All entries

mupscaling small models: Principled warm starts and hyperparameter transfer

arXiv:2602.10545v2 Announce Type: replace-cross Abstract: Modern large-scale neural networks are often trained and released in multiple sizes to accommodate diverse inference budgets. To improve effic

Agents

'My belief is that, three years down the road, people will not own PCs. You will stop buying PCs and laptops. You will just have a phone, or…

DGX agent

'My belief is that, three years down the road, people will not own PCs. You will stop buying PCs and laptops. You will just have a phone, or maybe you might have a tablet.' -- AGI Inc.'s Div Garg on w

agentsdiv-garg--x

3 Jul 2026

NarrativeTrack: Evaluating Entity-Centric Reasoning for Narrative Understanding

model-releasesarxiv-cs-lg

arXiv:2601.01095v3 Announce Type: replace-cross Abstract: Multimodal large language models (MLLMs) have achieved impressive progress in vision-language reasoning, yet their ability to understand tempo

NAVER LABS Europe Submission to the Instruction-following 2026 Short Track

arXiv:2607.01960v1 Announce Type: new Abstract: In this paper, we describe NAVER LABS Europe's submission to the instruction-following speech processing short track at IWSLT 2026. We participate again

researcharxiv-cs-cl

Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging

arXiv:2510.17426v3 Announce Type: replace-cross Abstract: The 'alignment tax' of post-training is typically framed as a drop in task accuracy. We show it also involves a severe loss of calibration, ma

NeoMap: Training-free Novel-View Synthesis from Single Images and Videos

arXiv:2607.01962v1 Announce Type: cross Abstract: We study the challenging problem of novel view video synthesis from single images or monocular videos. Existing methods, which operate under the assum

Neuro-Symbolic Safety Guidance for Vision-Language-Action Models via Constrained Flow Matching

model-releasesarxiv-cs-ro

arXiv:2607.01378v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models have demonstrated promising generalization capabilities across robotic manipulation tasks, yet their real-world depl

NeuroBridge: Bridging Multi-Task MRI Knowledge for Neurodegenerative Disease Diagnosis

arXiv:2607.01401v1 Announce Type: cross Abstract: INTRODUCTION: Accurate MRI-based identification of Alzheimer's disease (AD), mild cognitive impairment (MCI), and related dementias remains challengin

Neuron-Aware Active Few-Shot Learning for LLMs

arXiv:2607.02423v1 Announce Type: cross Abstract: Active Few-Shot Learning (AFSL) adapts LLMs to specialized domains by identifying the most valuable unlabeled samples for annotation and use as few-sh

Neuron-Aware Data Selection for Annotation-Free LLM Self-Distillation

arXiv:2607.02460v1 Announce Type: cross Abstract: Post-training large language models (LLMs) without real-world interaction feedback or human-labeled supervision remains challenging, particularly in s

NEUROSYMLAND: Neuro-Symbolic Landing-Site Assessment for Robust and Edge-Deployable UAV Autonomy

model-releasesarxiv-cs-ro

arXiv:2607.02277v1 Announce Type: new Abstract: Safe landing-site assessment in unstructured environments remains a key challenge for autonomous UAV deployment, as vision-only learning approaches ofte

NEW paper worth reading. (bookmark it) The basic idea is to pair a compressive recurrent state with a small exact memory, which helps to rec…

NEW paper worth reading. (bookmark it) The basic idea is to pair a compressive recurrent state with a small exact memory, which helps to recover long-range recall without giving up the efficiency of l

tutorialsdair-ai--x

Non-synchronism in Global Usage of Research Methods in Library and Information Science from 1990 to 2019

arXiv:2607.01833v1 Announce Type: cross Abstract: The global development of Library and Information Science (LIS) is influenced by various factors such as the economy, society, culture, discipline, tr

tutorialsarxiv-cs-cl

Tools

None of it was an accident. A team that gave up recharge week, and a partnership @GeoffBibby built with @swyx + the AI Engineer crew that tu…

None of it was an accident. A team that gave up recharge week, and a partnership @GeoffBibby built with @swyx + the AI Engineer crew that turned a year-old idea into the biggest stage in AI eng. Thank

toolsswyx--x

Object Aligner: A Configurable JSON Schema Similarity Score for Graphs, Applied to LLM Prompt Optimization

arXiv:2607.01972v1 Announce Type: cross Abstract: Large language models (LLMs) are often asked to produce JSON conforming to a fixed schema, powering information extraction, tool calling, agentic plan

Object-centric LeJEPA

arXiv:2607.02404v1 Announce Type: cross Abstract: Image encoders trained with LeJEPA can deliver strong features for downstream tasks, but, like other image-level self-supervised methods, typically re

researcharxiv-cs-lg

Office Comprehension Benchmark

arXiv:2607.01245v1 Announce Type: cross Abstract: We introduce Office Comprehension Bench (OCB), the first public benchmark to jointly evaluate LLM systems on Word, Excel, and PowerPoint comprehension

OmniGAIA: Towards Native Omni-Modal AI Agents

arXiv:2602.22897v3 Announce Type: replace Abstract: Human intelligence naturally intertwines omni-modal perception -- spanning vision, audio, and language -- with complex reasoning and tool usage to i

On the Asymptotics of Self-Supervised Pre-training: Two-Stage M-Estimation and Representation Symmetry

arXiv:2603.27631v2 Announce Type: replace Abstract: Self-supervised pre-training, where large corpora of unlabeled data are used to learn representations for downstream fine-tuning, has become a corne

tutorialsarxiv-cs-lg

On the Dimension-Free Approximation of Deep Neural Networks for Symmetric Korobov Functions

arXiv:2511.12398v2 Announce Type: replace Abstract: Deep neural networks have been widely used as universal approximators for functions with inherent physical structures, including permutation symmetr

researcharxiv-cs-lg

On the Limits of Steering Vectors for Preference-Aligned Generation

model-releasesarxiv-cs-cl

arXiv:2607.01802v1 Announce Type: new Abstract: Steering vectors have emerged as a promising approach to controlled text generation, offering interpretable, training-free mechanisms for shaping model

On the Role of Computation in Reinforcement Learning

arXiv:2602.05999v4 Announce Type: replace Abstract: How does the amount of compute available to a reinforcement learning (RL) policy affect its learning? Can policies using a fixed amount of parameter

On the Role of Directionality in Structural Generalization

arXiv:2607.02307v1 Announce Type: new Abstract: Several SLOG test categories explicitly involve directional distinctions (modifier position shifts, argument extraction positions), yet AM-Parser, the p

researcharxiv-cs-cl

On the Sample Efficiency of Inverse Dynamics Models for Semi-Supervised Imitation Learning

arXiv:2602.02762v2 Announce Type: replace Abstract: Semi-supervised imitation learning (SSIL) consists in learning a policy from a small dataset of action-labeled trajectories and a much larger datase

On the Utility and Factual Reliability of Pruned Mixture-of-Experts Models in the Biomedical Domain

arXiv:2607.01444v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) models offer inference speedups via selective activation but impose substantial memory requirements because the whole network

One Demonstration Is Enough for Real-World Robotic Reinforcement Learning

arXiv:2607.01651v1 Announce Type: new Abstract: Learning effective robot control policies on physical hardware is challenging due to costly data collection and the difficulty of reward specification.

safetyarxiv-cs-ro

One More Time: Revisiting Neural Quantum States from a Reinforcement Learning Perspective

model-releasesarxiv-cs-lg

arXiv:2607.02292v1 Announce Type: new Abstract: Neural quantum states (NQS) provide a flexible and scalable framework for approximating quantum many-body wavefunctions. Among NQS parameterizations, au

One way to live 400 years is to do (and experience) 5x more things per day

researchfrancois-chollet--x

This post explores a philosophical perspective on subjective time perception, suggesting that by engaging in more diverse activities and novel experiences daily, one can psychologically expand their l

Online Resource Allocation with Continuous Random Consumption: Regret under Degeneracy

arXiv:2607.02196v1 Announce Type: new Abstract: We study online resource allocation when both rewards and consumption sizes may be continuously distributed. Requests arrive sequentially and must be ac

Online Safety Monitoring for LLMs

arXiv:2607.02510v1 Announce Type: new Abstract: Despite alignment training, LLMs remain prone to generating unsafe outputs at deployment time. Monitoring outputs online and raising an alarm when safet

OntoLearner: A Modular Python Library for Ontology Learning with Large Language Models

arXiv:2607.01977v1 Announce Type: new Abstract: Ontology learning (OL) aims to automatically construct structured knowledge models from text, yet progress remains fragmented across methods, domains, a

Tools

Open model usage has gone from 10% of AI tokens to 30% in a year. The shift to open, modular AI is here to stay. Our founders on what's driv…

Open source AI model usage has tripled from 10% to 30% of total AI tokens consumed within a year, reflecting a significant market shift toward open and modular AI architectures. This trend indicates g

toolstogether-ai--x

Open Source AI Gap Map

model-releasessimon-willison

Open Source AI Gap Map Current AI is 'a global partnership building a public option for AI', founded as a non-profit at the AI Action Summit in Paris in February 2025 and backed by serious capital ($4

OpenAI offers feds a stake, Anthropic gets out of AI model jail and Meta wants to be a neocloud

OpenAI reportedly has floated giving the U.S. government a 5% stake in the company, perhaps the start of a series of such stakes in other AI companies as well. This no doubt has traditional anti-indus

safetysiliconangle

OpenSafeIntent: Evaluating Intent-Calibrated Safe Completion Across Dual-Use Prompt Sets

arXiv:2607.02047v1 Announce Type: cross Abstract: Safe completion requires models to provide useful assistance without enabling harm, but this behavior is difficult to evaluate with isolated prompts.

OpenWiki is at 1.7k stars in just 3 days! Right now it's just for codebases, but we're working to expand it to everything for memory. What d…

OpenWiki is at 1.7k stars in just 3 days! Right now it's just for codebases, but we're working to expand it to everything for memory. What do you want to see in a general purpose memory wiki agent? ht

agentsharrison-chase--x

OpenWiki up to 1.7k GitHub stars in two days Most common ask it to make it more general purpose (not just coding) What sources do people wan…

OpenWiki up to 1.7k GitHub stars in two days Most common ask it to make it more general purpose (not just coding) What sources do people want to see? Notion? GMail? Slack? GDrive? Internet search? Oth

agentsharrison-chase--x

OpenWiki 📈📈 What do you want to see from a general purpose wiki?

OpenWiki 📈📈 What do you want to see from a general purpose wiki? OpenWiki up to 1.7k GitHub stars in two days Most common ask it to make it more general purpose (not just coding) What sources do peopl

agentsharrison-chase--x

Ophiuchus: Incentivizing Tool-augmented 'Think with Images' for Joint Medical Segmentation, Understanding and Reasoning

arXiv:2512.14157v2 Announce Type: replace Abstract: Recent medical MLLMs have made significant progress in generating step-by-step textual reasoning chains. However, they still struggle with complex c

agentsarxiv-cs-ai

OPINE-World: Programmatic World Modeling with Ontology-error-Prioritized Interactive Exploration

arXiv:2607.01531v1 Announce Type: new Abstract: Learning how an environment behaves from interaction is central to building agents that adapt to unfamiliar tasks. World models learned with deep networ

Optimal Stabilizer Testing and Learning with Limited Quantum Memory

arXiv:2607.02444v1 Announce Type: cross Abstract: We study stabilizer state testing and learning with limited coherent quantum memory. Here an algorithm sequentially receives copies of an unknown n-qu

tutorialsarxiv-cs-lg

Optimizing RAG Rerankers with LLM Feedback via Reinforcement Learning

arXiv:2604.02091v2 Announce Type: replace-cross Abstract: Rerankers play a pivotal role in refining retrieval results for Retrieval-Augmented Generation. However, current reranking models are typicall

Optimizing Visual Generative Models via Distribution-wise Rewards

arXiv:2607.02291v1 Announce Type: new Abstract: Conventional reinforcement learning strategies for visual generation typically employ sample-wise reward functions, yet this practice frequently results

OrbitQuant: Data-Agnostic Quantization for Image and Video Diffusion Transformers

arXiv:2607.02461v1 Announce Type: cross Abstract: Diffusion transformers (DiTs) achieve state-of-the-art image and video generation, but their multi-step sampling and growing parameter count make infe

Overthink-Triggered Slowdown Attacks on LVLM-Based Robotic Systems

arXiv:2607.01518v1 Announce Type: cross Abstract: Large Vision-Language Models (LVLMs) have been increasingly integrated into robotic systems. However, these models may exhibit overthinking behaviors,

safetyarxiv-cs-ro

Overview of Risk Assessment and Management for Intelligent Systems under the AI Act and Beyond

arXiv:2607.02197v1 Announce Type: cross Abstract: The society and emerging risk-based regulatory frameworks for AI underscore the need for rigorous risk assessment to ensure safe and reliable AI syste

Applications

PACE: A Neuro-Symbolic Framework for Plausible and Actionable Counterfactual Explanations

arXiv:2607.01306v1 Announce Type: new Abstract: Counterfactual explanations explain machine learning predictions by identifying minimal input changes that would alter a model's decision. Although many

applicationsarxiv-cs-ai

PACE: A Proxy for Agentic Capability Evaluation

arXiv:2607.02032v1 Announce Type: new Abstract: Evaluating LLM agents on benchmarks like SWE-Bench and GAIA can be expensive, time-consuming, and requires complex infrastructure. A single evaluation c