Search: “models”

DUET: Optimize Token-Budget Allocation for Reinforcement Learning with Verifiable Rewards

arXiv:2605.08441v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) generates hundreds of thousands of tokens per training step, with rollout generation dominating

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

DuetFair: Coupling Inter- and Intra-Subgroup Robustness for Fair Medical Image Segmentation

DGX agent

arXiv:2605.10521v1 Announce Type: cross Abstract: Medical image segmentation models can perform unevenly across subgroups. Most existing fairness methods focus on improving average subgroup performanc

safetyarxiv-cs-ai

12 May 2026

DynaMiCS: Fine-tuning LLMs with Performance Constraints using Dynamic Mixtures

arXiv:2605.10770v1 Announce Type: new Abstract: Multi-domain fine-tuning of large language models requires improving performance on target domains while preserving performance on constrained domains,

safetyarxiv-cs-lg

EchoPrune: Interpreting Redundancy as Temporal Echoes for Efficient VideoLLMs

arXiv:2605.10050v1 Announce Type: new Abstract: Long-form video understanding remains challenging for Video Large Language Models (VideoLLMs), as the dense frame sampling introduces massive visual tok

researcharxiv-cs-cv

EmbodiSkill: Skill-Aware Reflection for Self-Evolving Embodied Agents

arXiv:2605.10332v1 Announce Type: new Abstract: Embodied agents can benefit from skills that guide object search, action execution, and state changes across diverse environments. Since embodied enviro

Enabling Structure-Only Initialization and Out-of-Distribution Generalization in GNN-based Molecular Dynamics Simulators

arXiv:2605.09495v1 Announce Type: cross Abstract: Machine learning-based simulators offer the potential to model the dynamics of complex systems more efficiently than classical approaches, while retai

researcharxiv-cs-lg

EpiGraph: A Knowledge Graph and Benchmark for Evidence-Intensive Reasoning in Epilepsy

arXiv:2605.09505v1 Announce Type: new Abstract: Epilepsy diagnosis and treatment require evidence-intensive reasoning across heterogeneous clinical knowledge, including biosignal patterns, genetic mec

Exact Unlearning from Proxies Induces Closeness Guarantees on Approximate Unlearning

arXiv:2605.10680v1 Announce Type: new Abstract: This paper proposes a paradigm shift linking machine unlearning directly to the structure of the data distributions rather than a mere update of the neu

researcharxiv-cs-lg

Execution Envelopes: A Shared Admission Contract for Backend AI Execution Requests

arXiv:2605.08267v1 Announce Type: cross Abstract: Enterprise AI backends increasingly admit heterogeneous execution requests across model deployment, inference, evaluation, data movement, and agentic

safetyarxiv-cs-ai

Local Ai

Explainability of Recurrent Neural Networks for Enhancing P300-based Brain-Computer Interfaces

arXiv:2605.10121v1 Announce Type: cross Abstract: Brain-Computer Interfaces (BCIs) based on P300 event-related potentials offer promising applications in health, education, and assistive technologies.

local-aiarxiv-cs-ai

Explanation-Aware Learning for Enhanced Interpretability in Biomedical Imaging

arXiv:2605.10054v1 Announce Type: new Abstract: Deep neural networks for medical image diagnosis often achieve high predictive accuracy while relying on spurious or clinically irrelevant visual cues,

safetyarxiv-cs-cv

Extending Confidence-Based Text2Cypher with Grammar and Schema Aware Filtering

arXiv:2605.10318v1 Announce Type: new Abstract: Large language models (LLMs) allow users to query databases using natural language by translating questions into executable queries. Despite strong prog

researcharxiv-cs-cl

Falcon 9 launches NROL-172 to orbit from pad 4E in California

model-releaseselon-musk--x

SpaceX's Falcon 9 rocket successfully launched the NROL-172 classified national reconnaissance payload from Space Launch Complex 4E at Vandenberg Space Force Base in California. This mission was condu

Fast mode for Claude Opus 4.7 is now available in Cursor! It's 2.5x the speed at 6x the cost. For most tasks, we recommend using the standar…

Cursor has released fast mode for Claude Opus 4.7, offering 2.5x faster processing speeds but at 6x the cost compared to standard mode. The announcement suggests that for most tasks, standard mode rem

model-releasescursor--x

Tools

Fast mode for Opus 4.7 available on AI Gateway

Vercel has introduced a fast mode option for Claude Opus 4.7 on its AI Gateway service, enabling faster response times for users integrating the model through their platform. This feature likely provi

toolsvercel-blog

Feature Repulsion and Spectral Lock-in: An Empirical Study of Two-Layer Network Grokking

arXiv:2605.08119v1 Announce Type: cross Abstract: Tian (2025) proves a repulsion theorem (Theorem 6) for the matrix B = (widetilde{F}^op widetilde{F} + eta I)^{-1} during the interactive feature-learn

Fetal Brain Imaging: A Composite Neural Network Approach for Keyframe Detection in Ultrasound Videos

arXiv:2605.09750v1 Announce Type: new Abstract: This article presents a novel approach to keyframe detection in ultrasound videos, with a particular focus on fetal brain imaging. The proposed model is

researcharxiv-cs-cv

Few-Click-Driven Interactive 3D Segmentation with Semantic Embedding

arXiv:2605.08925v1 Announce Type: new Abstract: Interactive segmentation allows efficient label generation by leveraging user-provided clicks to progressively refine predictions, which is critical whe

safetyarxiv-cs-cv

FinMoji: A Framework for Emoji-driven Sentiment Analysis in Financial Social Media

arXiv:2605.09469v1 Announce Type: new Abstract: This paper explores the use of emojis in financial sentiment analysis, focusing on the social media platform StockTwits. Emojis, increasingly prevalent

researcharxiv-cs-cl

Fitting Multilinear Polynomials for Logic Gate Networks

arXiv:2605.08657v1 Announce Type: cross Abstract: We study learnable logic gate networks that stack layers of 2-input Boolean gates to build combinational circuits. Every 2-input gate has a unique mul

Fix the Loss, Not the Radius: Rethinking the Adversarial Perturbation of Sharpness-Aware Minimization

model-releasesarxiv-cs-lg

arXiv:2605.10183v1 Announce Type: new Abstract: Sharpness-Aware Minimization (SAM) improves generalization by minimizing the worst-case loss within a fixed parameter-space radius neighborhood. SAM and

FPGA-Based Hardware Architecture for Contrast Maximization in Event-Based Vision

model-releasesarxiv-cs-cv

arXiv:2605.09581v1 Announce Type: new Abstract: This paper presents a hardware architecture that implements the Contrast Maximization (CM) algorithm in Field-Programmable Gate Array (FPGA) resources f

Free Energy Manifold: Score-Based Inference for Hybrid Bayesian Networks

arXiv:2605.09839v1 Announce Type: cross Abstract: We introduce the Free Energy Manifold (FEM), a score-trained conditional energy model specialized for inference in hybrid Bayesian networks with discr

researcharxiv-cs-ai

FreeMOCA: Memory-Free Continual Learning for Malicious Code Analysis

model-releasesarxiv-cs-lg

arXiv:2605.09664v1 Announce Type: cross Abstract: As over 200 million new malware samples are identified each year, antivirus systems must continuously adapt to the evolving threat landscape. However,

GAMBIT: A Three-Mode Benchmark for Adversarial Robustness in Multi-Agent LLM Collectives

model-releasesarxiv-cs-cl

arXiv:2605.09027v1 Announce Type: new Abstract: In multi-agent systems (MAS), a single deceptive agent can nullify all gains of an agentic AI collective and evade deployed defenses. However, existing

Gemini’s biggest new features are all about controlling your phone

model-releasesthe-verge-ai

It is, once again, Gemini season. Google is announcing a host of new Gemini features during its pre-I/O Android showcase, many of which aim to help use your phone for you. You'll find Gemini in more p

Generalized Wasserstein Flow Matching: Transport Plans, Everywhere, All at Once

arXiv:2605.08424v1 Announce Type: new Abstract: Flow matching has recently emerged as a flexible and efficient framework for generative modelling by learning deterministic transport dynamics between p

researcharxiv-cs-lg

Generative AI Fuels Solo Entrepreneurship, but Teams Still Lead at the Top

arXiv:2605.10291v1 Announce Type: cross Abstract: Recent advances in generative artificial intelligence (AI) are reshaping who enters entrepreneurship, but not who reaches the top of the quality distr

Geometrically Constrained Stenosis Editing in Coronary Angiography via Entropic Optimal Transport

arXiv:2605.08851v1 Announce Type: cross Abstract: The scarcity of high-quality imaging data for coronary angiography (CAG) stenosis limits the clinical translation of automated stenosis detection. Syn

Google announces Gemini Intelligence, bundling existing and new Gemini features, including task automation across apps and vibe-coding own Android widgets (Allison Johnson/The Verge)

Allison Johnson / The Verge: Google announces Gemini Intelligence, bundling existing and new Gemini features, including task automation across apps and vibe-coding own Android widgets — We're one ste

model-releasestechmeme

Google debuts Gemini Intelligence automation features, Googlebook laptop series

model-releasessiliconangle

Google LLC is upgrading its consumer device portfolio with a set of Android features called Gemini Intelligence and a new laptop series. The company debuted the products today during a virtual event c

Google DeepMind details a Gemini-powered mouse pointer that understands what it is pointing at, allowing users to perform tasks without using text-heavy prompts (Google DeepMind)

Google DeepMind: Google DeepMind details a Gemini-powered mouse pointer that understands what it is pointing at, allowing users to perform tasks without using text-heavy prompts — We are developing mo

model-releasestechmeme

Google launches Intrusion Logging, an Android feature developed in partnership with Amnesty International and others, on Android 16 Pixel devices for now (Tim Starks/CyberScoop)

Tim Starks / CyberScoop: Google launches Intrusion Logging, an Android feature developed in partnership with Amnesty International and others, on Android 16 Pixel devices for now — Intrusion Logging m

model-releasestechmeme

Agents

Grok Voice is #1!

Grok Voice is #1! Announcing agentic performance benchmarking for Speech to Speech models on Artificial Analysis. We use 𝜏-Voice to measure tool calling and customer interaction voice agent capabiliti

agentselon-musk--x

Grounded Satirical Generation with RAG

arXiv:2605.10853v1 Announce Type: new Abstract: Humor generation remains challenging task for Large Language Models (LLMs), due to their subjective nature. We focus on satire, a form of humor strongly

researcharxiv-cs-cl

GuardAD: Safeguarding Autonomous Driving MLLMs via Markovian Safety Logic

arXiv:2605.10386v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) are increasingly integrated into autonomous driving (AD) systems; however, they remain vulnerable to diverse sa

safetyarxiv-cs-ai

Local Ai

HeteroGenManip: Generalizable Manipulation For Heterogeneous Object Interactions

arXiv:2605.10201v1 Announce Type: cross Abstract: Generalizable manipulation involving cross-type object interactions is a critical yet challenging capability in robotics. To reliably accomplish such

local-aiarxiv-cs-ai

Hierarchical Mixture-of-Experts with Two-Stage Optimization

arXiv:2605.08292v1 Announce Type: cross Abstract: Sparse Mixture-of-Experts (MoE) models scale capacity by routing each token to a small subset of experts. However, their routers exhibit a fundamental

researcharxiv-cs-ai

Tutorials

Hierarchical Multi-Fidelity Learning for Predicting Three-Dimensional Flame Wrinkling and Turbulent Burning Velocity

arXiv:2605.08232v1 Announce Type: new Abstract: High-fidelity experimental characterization of turbulent premixed flames remains limited by the cost and complexity of advanced diagnostics, particularl

tutorialsarxiv-cs-lg

High Precision Hydraulic Excavator Control for Heavy-Duty Grading

model-releasesarxiv-cs-ro

arXiv:2605.09465v1 Announce Type: new Abstract: High-precision heavy-duty grading is a common step in earthworks, traditionally carried out manually by skilled operators. Removing a significant amount

Higher Resolution, Better Generalization: Unlocking Visual Scaling in Deep Reinforcement Learning

model-releasesarxiv-cs-lg

arXiv:2605.10546v1 Announce Type: new Abstract: Pixel-based deep reinforcement learning agents are typically trained on heavily downsampled visual observations, a convention inherited from early bench

Human-Inspired Memory Architecture for LLM Agents

arXiv:2605.08538v1 Announce Type: new Abstract: Current LLM agents lack principled mechanisms for managing persistent memory across long interaction horizons. We present a biologically-grounded memory

Human-LLM Dialogue Improves Diagnostic Accuracy in Emergency Care

arXiv:2605.08533v1 Announce Type: new Abstract: Clinical decision-making in emergency medicine demands rapid, accurate diagnoses under uncertainty. Despite benchmark progress, evidence for LLMs as int

HYPERPOSE: Hyperbolic Kinematic Phase-Space Attention for 3D Human Pose Estimation

arXiv:2605.10100v1 Announce Type: cross Abstract: We introduce HYPERPOSE, a novel 3D human pose estimation framework that performs spatio-temporal reasoning entirely within the Lorentz model of hyperb

researcharxiv-cs-ai

HyperSpace: A Generalized Framework for Spatial Encoding in Hyperdimensional Representations

arXiv:2604.15113v2 Announce Type: replace Abstract: Vector Symbolic Architectures (VSAs) provide a well-defined algebraic framework for compositional representations in hyperdimensional spaces. We int

I love seeing a new eval with such low scores. When we announced GPT-5.5, almost every benchmark had a score above 50%. It's time to retire …

model-releasesgary-marcus--x

I love seeing a new eval with such low scores. When we announced GPT-5.5, almost every benchmark had a score above 50%. It's time to retire evals like GQPA and bring in a new set. The first ProgramBen

I needed to book flights for a bunch of upcoming travel. As always, I used Claude Cowork to do it. In the past, Cowork has been decent at bo…

model-releasesboris-cherny--x

I needed to book flights for a bunch of upcoming travel. As always, I used Claude Cowork to do it. In the past, Cowork has been decent at booking flights, but with Opus 4.7, for the first time ever, i

I put my flight preferences in my Cowork instructions, then let Opus get to work. It opened my browser, navigated a bunch of websites, and b…

model-releasesboris-cherny--x

I put my flight preferences in my Cowork instructions, then let Opus get to work. It opened my browser, navigated a bunch of websites, and booked everything for me. The result: Cowork booked 8 flights