Search: “safety”

One of the things I hate the most about this site is the consistent lack of nuance. That’s why everything is an argument, and progress here …

Gary Marcus critiques social media platforms for lacking nuance in discourse, which he identifies as a root cause of persistent arguments and stalled progress on the site. The post reflects concerns a

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Online semi-supervised perception: Real-time learning without explicit feedback

DGX agent

arXiv:2604.27562v1 Announce Type: new Abstract: This paper proposes an algorithm for real-time learning without explicit feedback. The algorithm combines the ideas of semi-supervised learning on graph

safetyarxiv-cs-lg

1 May 2026

OpAgent: Operator Agent for Web Navigation

arXiv:2602.13559v2 Announce Type: replace Abstract: To fulfill user instructions, autonomous web agents must contend with the inherent complexity and volatile nature of real-world websites. Convention

Performance-Driven QUBO for Recommender Systems on Quantum Annealers

arXiv:2410.15272v3 Announce Type: replace-cross Abstract: Quantum annealers offer a promising hardware platform for solving combinatorial optimization problems, especially those formulated as Quadrati

Political Bias Audits of LLMs Capture Sycophancy to the Inferred Auditor

arXiv:2604.27633v1 Announce Type: new Abstract: Large language models (LLMs) are commonly evaluated for political bias based on their responses to fixed questionnaires, which typically place frontier

Preserving Temporal Dynamics in Time Series Generation

arXiv:2604.27182v1 Announce Type: cross Abstract: Time-series data augmentation plays a crucial role in regression-oriented forecasting tasks, where limited data restricts the performance of deep lear

Residual Gaussian Splatting for Ultra Sparse-View CBCT Reconstruction

arXiv:2604.27552v1 Announce Type: new Abstract: While 3D Gaussian splatting (3DGS) offers explicit and efficient scene representations for cone-beam computed tomography reconstruction, conventional ph

RHyVE: Competence-Aware Verification and Phase-Aware Deployment for LLM-Generated Reward Hypotheses

arXiv:2604.28056v1 Announce Type: new Abstract: Large language models (LLMs) make reward design in reinforcement learning substantially more scalable, but generated rewards are not automatically relia

Robot Learning from Human Videos: A Survey

arXiv:2604.27621v1 Announce Type: cross Abstract: A critical bottleneck hindering further advancement in embodied AI and robotics is the challenge of scaling robot data. To address this, the field of

Sam Altman is a master at insincerity, Author @GaryMarcus claims. 'He is a master of projecting insincerity, a master at telling the room wh…

Sam Altman is a master at insincerity, Author @GaryMarcus claims. 'He is a master of projecting insincerity, a master at telling the room what it wants to hear and not always truthful'. “Altman, sitti

Sample-efficient evidence estimation of score based priors for model selection

arXiv:2602.20549v2 Announce Type: replace-cross Abstract: The choice of prior is central to solving ill-posed imaging inverse problems, making it essential to select one consistent with the measuremen

SCOOP: A pro-AI dark money group backed by a powerful super PAC funded by execs tied to Palantir and OpenAI, has been secretly paying influe…

SCOOP: A pro-AI dark money group backed by a powerful super PAC funded by execs tied to Palantir and OpenAI, has been secretly paying influencers to push pro-AI, anti-China propaganda on TikTok and IG

Stable Behavior, Limited Variation: Persona Validity in LLM Agents for Urban Sentiment Perception

arXiv:2604.28048v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly used as proxies for human perception in urban analysis, yet it remains unclear whether persona prompting p

Stating that women do not have penises is conservative. Stating that the scientific method is superior to ancestral tribal dances for seekin…

Stating that women do not have penises is conservative. Stating that the scientific method is superior to ancestral tribal dances for seeking truth is conservative. Supporting a rational immigration p

safetyelon-musk--x

Supercharging Agenda Setting Research: The ParlaCAP Dataset of 28 European Parliaments and a Scalable Multilingual LLM-Based Classification

arXiv:2602.16516v2 Announce Type: replace Abstract: This paper introduces ParlaCAP, a large-scale dataset for analyzing parliamentary agenda setting across Europe, and proposes a cost-effective method

Taxon: Hierarchical Tax Code Prediction with Semantically Aligned LLM Expert Guidance

arXiv:2601.08418v2 Announce Type: replace-cross Abstract: Tax code prediction is a crucial yet underexplored task in automating invoicing and compliance management for large-scale e-commerce platforms

Test-Time Distillation for Continual Model Adaptation

arXiv:2506.02671v3 Announce Type: replace Abstract: Deep neural networks often suffer performance degradation upon deployment due to distribution shifts. Continual Test-Time Adaptation (CTTA) aims to

The Likelihood Ratio Wall: Structural Limits on Accurate Risk Assessment for Rare Violence

arXiv:2604.27282v1 Announce Type: cross Abstract: Pretrial risk assessment tools are used on over one million U.S. defendants each year, yet their use for predicting rare violent re-offense faces a ba

safetyarxiv-cs-lg

The Two Boundaries: Why Behavioral AI Governance Fails Structurally

arXiv:2604.27292v1 Announce Type: new Abstract: Every system that performs effects has two boundaries: what it can do (expressiveness) and what governance covers (governance). In nearly all deployed A

Tokenmaxxing is stupid. Change my mind?

Gary Marcus critiques the AI industry's focus on scaling model parameters and training data (tokenmaxxing) as an inefficient approach to advancing AI capabilities. He argues that simply increasing tok

TouchGuide: Inference-Time Steering of Visuomotor Policies via Touch Guidance

arXiv:2601.20239v4 Announce Type: replace Abstract: Fine-grained and contact-rich manipulation remain challenging for robots, largely due to the underutilization of tactile feedback. To address this,

safetyarxiv-cs-ro

Vision-Language Models Mistake Head Orientation for Gaze Direction: Nonverbal Conversation Cues

arXiv:2506.05412v3 Announce Type: replace-cross Abstract: Where someone looks is a nonverbal communication cue that children and adults readily use. How well can Vision-Language Models (VLMs) infer ga

When Does Structure Matter in Continual Learning? Dimensionality Controls When Modularity Shapes Representational Geometry

arXiv:2604.27656v1 Announce Type: cross Abstract: To preserve previously learned representations, continual learning systems must strike a balance between plasticity, the ability to acquire new knowle

who's ready to dance? 😆🕺🎶 uploaded 11 new favorite @suno songs to @spotify [search: 'captain yohei'] 💿 Captain Yohei Vol 1. 💿 ♫ Don't G…

who's ready to dance? 😆🕺🎶 uploaded 11 new favorite @suno songs to @spotify [search: 'captain yohei'] 💿 Captain Yohei Vol 1. 💿 ♫ Don't Go in the Office [hard rock] ♫ Nine-Nine-Six [hip-hop/rap] ♫ Fog F

safetyyohei-nakajima--x

A Multimodal Pre-trained Network for Integrated EEG-Video Seizure Detection

arXiv:2604.26379v1 Announce Type: new Abstract: Reliable seizure detection in mouse models is essential for preclinical epilepsy research, yet manual review of synchronized video-EEG recordings is lab

A Survey of Process Reward Models: From Outcome Signals to Process Supervisions for Large Language Models

arXiv:2510.08049v3 Announce Type: replace-cross Abstract: Although Large Language Models (LLMs) exhibit advanced reasoning ability, conventional alignment remains largely dominated by outcome reward m

Accelerating RL Post-Training Rollouts via System-Integrated Speculative Decoding

arXiv:2604.26779v1 Announce Type: cross Abstract: RL post-training of frontier language models is increasingly bottlenecked by autoregressive rollout generation, making rollout acceleration a central

alignment in 2016: obviously any real AI will be made inside a faraday cage magnetically suspended in a 10×10×10 cube of telekill alloy alig…

alignment in 2016: obviously any real AI will be made inside a faraday cage magnetically suspended in a 10×10×10 cube of telekill alloy alignment in 2026: yeah we can not make it stop talking about go

safetyconnor-leahy--x

ATLAS: An Annotation Tool for Long-horizon Robotic Action Segmentation

arXiv:2604.26637v1 Announce Type: cross Abstract: Annotating long-horizon robotic demonstrations with precise temporal action boundaries is crucial for training and evaluating action segmentation and

Atomic-Probe Governance for Skill Updates in Compositional Robot Policies

arXiv:2604.26689v1 Announce Type: cross Abstract: Skill libraries in deployed robotic systems are continually updated through fine-tuning, fresh demonstrations, or domain adaptation, yet existing type

Attribution-Guided Multimodal Deepfake Detection via Cross-Modal Forensic Fingerprints

arXiv:2604.26453v1 Announce Type: new Abstract: Audio-visual deepfakes have reached a level of realism that makes perceptual detection unreliable, threatening media integrity and biometric security. W

Beyond Shortcuts: Mitigating Visual Illusions in Frozen VLMs via Qualitative Reasoning

arXiv:2604.26250v1 Announce Type: new Abstract: While Vision-Language Models (VLMs) have achieved state-of-the-art performance in general visual tasks, their perceptual robustness remains remarkably b

Big Tech’s $700 billion spending on AI this year is called the ‘greatest capital misallocation in history’ https://trib.al/qUQIbrJ

Big Tech companies are projected to spend approximately $700 billion on AI infrastructure and development in the current year, a figure that AI researcher and entrepreneur Gary Marcus has criticized a

Classification of Public Opinion on the Free Nutritional Meal Program on YouTube Media Using the LSTM Method

arXiv:2604.26312v1 Announce Type: new Abstract: Public opinion towards the Free Nutritious Meal Program (MBG) on YouTube social media reflects diverse community responses. This study applies the Long

Co-Learning Port-Hamiltonian Systems and Optimal Energy-Shaping Control

arXiv:2604.26172v1 Announce Type: cross Abstract: We develop a physics-informed learning framework for energy-shaping control of port-Hamiltonian (pH) systems from trajectory data. The proposed approa

Consist-Retinex: One-Step Noise-Emphasized Consistency Training Accelerates High-Quality Retinex Enhancement

arXiv:2512.08982v2 Announce Type: replace-cross Abstract: Retinex-based low-light image enhancement benefits from separating reflectance and illumination, yet recent generative approaches often rely o

Correcting Performance Estimation Bias in Imbalanced Classification with Minority Subconcepts

arXiv:2604.26024v1 Announce Type: cross Abstract: Class-level evaluation can conceal substantial performance disparities across subconcepts within the same class, causing models that perform well on a

Data-Centric Foundation Models in Computational Healthcare: A Survey

arXiv:2401.02458v3 Announce Type: replace-cross Abstract: The advent of foundation models (FMs) as an emerging suite of AI techniques has struck a wave of opportunities in computational healthcare. Th

DC-Ada: Reward-Only Decentralized Sensor Adaptation for Heterogeneous Multi-Robot Teams

arXiv:2604.03905v2 Announce Type: replace-cross Abstract: Heterogeneity is a defining feature of deployed multi-robot teams: platforms often differ in sensing modalities, ranges, fields of view, and f

Delta Score Matters! Spatial Adaptive Multi Guidance in Diffusion Models

arXiv:2604.26503v1 Announce Type: new Abstract: Diffusion models have achieved remarkable success in synthesizing complex static and temporal visuals, a breakthrough largely driven by Classifier-Free

DORA: A Scalable Asynchronous Reinforcement Learning System for Language Model Training

arXiv:2604.26256v1 Announce Type: new Abstract: Reinforcement learning (RL) has become a critical paradigm for LLM post-training, yet the rollout phase -- accounting for 50--80% of total step time --

safetyarxiv-cs-lg

Efficient and Interpretable Transformer for Counterfactual Fairness

arXiv:2604.26188v1 Announce Type: new Abstract: The growing reliance of machine learning models in high-stakes, highly regulated domains such as finance and insurance has created a growing tension bet

safetyarxiv-cs-lg

Evaluating Strategic Reasoning in Forecasting Agents

arXiv:2604.26106v1 Announce Type: new Abstract: Forecasting benchmarks produce accuracy leaderboards but little insight into why some forecasters are more accurate than others. We introduce Bench to t

Evaluating the Alignment Between GeoAI Explanations and Domain Knowledge in Satellite-Based Flood Mapping

arXiv:2604.26051v1 Announce Type: cross Abstract: The increasing number of satellites has improved the temporal resolution of Earth observation, making satellite-based flood mapping a promising approa

EvoSelect: Data-Efficient LLM Evolution for Targeted Task Adaptation

arXiv:2604.26170v1 Announce Type: new Abstract: Adapting large language models (LLMs) to a targeted task efficiently and effectively remains a fundamental challenge. Such adaptation often requires ite

FedPF: Accurate Target Privacy Preserving Federated Learning Balancing Fairness and Utility

arXiv:2510.26841v2 Announce Type: replace-cross Abstract: Federated Learning (FL) enables collaborative model training without data sharing, yet participants face a fundamental challenge, e.g., simult

FLARE: Fully Integration of Vision-Language Representations for Deep Cross-Modal Understanding

arXiv:2504.09925v3 Announce Type: replace Abstract: We introduce FLARE, a family of vision language models (VLMs) with a fully vision-language alignment and integration paradigm. Unlike existing appro

For better or worse, regulation for closed-source models served by a few (quite large) companies is easy. It is not as easy to imagine how y…

For better or worse, regulation for closed-source models served by a few (quite large) companies is easy. It is not as easy to imagine how you regulate open-source models that can be served by a range

safetyethan-mollick--x