Search: “pretraining”

Isotropic Embedding Perturbations for Robust Vision Language Encoders

arXiv:2609.10292v1 Announce Type: new Abstract: Data augmentation is fundamental to training modern deep vision and multimodal models. While individual methods, such as RandAug, CutMix, Mixup, RandEra

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Jacob was a senior researcher who joined Anthropic in May. My Anthropic colleagues and I had been trying to recruit him for ~2 years, becaus…

DGX agent

Jacob was a senior researcher who joined Anthropic in May. My Anthropic colleagues and I had been trying to recruit him for ~2 years, because we knew he was a strong researcher at OpenAI. Before he le

safetygary-marcus--x

10 Sept 2026

Learning to Adapt and Calibrate: Score Distribution Alignment for Few-Shot Uncertainty Prediction in Medical VLMs

arXiv:2609.10333v1 Announce Type: new Abstract: Uncertainty estimation for medical vision--language models (VLMs) using conformal prediction has gained increasing attention due to its distribution-fre

Learning with Synthetic Data via SGD in High-Dimensional Linear Regression

arXiv:2609.09572v1 Announce Type: cross Abstract: Synthetic data has become a promising way to scale model training beyond limited human-generated data but it may also induce strong model collapse (Do

safetyarxiv-cs-lg

No Free Checker: A Survey of Verifiers for Robot Policies

arXiv:2609.09250v1 Announce Type: cross Abstract: A verifier for robot policies reads a candidate behavior and returns a score for how well it did, used both to evaluate vision-language-action policie

Noise Adaptive Streaming Audio-Visual Speech Token Enhancement for Robust Full-Duplex Spoken Dialogue Models

arXiv:2609.08390v1 Announce Type: cross Abstract: Full-duplex spoken dialogue systems enable simultaneous listening and speaking, but their audio-only perception often fails under background noise and

researcharxiv-cs-ai

Programmable World Model

arXiv:2609.10540v1 Announce Type: new Abstract: Recent video world models generate increasingly realistic and interactive visual experiences, yet lack reliable mechanisms for maintaining persistent wo

Local Ai

Show-Harness: Just a VLM Agent Can Play Robots

arXiv:2609.10522v1 Announce Type: cross Abstract: Foundation vision-language models (VLMs) exhibit broad intelligence about the world, yet translating this intelligence into robot control remains chal

local-aiarxiv-cs-cv

StreamAlign: Streaming Text-Aligned Speech Tokenization

arXiv:2609.09719v1 Announce Type: new Abstract: Text-aligned speech tokenization methods have emerged to better align speech tokens with LLM token spaces, enabling more effective utilization of pretra

safetyarxiv-cs-cl

SynThermFace: Amplifying Limited Paired Data for Visible-Thermal Face Recognition via Synthetic Data Generation

arXiv:2609.10303v1 Announce Type: new Abstract: Face recognition (FR) is a widely used modality for biometric authentication, but conventional models rely on visible-spectrum imagery and degrade when

Test-time Prompt Refinement for Text-to-Image Models

arXiv:2507.22076v2 Announce Type: replace Abstract: Text-to-image (T2I) generation models have made significant strides but still struggle with prompt sensitivity: even minor changes in prompt wording

Time-Frequency Geometric Cross-Attention for Chunked Vision-Language-Action Models

arXiv:2609.09925v1 Announce Type: cross Abstract: Modern vision-language-action (VLA) policies predict a whole chunk of actions: one to two seconds of coordinated motion emitted in a single forward pa

safetyarxiv-cs-ro

Why Do LLM Agents Fail in Exploring New Environments? A World-Modeling Perspective

arXiv:2510.15047v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) as agents often fail to improve in new environments. We identify and characterize a failure mode we call explorat

safetyarxiv-cs-cl

Agents

WorldAgen: Unified State-Action Prediction with Test-Time World Model Training

arXiv:2609.08162v1 Announce Type: new Abstract: How can vision-language-action (VLA) models adapt to new environments where world dynamics shift? While recent research has combined world modeling and

agentsarxiv-cs-ai

Adapting Vision Foundation Models to Acoustics for Pose-Free 3D Sonar Reconstruction

arXiv:2609.06261v1 Announce Type: new Abstract: Vision foundation models trained on Internet-scale RGB datasets enable remarkable capabilities across a range of tasks, from text-to-video generation to

Anyone still denying the risks posed by unregulated AI should read this thread. It’s time to pump the brakes & take action NOW. We need: - M…

Anyone still denying the risks posed by unregulated AI should read this thread. It’s time to pump the brakes & take action NOW. We need: - Mandatory safeguards - A comprehensive testing regime - Urgen

safetygary-marcus--x

ARC-Bench: Closed-Loop Replanning Masks Broken Action Ranking in Frozen JEPA World Models

arXiv:2609.05461v1 Announce Type: cross Abstract: Reward-free latent world models plan by scoring candidate actions with distances in a frozen latent space: an action is preferred if its predicted fut

Are Image Generators Zero-Shot Perceivers? A Rigorous Evaluation

arXiv:2609.07884v1 Announce Type: new Abstract: Recent work, such as Vision Banana, shows that lightweight instruction tuning can enable an image generator to achieve state-of-the-art performance acro

Assessing Covariate-Informed Grid Load Forecasting with a Time-Series Foundation Model

arXiv:2609.06656v1 Announce Type: new Abstract: Modern power systems are growing increasingly complex as they integrate diverse generation sources to meet rising demand, making accurate load forecasti

Local Ai

AstraMoE-SR: Trajectory-Guided Diffusion for Blind Satellite Jitter Deblurring and Super-Resolution

arXiv:2609.07012v1 Announce Type: cross Abstract: Pushbroom satellite imaging couples limited spatial resolution with platform attitude instability. Platform jitter produces spatially varying motion b

local-aiarxiv-cs-cv

Back to the Feature: Zero-Shot 6DoF Pose Estimation via Dense Local Features

arXiv:2609.06726v1 Announce Type: new Abstract: We present B2TFPose, a training-free zero-shot method for 6DoF pose estimation of unseen objects from RGB images. Using a single frozen DINOv3 vision tr

ChildGaze: A Benchmark Dataset for Collaborative Behavior Understanding in Children

arXiv:2609.06353v1 Announce Type: new Abstract: Understanding collaborative behavior in children is important for analyzing social participation, peer interaction, shared attention, and engagement dur

Clinician-Friendly Foundation Models for Ophthalmic Image Diagnostics without Fine-Tuning or Technical Barriers

arXiv:2504.15928v3 Announce Type: replace Abstract: Artificial intelligence (AI) shows remarkable potential in medical imaging diagnostics, yet most current models require retraining when applied acro

Cross-modal learning for SAR target recognition using optical vision foundation models

arXiv:2609.07753v1 Announce Type: new Abstract: Synthetic Aperture Radar (SAR) is an important modality in a wide range of imaging applications due to its versatile, long range and near all weather op

CVT-GS: Learning to Simplify 3D Gaussian Splatting with Centroidal Voronoi Tessellation

arXiv:2609.08730v1 Announce Type: new Abstract: While 3D Gaussian Splatting (3DGS) has emerged as a powerful representation for real-time novel view synthesis, rendering high-fidelity scenes often rel

Diagnosing and Dynamically Filtering Occupancy World Models for Active Mapping

arXiv:2609.06820v1 Announce Type: cross Abstract: Active mapping requires a robot to select camera viewpoints that efficiently reconstruct an unknown 3D scene. To reason about unobserved regions, rece

DriveZero: End-to-End Driving Beyond Human Demonstrations

arXiv:2609.06055v1 Announce Type: new Abstract: Most end-to-end autonomous-driving systems learn by imitating human driving logs, leaving their learned behavior constrained by the quality and behavior

DrugReason: Dynamic Multi-View Reasoning over Knowledge Graph and Language Evidence for Drug Repurposing

arXiv:2609.06779v1 Announce Type: new Abstract: Drug repurposing aims to identify new therapeutic uses for existing compounds and, compared with de novo drug discovery, offers a faster and more cost-e

researcharxiv-cs-lg

DSE-VTG: Dual-Side Enhancement for Training-Free Video Temporal Grounding

arXiv:2609.08850v1 Announce Type: new Abstract: Text-guided Video Temporal Grounding (VTG) aims to localize the relevant segments in an untrimmed video based on text queries, yet collecting dense temp

Tutorials

Earth System World Model for What-If Simulations: A Case Study for Terrestrial Ecosystems

arXiv:2609.08855v1 Announce Type: new Abstract: Machine learning emulators have become essential for accelerating expensive Earth-system simulations, but most existing approaches remain passive foreca

tutorialsarxiv-cs-lg

Local Ai

Efficient and Microphone-Fault-Tolerant 3D Sound Source Localization

arXiv:2505.20961v2 Announce Type: replace-cross Abstract: Sound source localization (SSL) is a critical technology for determining the position of sound sources in complex environments. However, exist

local-aiarxiv-cs-lg

Applications

EquiGQNet: Fast Grasp Quality Evaluation via Shared Equivariant Point Cloud Encoding

arXiv:2609.07145v1 Announce Type: new Abstract: Planning six-degree-of-freedom (6-DoF) grasps for unseen objects in cluttered tabletop scenes from a single-view depth image requires accurate and effic

applicationsarxiv-cs-ro

Flow3D-OPD: Multi-Teacher On-Policy Distillation for 3D Geometry Generation with Flow-Matching Diffusion Transformer

arXiv:2609.07137v1 Announce Type: new Abstract: Recent image-to-3D generation models built on flow-matching diffusion Transformers (DiT) can produce high-fidelity meshes, yet their post-training strat

Tutorials

Foundation Models for Generalizable Semantic and Goal-Oriented Communication

arXiv:2609.07853v1 Announce Type: new Abstract: Semantic and goal-oriented communication is increasingly studied for 6G, but generalization beyond seen data remains a key weakness under tight rate bud

tutorialsarxiv-cs-lg

Agents

From Gaze to Meaning: A Training-Free AI Agent for Unified Grounding and Explanation

arXiv:2609.06208v1 Announce Type: new Abstract: Understanding human attention is fundamental for scene interpretation, yet existing approaches often rely on heavily trained models that lack interpreta

agentsarxiv-cs-cv

Generalist Open-World Temporal Perception

arXiv:2609.06823v1 Announce Type: new Abstract: The next generation of artificial intelligence systems will likely be natively temporal and multimodal in both inputs and outputs: able to converse, per

GRAN-TED: Generating Robust, Aligned, and Nuanced Text Embedding for Diffusion Models

arXiv:2512.15560v3 Announce Type: replace Abstract: The text encoder is a critical component of text-to-image and text-to-video diffusion models, fundamentally determining the semantic fidelity of the

Harnessing CLIP and DINO: An Uncertainty-Aware Cascaded Fusion Network for Generalizable Deepfake Image Detection

arXiv:2609.07670v1 Announce Type: new Abstract: The growing realism and accessibility of manipulated and generated faces threaten the trustworthiness of digital media. To detect such forgeries, deepfa

Hidden in Plain Sight: The Overlooked Significance of Canonical Elements for Extreme LLM Sparsity

arXiv:2609.06557v1 Announce Type: new Abstract: Large language models (LLMs) are often considered fragile under aggressive sparsification, and maintaining reliable performance typically requires stick

I dropped out of a PhD in AI safety partially for the opposite reason. I didn’t believe AI existential risk was as important as the doomers …

I dropped out of a PhD in AI safety partially for the opposite reason. I didn’t believe AI existential risk was as important as the doomers think. My biggest pet peeve is that no one can really provid

safetygary-marcus--x

I worked at Google DeepMind and now at Anthropic. This is a common sentiment amongst my peers. (I write this in personal capacity.) There is…

I worked at Google DeepMind and now at Anthropic. This is a common sentiment amongst my peers. (I write this in personal capacity.) There is not yet a viable scientific plan to solve risks from recurs

safetygary-marcus--x

Image-Scale Robustness and Visual Recognition Performance: A Cross-Architecture Analysis

arXiv:2609.06051v1 Announce Type: new Abstract: The sensitivity of visual recognition models to changes in image scale is well established, yet the factors governing this sensitivity across heterogene

Inclusive electron-nucleus cross section models from domain adaptation

arXiv:2609.08463v1 Announce Type: cross Abstract: We apply transfer learning (TL) to construct data-driven models of inclusive electron-nucleus cross sections. Starting from an ensemble of deep neural

researcharxiv-cs-lg

Infrastructure-based Monocular 3D Vehicle Localization Framework with Experimental Validation

arXiv:2609.05523v1 Announce Type: new Abstract: This paper presents a one-stage learning framework that maps monocular roadside-camera images directly to vehicle states in a ground-fixed coordinate fr

JEDI: JEPA-to-Edge Distillation for Efficient Cropland Segmentation from Satellite Imagery

arXiv:2609.07915v1 Announce Type: new Abstract: Large vision models provide useful representations for remote-sensing segmentation but are often too expensive for deployment at the satellite or field

Hardware

Kalman Delta Networks: Uncertainty-aware Associative Memory

arXiv:2609.07816v1 Announce Type: new Abstract: Linear attention is increasingly used in frontier language models for efficient long-context inference and constant-memory decoding. Its fixed-size recu

hardwarearxiv-cs-lg

Learn to Rank: Visual Attribution by Learning Importance Ranking

arXiv:2604.05819v2 Announce Type: replace Abstract: Interpreting the decisions of complex computer vision models is crucial to establish trust and accountability, especially in safety-critical domains

LLM-Aided Design for Manufacturing: A Multi-Agent System for Intent-Preserving Redesign of CAD for Improved Manufacturability

arXiv:2609.05559v1 Announce Type: cross Abstract: We introduce autonomous, intent-preserving Design for Manufacturing (DFM) redesign of CAD parts: given an engineer's CAD model, the method returns a v