AI safety and alignment

Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.

17,788 stories · RSS feed

arXiv Machine Learning

Oct 15

Large-Step Last-Iterate Convergence of Random-Reshuffling Stochastic Gradient Descent

arXiv:2610.19444v1 Announce Type: cross Abstract: Random reshuffling stochastic gradient descent (RR-SGD) is widely used for finite-sum optimization in machine learning. For convex $L$-smooth compone...

By Zijun Chen, Zihan Zhang

safety

arXiv Machine Learning

Oct 15

Fractional is Better: Learnable Derivative Orders in Neural Operator Learning

arXiv:2610.20363v1 Announce Type: new Abstract: Neural operators learn mappings between function spaces, enabling fast surrogate solutions to partial differential equations. Despite remarkable archit...

By Fares B. Mehouachi, Saif Eddin Jabari

benchmarks safety

arXiv Machine Learning

Oct 15

Bayesian Causal Forests for Zero-Inflated Continuous Outcomes

arXiv:2610.18889v1 Announce Type: cross Abstract: Estimating (heterogeneous) treatment effects is still challenging when the outcome is semicontinuous (i.e., exhibiting both a discrete point mass at...

By Hugo Gobato Souto, Eric Wang, Igor Shuryak

benchmarks safety

arXiv Machine Learning

Oct 15

Noise-Aware Differentially Private Variational Inference

arXiv:2410.19371v4 Announce Type: replace-cross Abstract: Differential privacy (DP) provides robust privacy guarantees for statistical inference, but this can lead to unreliable results and biases in...

By Talal Alrawajfeh, Joonas J\"alk\"o, Antti Honkela

safety

arXiv Machine Learning

Oct 15

Bridging the Topological Divide: Dual-Branch Graph Distillation with Semantic Priors for Robust Cold-Start Recommendation

arXiv:2610.19578v1 Announce Type: cross Abstract: Graph Collaborative Filtering models have achieved remarkable success in recommendation systems. However, they inherently suffer from a catastrophic...

By Yuanfeng Qiu

efficiency multimodal benchmarks safety

arXiv Machine Learning

Oct 15

Continual pretraining Prostate MRI ViT Encoder for task alignment

arXiv:2610.20217v1 Announce Type: cross Abstract: Building foundation models for MRI has emerged as a promising direction for advancing medical image analysis and improving accessibility of task spec...

By Bilal A. Sidiqi, Yipei Wang, Shaheer U. Saeed, Shiqi Huang, Daniel C. Alexander, Yipeng Hu

llms computer-vision fine-tuning safety

arXiv Machine Learning

Oct 15

SafeAdapt-Poker: Counterfactual Risk Calibration for Budgeted Adaptation under Opponent Shift

arXiv:2610.20473v1 Announce Type: cross Abstract: Opponent histories can mislead after behavioral shifts. SafeAdapt-Poker formulates budgeted adaptation under opponent shift as a dual-counterfactual...

By Miaobo Hu, Shuhao Hu, Xiaobo Guo, Xin Wang, Bokun Wang, Rui Chen, Daren Zha, Jun Xiao

safety

arXiv Machine Learning

Oct 15

Generation of data on discontinuous manifolds via continuous stochastic non-invertible networks

arXiv:2112.09646v2 Announce Type: replace Abstract: The generation of discontinuous distributions is a difficult task for most known frameworks such as generative autoencoders and generative adversar...

By Mariia Drozdova, Vitaliy Kinakh, Guillaume Qu\'etant, Tobias Golling, Slava Voloshynovskiy

diffusion safety

arXiv Machine Learning

Oct 15

GLL: A Differentiable Graph Learning Layer for Neural Networks

arXiv:2412.08016v3 Announce Type: replace Abstract: Standard deep learning architectures used for classification generate label predictions with a projection head and softmax activation function. Alt...

By Jason Brown, Bohan Chen, Harris Hardiman-Mostow, Jeff Calder, Andrea L. Bertozzi

safety

arXiv Machine Learning

Oct 15

SWE-Spot: Building Small Repo-Experts with Repository-Centric Learning

arXiv:2601.21649v2 Announce Type: replace Abstract: The deployment of coding agents in privacy-sensitive and resource-constrained environments drives the demand for capable open-weight Small Language...

By Jinjun Peng, Magnus Saebo, Tianjun Zhong, Junfeng Yang, Shi-Xiong Zhang, Baishakhi Ray, Simin Chen, Yangruibo Ding

llms agents safety

arXiv Machine Learning

Oct 15

Causally-Constrained Probabilistic Forecasting for Time-Series Anomaly Detection

arXiv:2604.17998v2 Announce Type: replace Abstract: Anomaly detection in multivariate time series is a central challenge in industrial monitoring, as failures frequently arise from complex temporal d...

By Pooyan Khosravinia, Jo\~ao Gama, Bruno Veloso

llms benchmarks safety

arXiv Machine Learning

Oct 15

Privacy-Preserving GNN Inference on Financial Transaction Graphs via Homomorphic Encryption

arXiv:2610.18054v2 Announce Type: replace-cross Abstract: Anti-money laundering (AML) in financial networks requires modeling transaction patterns that span multiple institutions. In practice, howeve...

By H. \c{C}a\u{g}r{\i} Bilgi, Kemal Mutluergil, Ceren Y{\i}ld{\i}r{\i}m, Kubilay Atasu, Erkay Savas

efficiency benchmarks safety

arXiv Machine Learning

Oct 15

TabFORGE: A Tabular Foundation Model for Generative Modelling

arXiv:2605.09424v2 Announce Type: replace Abstract: Generative modelling is a demanding test of foundation models, because it requires robust, holistic representation learning for a given data modali...

By Xiangjian Jiang, Mingxuan Liu, Nikola Simidjievski, Tassilo Klein, Mateja Jamnik

llms rag diffusion benchmarks safety data-engineering

arXiv Machine Learning

Oct 15

LEAP: Unlocking dLLM Parallelism via Lookahead Early-Convergence Token Detection

arXiv:2605.10980v2 Announce Type: replace Abstract: Diffusion Language Models (dLLMs) have garnered significant attention for their potential in highly parallel processing. The parallel capabilities...

By Haohui Zhang, Zhiye Wang, Xiaoying Gan, Xinbing Wang, Bo Jiang

llms diffusion efficiency benchmarks safety

arXiv Machine Learning

Oct 15

Activation Steering Induces Emergent Misalignment: Steering Vector Construction and Comprehensive Evaluation

arXiv:2606.08682v2 Announce Type: replace Abstract: Activation steering (AS) has emerged as a popular inference-time technique for modulating the behavior of large language models (LLMs). AS modulate...

By Qi Cao, Jian Lou, Meiting Liu, Wenjie Feng, Dan Li, Anh Tuan Luu

llms fine-tuning safety

arXiv Machine Learning

Oct 15

Deep kernel hedging

arXiv:2609.34474v2 Announce Type: replace Abstract: We introduce a deep kernel hedging framework that combines the flexibility of deep learning with the structural inductive bias of kernel methods. T...

By Jean-Loup Dupret, Donatien Hainaut, Edouard Motte

rag safety

arXiv Machine Learning

Oct 15

Imaging-anchored Multiomics in Cardiovascular Disease: Integrating Cardiac Imaging, Bulk, Single-cell, and Spatial Transcriptomics

arXiv:2601.07871v2 Announce Type: replace-cross Abstract: Cardiovascular disease arises from interactions between inherited risk, molecular programmes, and tissue-scale remodelling that are observed...

By Minh H. N. Le, Tuan Vinh, Thanh-Huy Nguyen, Tao Li, Bao Quang Gia Le, Han H. Huynh, Monika Raj, Carl Yang, Min Xu, Nguyen Quoc Khanh Le

multimodal safety

arXiv Machine Learning

Oct 15

Sharp Statistical Thresholds for Attributed Gaussian Graph Matching

arXiv:2603.23305v2 Announce Type: replace-cross Abstract: We investigate the problem of matching networks where both edge weights and node features are correlated following a specific node correspond...

By Mohammad Hassan Ahmad Yarandi, Luca Ganassali

benchmarks safety

arXiv Computation and Language

Oct 15

Can LLMs Access their World Knowledge for Event Prediction?

arXiv:2610.19021v1 Announce Type: new Abstract: Recent work has questioned whether large language models (LLMs) can be useful for temporal event prediction. Despite vast pretrained world knowledge, L...

By Zijun Wu, Zhitian Zhang, Hossein Hajimirsadeghi, Nikita Dvornik, Alexander Pashevich

llms benchmarks safety

arXiv Computation and Language

Oct 15

Vision-Language Models Integrate Conflicting Evidence Like Biased Bayesian Observers

arXiv:2610.19407v1 Announce Type: new Abstract: Vision-language models must reconcile conflicting text and images. Prior work has characterized the resulting behavior in terms of modality biases--e.g...

By Josh Barua, Kyle Mahowald, William Rudman

llms multimodal safety

Get the daily briefing