Adel, et al.
Veröffentlicht: (2026)

Parallelly Tempered Generative Adversarial Nets: Toward Stabilized Gradients
von: Sohn, Jinwon, et al.
Veröffentlicht: (2024)

Pre-Trained AI Model Assisted Online Decision-Making under Missing Covariates: A Theoretical Perspective
von: Hu, Haichen, et al.
Veröffentlicht: (2025)

Improving Clean Accuracy via a Tangent-Space Perspective on Adversarial Training
von: Yi, Bongsoo, et al.
Veröffentlicht: (2024)

HGAN-SDEs: Learning Neural Stochastic Differential Equations with Hermite-Guided Adversarial Training
von: Xu, Yuanjian, et al.
Veröffentlicht: (2025)

When Muon Optimizer Meets Adversarial Training: A Theoretical and Empirical Study
von: Yan, Jun, et al.
Veröffentlicht: (2026)

Stability and Generalization in Free Adversarial Training
von: Cheng, Xiwei, et al.
Veröffentlicht: (2024)

Latent Adversarial Training Improves the Representation of Refusal
von: Abbas, Alexandra, et al.
Veröffentlicht: (2025)

CTSyn: A Foundation Model for Cross Tabular Data Generation
von: Lin, Xiaofeng, et al.
Veröffentlicht: (2024)

Cross-Receiver Generalization for RF Fingerprint Identification via Feature Disentanglement and Adversarial Training
von: Pan, Yuhao, et al.
Veröffentlicht: (2025)

Information-Theoretic Policy Pre-Training with Empowerment
von: Schneider, Moritz, et al.
Veröffentlicht: (2025)

Better Training Data Attribution via Better Inverse Hessian-Vector Products
von: Wang, Andrew, et al.
Veröffentlicht: (2025)

OptiMer: Optimal Distribution Vector Merging Is Better than Data Mixing for Continual Pre-Training
von: Song, Haiyue, et al.
Veröffentlicht: (2026)

Rethinking Pre-Training in Tabular Data: A Neighborhood Embedding Perspective
von: Ye, Han-Jia, et al.
Veröffentlicht: (2023)

Information Theoretic Adversarial Training of Large Language Models
von: Zhang, Yiwei, et al.
Veröffentlicht: (2026)

Two Heads Are Better Than One: Boosting Graph Sparse Training via Semantic and Topological Awareness
von: Zhang, Guibin, et al.
Veröffentlicht: (2024)

HG-Adapter: Improving Pre-Trained Heterogeneous Graph Neural Networks with Dual Adapters
von: Mo, Yujie, et al.
Veröffentlicht: (2024)

Relational Learning in Pre-Trained Models: A Theory from Hypergraph Recovery Perspective
von: Chen, Yang, et al.
Veröffentlicht: (2024)

Towards a Better Theoretical Understanding of Independent Subnetwork Training
von: Shulgin, Egor, et al.
Veröffentlicht: (2023)

FairRR: Pre-Processing for Group Fairness through Randomized Response
von: Zeng, Xianli, et al.
Veröffentlicht: (2024)

Scaling Adversarial Training via Data Selection
von: Ye, Youran, et al.
Veröffentlicht: (2025)

Adversarial Training from Mean Field Perspective
von: Kumano, Soichiro, et al.
Veröffentlicht: (2025)

SQS: Bayesian DNN Compression through Sparse Quantized Sub-distributions
von: Wang, Ziyi, et al.
Veröffentlicht: (2025)

Representation Learning of Multivariate Time Series using Attention and Adversarial Training
von: Scharwächter, Leon, et al.
Veröffentlicht: (2024)

Adjustment for Confounding using Pre-Trained Representations
von: Schulte, Rickmer, et al.
Veröffentlicht: (2025)

An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learning
von: Yoon, Jaesik, et al.
Veröffentlicht: (2023)

Adversarial Training for Graph Neural Networks via Graph Subspace Energy Optimization
von: Liu, Ganlin, et al.
Veröffentlicht: (2024)

Stochastic Forward-Backward Deconvolution: Training Diffusion Models with Finite Noisy Datasets
von: Lu, Haoye, et al.
Veröffentlicht: (2025)