AI safety and alignment

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

18,070 stories · RSS feed

arXiv Machine Learning

Jul 27

Toward User-Conditioned Evaluation of Personal LLM Agents under Temporal Interventions

arXiv:2607. 21635v1 Announce Type: new Abstract: Personal agents maintain memories, learned skills, tool configurations, and policy state that evolve with each user.

By Pin Qian, Su Wang, Yihang Chen, Qiaolin Yu, Xiaoyuan Wang, Zhitong Guo, Zhicheng Wang, Junxian You

llms agents benchmarks safety

arXiv Machine Learning

Jul 27

The pretraining domain outweighs the training objective in setting the privacy-utility trade-off of differentially private medical image analysis

arXiv:2601. 19618v2 Announce Type: replace-cross Abstract: Differential privacy protects the patients whose images train medical imaging models, but it lowers diagnostic accuracy, and the initialization is the strongest known remedy.

By Soroosh Tayebi Arasteh, Mina Farajiamiri, Mahshad Lotfinia, Behrus Hinrichs-Puladi, Jonas Bienzeisler, Mohamed Alhaskir, Mirabela Rusu, Christiane Kuhl, Sven Nebelung, Daniel Truhn

safety

arXiv Machine Learning

Jul 27

Spatially-Enhanced Temporal Fusion Transformer: Interpretable Multi-Output Prediction for Parametric Dynamical Systems with Time-Varying Inputs

arXiv:2505. 00473v2 Announce Type: replace Abstract: We explore the promising performance of a transformer model in predicting outputs of parametric dynamical systems with external time-varying input signals.

By Shuwen Sun, Lihong Feng, Peter Benner

llms safety

arXiv Machine Learning

Jul 27

Unbiased Open World Regularization for Fair Self-Supervised Learning

arXiv:2607. 22149v1 Announce Type: new Abstract: Despite recent advances, self-supervised learning (SSL) models and Joint-Embedding Predictive Architectures (JEPAs) remain susceptible to learning spurious biases in the dataset.

By L{\'e}o Nicollier (CB, ATT), Marc Pic (ATT), Pablo Mus{\'e} (CB, IFUMI), Enric Meinhardt-Llopis (CB), Gabriele Facciolo (CB)

rag diffusion benchmarks safety

arXiv Machine Learning

Jul 27

Safe In-Context Reinforcement Learning

arXiv:2509. 25582v4 Announce Type: replace Abstract: In-context reinforcement learning (ICRL) is an emerging RL paradigm where an agent, after pretraining, can adapt to out-of-distribution test tasks without any parameter updates, instead relying on an expanding context of interaction history.

By Amir Moeini, Minjae Kwon, Alper Kamil Bozkurt, Yuichi Motai, Rohan Chandra, Lu Feng, Shangtong Zhang

agents reinforcement-learning benchmarks safety

arXiv Machine Learning

Jul 27

MA-DAR: Manifold-Aligned Dynamic Adaptive Routing for Continual Temporal Knowledge Graph Reasoning

arXiv:2607. 21949v1 Announce Type: new Abstract: Continual temporal knowledge graph (TKG) reasoning aims to continuously incorporate newly emerging facts while preserving previously acquired knowledge.

By Xiangjun Shi, Chong Mu, Jinchuan Zhang, Lizong Zhang, Yuefeng He, Shang Liu

benchmarks safety

arXiv Machine Learning

Jul 27

Industrial Tokenization for LLM-Based Health Intelligence: A Federated Architecture for Industrial Evidence Integration

arXiv:2607. 22153v1 Announce Type: cross Abstract: Industrial health management increasingly relies on heterogeneous information sources, including condition monitoring systems, supervisory control and data acquisition systems, maintenance records, inspection results, and prognostic models.

By Deshui Li, Xiao-Ming Yuan, Zishun Wang

llms agents nlp safety

arXiv Machine Learning

Jul 27

Optimal generalisation and learning transition in extensive-width shallow neural networks near interpolation

arXiv:2501. 18530v3 Announce Type: replace-cross Abstract: We consider a teacher-student model of supervised learning with a fully-trained two-layer neural network whose width $k$ and input dimension $d$ are large and proportional.

By Jean Barbier, Francesco Camilli, Minh-Toan Nguyen, Mauro Pastore, Rudy Skerk

safety

arXiv Machine Learning

Jul 27

Autoregressive EHR Foundation Models with Multimodal Inputs

arXiv:2607. 22264v1 Announce Type: new Abstract: Autoregressive foundation models trained on tokenized electronic health records (EHRs) can support zero-shot clinical prediction, yet most operate on structured event codes alone, and do not incorporate multiple modalities in a principled way.

By Yuxuan Liu, Joshua Placidi, Jinpei Han, Alfred John Balston, Marek Rei, A. Aldo Faisal

multimodal safety

arXiv Machine Learning

Jul 27

Carpe Diem: Critical Learning Period-Aware Contract-Based Incentives for Federated Learning

arXiv:2503. 07869v4 Announce Type: replace Abstract: Critical learning periods (CLPs) in federated learning (FL) refer to early stages during which low-quality contributions (e.

By Thanh Linh Nguyen, Dinh Thai Hoang, Diep N. Nguyen, Quoc-Viet Pham

safety

arXiv Machine Learning

Jul 27

Reasoning Denoiser: Denoising Reasoning Traces for Hallucination Detection in Large Reasoning Models

arXiv:2607. 22098v1 Announce Type: cross Abstract: Large reasoning models (LRMs) generate long reasoning traces before producing final answers.

By Junlin Fang, Do Nguyen-Thanh, Xiaogang Xu, Zhen Fang, Sean Du

rag benchmarks safety

arXiv Machine Learning

Jul 27

Hopformer: Homogeneity-Pursuit Transformer for Time Series Forecasting

arXiv:2607. 22299v1 Announce Type: cross Abstract: Forecasting multiple time-series with high-dimensional covariates presents a core challenge: unifying common temporal patterns while retaining meaningful series-specific information.

By Wan Zhang, Qinjie Lin, Chan Lee, Weijian Li, Han Liu, Kai Zhang

llms fine-tuning efficiency benchmarks safety

arXiv Machine Learning

Jul 27

Quantum Spectral Model: Data Reuploading with Input-Conditioned Frequency Support

arXiv:2607. 22516v1 Announce Type: cross Abstract: A central design principle in modern machine learning and artificial intelligence is to align a model's inductive bias with the structure of its input data.

By Peiyong Wang, Udaya Parampalli, Casey R. Myers

benchmarks safety

arXiv Machine Learning

Jul 27

Security Without Detection: Economic Denial as a Primitive for Edge and IoT Defense

arXiv:2512. 23849v2 Announce Type: replace-cross Abstract: Sophisticated attackers can evade detection-based security by using encryption, stealth tactics, and low-rate attack patterns.

By Samaresh Kumar Singh, Joyjit Roy, Sriharsha Anand Pushkala

safety

arXiv Machine Learning

Jul 27

FBLayout: Optimizing Memory Layout for Efficient LLM Finetuning on Mobile GPUs

arXiv:2607. 21624v1 Announce Type: cross Abstract: Transformer-based models have enabled unprecedented capabilities across language, vision, and multimodal tasks.

By Kahou Tam, Wei Niu, Yu Bao, Xiaomin Ouyang, Chengzhong Xu, Li Li

llms fine-tuning efficiency multimodal safety

arXiv Machine Learning

Jul 27

Neural Feature Governance: Extending Atom Prevalence

arXiv:2607. 21671v1 Announce Type: new Abstract: Neural network compression and interpretability remain open challenges in modern deep learn- ing, where billion-parameter architectures deliver impressive accuracy at the cost of trans- parency, computational efficiency, and reliable uncertainty quantification.

By Idris Karel Seunda Ekwe, Patrick Tenga Shako, Ernest Parfait Fokou\'e

computer-vision fine-tuning efficiency benchmarks safety

arXiv Machine Learning

Jul 27

Interpretable EEG biomarkers with bag-of-waves: Spatial and temporal waveform dictionaries for low-data regimes

arXiv:2607. 22508v1 Announce Type: new Abstract: Electroencephalography (EEG) is widely used to diagnose neurological conditions, but its analysis usually relies on either predefined spectral features or deep neural networks.

By Athanasios Papastathopoulos-Katsaros, Steven T. Lee, Lin Yao, Ajay Thomas, Junseok Park, Matthew J. McGinley, Zhandong Liu

benchmarks safety

arXiv Machine Learning

Jul 27

Reliability Scales Inversely: Bigger Language Models Compound Mistakes Faster

arXiv:2607. 18292v2 Announce Type: replace Abstract: As language models scale, answers start truer but degrade faster: scaling buys capability but erodes reliability.

By Kushal Chakrabarti

llms safety

arXiv Machine Learning

Jul 27

Dysphagia Risk Stratification in Head and Neck Cancer via Two-Stage PRO-Clinical Stacking

arXiv:2607. 22514v1 Announce Type: new Abstract: Dysphagia is a debilitating late effect of head and neck cancer (HNC) treatment, yet timely identification of at-risk patients remains challenging in survivorship care.

By Siyuan Zhao, Eric Ababio Anyimadu, Zachary G. Brumm, Yue Ma, Clifton David Fuller, Xinhua Zhang, G. Elisabeta Marai, Guadalupe Canahuate

safety

arXiv Machine Learning

Jul 27

Cloud-Native Evaluation-as-a-Service: A Microservices Architecture for Scalable AI Monitoring with Conformal Guarantees

arXiv:2607. 21623v1 Announce Type: new Abstract: We present EaaS, a cloud-native reference architecture that operationalizes AI evaluation methods as six stateless Kubernetes microservices: conformal prediction with finite-sample-corrected Adaptive Prediction Sets, calibration assessment, drift detection via RFF-approximated Maximum Mean Discrepancy, fairness monitoring with bootstrap confidence intervals, a DAG-based pipeline orchestrator, and a result storage API.

By Lei Yang

agents benchmarks safety

Get the daily briefing