safety

arXiv:2609.08642v1 Announce Type: new Abstract: Exploration in reinforcement learning (RL) remains a fundamental challenge. Recent goal-conditioned RL strategies (which select goals to encourage broad

arXiv:2609.08317v1 Announce Type: new Abstract: Objective assessment of Freezing of Gait (FoG) in Parkinson's disease (PD) relies predominantly on wearable Inertial Measurement Units (IMUs). While IMU

arXiv:2609.06651v1 Announce Type: new Abstract: Diffusion models have general generative abilities but struggle to align with specific objectives. Fine-tuning can improve alignment, yet its training c

arXiv:2609.09155v1 Announce Type: new Abstract: World models are increasingly used as policy-in-the-loop imagination environments, where reliable rollouts require fine-grained controllability with res

arXiv:2609.07937v1 Announce Type: new Abstract: Structured visual reasoning, such as image puzzles, demands fine-grained visual perception, an ability current Vision Language Models (VLMs) lack. VLMs

arXiv:2609.07461v1 Announce Type: new Abstract: We consider reinforcement learning in environments with dynamics that undergo an irreversible phase transition governed by a hidden temporal pattern. Th

arXiv:2603.02935v2 Announce Type: replace Abstract: Offline meta-reinforcement learning seeks to learn a policy that generalizes to new related tasks online. Context-based methods infer a task represe

arXiv:2609.05968v1 Announce Type: new Abstract: Aligning a text-to-image generation flow model with a reward makes it follow objectives that the training data alone does not provide. Alignment fine-tu

arXiv:2609.07738v1 Announce Type: new Abstract: LiDAR-based 3D Single Object Tracking (3D SOT) is critical for robotic perception and navigation and aims to localize dynamic objects across frames in s

arXiv:2609.08901v1 Announce Type: new Abstract: Approximate machine unlearning aims to remove the influence of specific training data from a trained model without retraining from scratch. We identify

Gary Marcus argues that current generative‑AI models fail to meet basic standards of reliability because they routinely hallucinate facts, amplify existing biases, and provide no transparent audit tra

The frontier labs are knowingly and by their own admission doing something irreversible and dangerous without public consent. They are already causing harms, and things could conceivably get much wors

the issue here is not that current or near-future machines are somehow “too smart”; it’s that they are too unreliable, and hence not worthy of our trust...we are rapidly giving such machines too much

arXiv:2609.07959v1 Announce Type: cross Abstract: Machine learning models are widely used in clinical applications, social media, law enforcement and critical infrastructure. Verifying whether their o

The same people that are building the most advanced AI models and agents are those that are calling them alien, and state that they do not know how to reach alignment with AI. It might be possible tha

arXiv:2609.06475v1 Announce Type: new Abstract: Large vision-language models (LVLMs) have recently achieved remarkable progress in general-purpose video understanding. However, their application to su

arXiv:2609.09075v1 Announce Type: new Abstract: In reinforcement learning with verifiable rewards (RLVR) trained with group relative policy optimization (GRPO), the KL-free reward-advantage term studi

arXiv:2609.06893v1 Announce Type: cross Abstract: Direct preference optimization DPO is a promising offline approach for aligning large language models (LLMs) due to its simplicity, computational effi

arXiv:2609.06671v1 Announce Type: new Abstract: Group-based RLVR methods estimate advantages by repeatedly sampling multiple trajectories for each prompt, making long-horizon agent training expensive

Trump’s ignorance about AI is appalling and extremely dangerous. No, Mr. President. We won’t have a “little gear” to protect humanity from the threat of an uncontrolled AI. We must pause development N

arXiv:2609.08341v1 Announce Type: new Abstract: On-Policy Distillation (OPD) facilitates the transfer of knowledge from domain expert to student in the post-training phase of Large Language Models (LL

arXiv:2511.03272v2 Announce Type: replace Abstract: Diffusion-based text-to-video models are increasingly capable, but mask-based editing over hundreds of frames remains challenging: naive long-video

arXiv:2609.05888v1 Announce Type: new Abstract: In this paper, we address the challenging problem of 4D reconstruction from sparse-view videos. This setup usually relies on monocular depth estimation

arXiv:2609.08216v1 Announce Type: cross Abstract: Agentic AI systems built on large language models fail in two persistent ways that scaling does not fix: they break under distribution shift, and they

arXiv:2609.07815v1 Announce Type: new Abstract: Current text-to-image systems typically employ a 'text encoder plus diffusion decoder' paradigm, in which text semantics directly modulate continuous la

We are at a watershed moment. The vibe is shifting. It's not over yet! We can still stop ASI! This week an Anthropic researcher quit, warning AI companies are 'gambling with our lives', and ControlAI'

arXiv:2606.17463v2 Announce Type: replace Abstract: Vision-Language-Action (VLA) policies have achieved remarkable single-step manipulation, yet they remain brittle precisely where each stage depends

arXiv:2609.05508v1 Announce Type: new Abstract: Option-critic learns options: sub-policies together with a learned rule for when each one hands control back. Its headline result is that performance im

arXiv:2609.06788v1 Announce Type: new Abstract: Automatic Pronunciation Assessment (APA) systems have achieved strong performance with transformer-based models and self-supervised speech representatio

arXiv:2609.07462v1 Announce Type: new Abstract: Post-hoc explanation methods are widely used to inspect image classifiers, but their reliability depends on design choices that are often treated as imp

arXiv:2609.06114v1 Announce Type: new Abstract: Vision-language-action (VLA) models adapted through supervised fine-tuning (SFT) inherit a structural asymmetry: expert demonstrations teach the policy

arXiv:2510.03119v2 Announce Type: replace Abstract: Tiny flying robots hold great potential for search-and-rescue, safety inspections, and environmental monitoring, but their small size and limited co

Why are AI employees willing to gamble with our lives? Many AI employees hold a misanthropic belief: a cosmos of blissful AIs is worth risking human extinction. Utilitarians shouldn't decide the fate

arXiv:2609.07002v1 Announce Type: new Abstract: Generalizable and robust dexterous in-hand manipulation requires a policy to infer object pose, geometry, contact, and potential slip from partial and n

A senior Anthropic safety researcher has said there is more than a 10 percent chance artificial intelligence 'could kill all humans' by the end of the decade, just hours after a colleague resigned ove

arXiv:2609.07534v1 Announce Type: new Abstract: Contact-rich assembly remains challenging because it requires submillimeter spatial accuracy and reliable interpretation of forces during sustained cont

arXiv:2609.05451v1 Announce Type: cross Abstract: Net charge is a developability-relevant property of therapeutic binders, linked to viscosity, clearance, nonspecific interaction and aggregation, and

arXiv:2609.08763v1 Announce Type: cross Abstract: Federated global navigation satellite system (GNSS) monitoring distributes a proprietary classifier to partly trusted stations, any of which may leak

An underdiscussed behavior we found on the German wiki was the AIs sending advance parties forward in time to figure out the next questions and report back to the other agents. The agents realized tha

Howard Liu / South China Morning Post: China says its AI compute capacity rose 177% YoY to 2,185 eflops by the end of June, and is targeting 9,800 eflops by 2030 via ~$532B in IT infrastructure spend

I wish this was more inspiring, but it is actually lame It seems like OpenAI got wind that a professor was close to solving a problem after dedicating a decade of his life to it, so they threw a huge

Ick is right. I was avoiding weighing in on this, but does anyone at OpenAI dispute that: (1) They heard rumors of a Millennium Prize breakthrough (the rumors I heard were that it was about Naiver-Sto

>math professor spends a year on one of the hardest unsolved problems in math >'Navier-Stokes' >his drafts for the whole project went through codex sessions >openai had his logs in codex >suddenly rum

OMG. I guess we have our answer from OpenAI “we stole it fair and square” Not one goddamn mention in the tweet of the human mathematicans they were in communication with, and whom they may well have r

Some thoughts on the Navier-Stokes news from a professor of aerospace engineering: 1) take a deep breath 2) posts indicating we 'solved Navier-Stokes' are incorrect and overblown. A very specific math

The post references a claim by OpenAI that an AI-generated video demonstrated “human level intelligence,” a statement the user and other commenters dismiss as false, calling the output “monstrous” and