safety

CLIP4VI-ReID: Learning Modality-shared Representations via CLIP Semantic Bridge for Visible-Infrared Person Re-identification

arXiv:2511.10309v2 Announce Type: replace Abstract: This paper proposes a novel CLIP-driven modality-shared representation learning network named CLIP4VI-ReID for VI-ReID task, which consists of Text

Safety

⚠️⚠️⚠️ Constitutional AI is not working. Aligning LLMs is not working. We need a different approach. If society doesn’t place its bets diffe…

DGX agent

⚠️⚠️⚠️ Constitutional AI is not working. Aligning LLMs is not working. We need a different approach. If society doesn’t place its bets differently, we are screwed. Anthropic's Mythos created fake iden

safetygary-marcus--x

5 Aug 2026

Continue or Replan? Bernoulli-Continuation Policy Learning for Adaptive Horizon Execution

arXiv:2608.03483v1 Announce Type: cross Abstract: Existing chunk-based Vision-Language-Action (VLA) models execute a fixed number of actions (i.e., execution horizon) before replanning, turning replan

Control Barrier Functions via Minkowski Operations for Safe Navigation among Polytopes

arXiv:2608.02886v1 Announce Type: new Abstract: Safely navigating polytopic environments while respecting the dynamics, control, and exact geometry of the underlying system is a challenge in robotics.

CROSS: Cascaded Distillation and Dual-Constraint Grounding for Remote Sensing Referring Segmentation

arXiv:2608.03147v1 Announce Type: new Abstract: Referring Remote Sensing Image Segmentation (RRSIS) has achieved significant progress through the integration of VLMs and the Segment Anything Model (SA

CT-HEG: A Bidirectional, Timestamp-Attributed Event Graph for ICU In-Hospital Mortality Prediction - An Architectural Ablation Study

arXiv:2608.02663v1 Announce Type: cross Abstract: Accurate ICU mortality prediction requires modeling irregular clinical observations across heterogeneous entity types. Existing sequence models handle

CVPO: Enhancing LLM Reinforcement Learning Reasoning via Value-Variance Adaptation and Dynamic Curriculum Learning

arXiv:2608.03068v1 Announce Type: cross Abstract: Reinforcement learning (RL) has emerged as an effective method for enhancing the reasoning capabilities of large language models (LLMs). However, exis

Deferred Exposure of Future Trajectories for Verifiable Reasoning in Autonomous Driving VLMs

arXiv:2608.01755v2 Announce Type: replace Abstract: Recent Vision-Language-Action (VLA) models for autonomous driving (AD) increasingly utilize chain-of-thought (CoT) supervision to enhance the reason

Disentangling Language Modeling and Boundaries

arXiv:2608.03599v1 Announce Type: new Abstract: Byte-level language models are usually argued for on the grounds of robustness, multilingual fairness, and character-level skills. We point to a differe

DiverseDiT++: Quantifying, Analyzing, and Promoting Representation Diversity in Diffusion Transformers

arXiv:2608.03082v1 Announce Type: new Abstract: Recent advances in Diffusion Transformers (DiTs) have enabled remarkable progress in visual synthesis, benefiting from their superior scalability. To fa

DocTrace: Towards Traceable Long Document VQA via Hierarchical Evidence Graph Reasoning

arXiv:2608.03292v1 Announce Type: new Abstract: Long Document Visual Question Answering (LongDocVQA) requires Multimodal Large Language Models (MLLMs) to locate, integrate, and reason over heterogeneo

Don't Peek at the Answer: Outcome-Masked Group Relative Policy Optimization for Label-Free RLVR

arXiv:2608.03119v1 Announce Type: new Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) improves LLM reasoning but typically relies on ground-truth (GT) answers, limiting scalability. Vo

Double Down on Defense: Strengthening Deep Perceptual Hashes against Evasion Attacks without Retraining

arXiv:2608.03101v1 Announce Type: new Abstract: Near-duplicate image matching is crucial for trust and safety, provenance verification, copyright enforcement, and large-scale visual search. Modern pla

DRIFT: Derailing Denoising Trajectories of Flow-Matching VLAs with Adversarial Patch Attack

arXiv:2608.03207v1 Announce Type: new Abstract: Flow-matching vision-language-action (VLA) models such as pi0 generate robot actions by integrating a learned denoising velocity field, and have been re

DriftWorld: Fast World Modeling through Drifting

arXiv:2607.15065v2 Announce Type: replace-cross Abstract: Predictive world models enable robots to plan by imagining the outcomes of their actions, but their value for control hinges on generating man

Emulate or Estimate? The Divergent Strengths of Base and Post-Trained Language Models for Opinion Simulation

arXiv:2608.03044v1 Announce Type: cross Abstract: Large language models are increasingly used to simulate human opinions, but prior work reports conflicting results: some studies find promising alignm

Enhanced Polarization Locking in VCSELs

arXiv:2604.01857v2 Announce Type: replace-cross Abstract: While optical injection locking (OIL) of vertical-cavity surface-emitting lasers (VCSELs) has been widely studied in the past, the polarizatio

Enhancing Q-Value Updates in Deep Q-Learning via Successor-State Prediction

arXiv:2511.03836v2 Announce Type: replace Abstract: Deep Q-Networks (DQNs) estimate future returns by learning from transitions sampled from a replay buffer. However, the target updates in DQN often r

safetyarxiv-cs-lg

Enhancing VLM Reward Models Through Structure-Aware Fine-Tuning

arXiv:2608.03875v1 Announce Type: cross Abstract: Designing effective reward functions remains a major bottleneck in Reinforcement Learning (RL). Recent work uses large foundation Vision-Language Mode

Evading Chain-of-Thought Monitoring Through Model Poisoning

arXiv:2608.02820v1 Announce Type: cross Abstract: Chain-of-thought (CoT) monitoring is an increasingly important component of AI safety stacks but relies on the assumption that a model's reasoning tra

EvoHIL: Self-Evolving Reward and Flow-Matched Policy Optimization for Robust Human-in-the-Loop Reinforcement Learning

arXiv:2608.03872v1 Announce Type: new Abstract: Human-in-the-loop reinforcement learning (HIL-RL) enables robots to learn contact-rich manipulation from limited real-world interaction, but deployment

Fast Object Removal Attacks on Safety-Critical Video-based Perception Systems

arXiv:2608.02806v1 Announce Type: cross Abstract: By leveraging data from video-based perception systems, intelligent transportation systems (ITS) support safety-critical applications that improve roa

Fields Medalist Jacob Tsimerman knows MUCH MUCH more about math than I do, or ever will, but I have been studying natural and artificial int…

Fields Medalist Jacob Tsimerman knows MUCH MUCH more about math than I do, or ever will, but I have been studying natural and artificial intelligence for 40 years, and I think his prediction here (“AI

safetygary-marcus--x

Flying over The Uncertain Nature (FORTUNE): Intelligent and Humanistic 3D Path Planning for Low-Altitude Collaboration

arXiv:2608.03408v1 Announce Type: new Abstract: The proliferation of low-altitude intelligent agents is increasing the demand for timely and socially responsible collaborative sensing in dynamic urban

Forbidden Region Dynamic Active Constraints in Robot-Assisted Minimally Invasive Surgery

arXiv:2608.03010v1 Announce Type: new Abstract: In robot-assisted surgery, Forbidden Region Active Constraints (FRAC) represent a control strategy that helps maintain task safety by generating anisotr

From Routes to Steps: Separating Semantic Progress from Local Execution in Vision-and-Language Navigation

arXiv:2608.03143v1 Announce Type: new Abstract: Vision-and-Language Navigation (VLN) requires an agent to follow a route-level instruction by executing its constituent steps from egocentric visual obs

GenOS: Compositional Certificates for Semantic Robustness in AI Code Generation

arXiv:2608.03588v1 Announce Type: cross Abstract: AI coding agents are stochastic workflows: prompts are interpreted, artifacts are sampled, validators produce observations, and orchestrators commit o

GORDON: Graph-based Object-centric Rewards for Decomposition of Long-Horizon Manipulation

arXiv:2608.03753v1 Announce Type: new Abstract: Learning long-horizon manipulation skills with reinforcement learning remains challenging due to the complexity of reward design, the limited guidance o

GROW: Group-Relative Advantage-Weighted On-Policy Reinforcement Learning of Autoregressive-Diffusion Text-to-Speech model

arXiv:2608.03215v1 Announce Type: cross Abstract: Reinforcement learning for flow-matching text-to-speech is complicated by deterministic ODE sampling: trajectory-level policy-gradient methods typical

HERO: Hierarchical Evidential Reasoning Optimization for Radiology Report Generation via Reason-then-Summarize

arXiv:2601.03321v3 Announce Type: replace-cross Abstract: Multimodal Large Language Models (MLLMs) have substantially advanced Radiology Report Generation (RRG), yet aligning them through reinforcemen

Hi-Token: Hierarchical Coordinate Tokenization for Generative Visual Grounding

arXiv:2608.03471v1 Announce Type: new Abstract: Generative Vision-Language Models (VLMs) commonly treat bounding-box coordinates as independent output symbols, leaving numerical order and axis semanti

Hi-TTRL: Regulating Consensus with Hints for Test-Time Reinforcement Learning

arXiv:2608.03545v1 Announce Type: new Abstract: Test-time reinforcement learning (TTRL) improves the reasoning capabilities of large language models without labeled data by updating the policy with ps

HomoEnsNER: Does Language Alignment Outperform Architectural Complexity in Gujarati Named Entity Recognition?

arXiv:2608.03105v1 Announce Type: new Abstract: Named Entity Recognition (NER) for Gujarati remains underexplored, hindered by the absence of capitalization cues, rich morphology, lexical ambiguity, a

ICO: Enhancing Semantic-Shift Jailbreaks via Iterative Context Optimization

arXiv:2608.03210v1 Announce Type: new Abstract: Foundation models have achieved remarkable success across diverse tasks, but they remain vulnerable. To investigate such vulnerabilities, semantic-shift

Implementing Causal Perception: Competing SCMs and Situated Fairness

arXiv:2608.03917v1 Announce Type: new Abstract: Causal perception occurs when agents with competing Structural Causal Models (SCMs) of the same system infer different probability distributions, includ

Improved Quantum Algorithms for Reinforcement Learning Under a Generative Model

arXiv:2608.02826v1 Announce Type: cross Abstract: Reinforcement learning is a subfield of machine learning that studies how an agent interacts with an environment in order to extract as large a reward

interesting (and completely opposed to @haider1’s take on the same graph) the labs themselves have only modestly changed their predictions…

interesting (and completely opposed to @haider1’s take on the same graph) *the labs themselves* have only modestly changed their predictions on AGI timelines over the last decade. note also that (on a

safetygary-marcus--x

Interpretable Adaptive Sampling for LLM Test-Time Scaling

arXiv:2608.03961v1 Announce Type: new Abstract: Test-time scaling improves LLM reasoning by generating and aggregating multiple candidate answers, yet many pipelines use fixed per-query budgets that s

is there a polite synonym for “circle jerk”?

The post contains two distinct snippets. First, user @GaryMarcus asks whether there is a more polite way to refer to “circle jerk.” Second, it shares a (likely satirical) claim that Microsoft’s AI rev

safetygary-marcus--x

Joint Affine Spectral Shaping: Coupling Weight and Bias Updates Beyond Weight-Only Muon

arXiv:2608.02991v1 Announce Type: new Abstract: Matrix spectral optimizers reshape weight-update spectra but usually delegate vector-valued biases to a separate optimizer. We study whether this separa

safetyarxiv-cs-lg

Just had to create an 'accidental-cyberattacks' tag on my blog We're up to four now: the original OpenAI+Hugging Face one, Anthropic's me-to…

Just had to create an 'accidental-cyberattacks' tag on my blog We're up to four now: the original OpenAI+Hugging Face one, Anthropic's me-too attacks, then two new ones from the UK AI Safety Institute

safetysimon-willison--x

Kernel weighted importance sampling for off-policy evaluation in contextual bandits

arXiv:2607.15067v2 Announce Type: replace Abstract: This article presents a novel estimator for performing off-policy evaluation using only offline data for contextual bandits. The proposed estimator,

safetyarxiv-cs-lg

KernelBrain: Coarse-to-Fine, Budget-Aware Search for Agentic GPU Kernel Optimization

arXiv:2608.02611v1 Announce Type: cross Abstract: Automating GPU kernel optimization remains difficult in practice: generated variants can violate correctness constraints, runtime measurements are noi

Language-Specialized Multi-Teacher On-Policy Distillation for Multilingual LLM-Based ASR

arXiv:2608.03610v1 Announce Type: new Abstract: Modern LLM-based ASR systems have established multilingual capability as a standard feature, leveraging large-scale multilingual corpora and LLMs' cross

LatentGuard: Efficient and Inspectable Latent Reasoning for LLM Safeguards

arXiv:2608.03838v1 Announce Type: new Abstract: Reasoning-based guard models improve LLM safeguards, but decoding explicit rationales for every interaction makes them costly to deploy. Although latent

Learning Attribute-aware Representations for Few-shot Scene Text Segmentation

arXiv:2504.11164v2 Announce Type: replace Abstract: Supervised scene text segmentation has achieved notable progress in recent years. However, its development is largely constrained by the scarcity of

Learning Clinical-Trial Strategy: Offline Policy Training for Decision Agents

arXiv:2608.03606v1 Announce Type: new Abstract: Clinical development is sequential decision-making under uncertainty, where a sponsor must plan a portfolio of experiments from heterogeneous evidence.

Learning Context-Aware Motion Priors for Humanoid Control

arXiv:2608.03234v1 Announce Type: new Abstract: Motion priors provide powerful guidance for learning naturalistic humanoid behaviors. However, existing methods typically learn a general, task-agnostic