AI summaries of arXiv papers
Paste any arXiv URL or paper ID. Or just swap arxiv.org → arxivtldr.org in your address bar.
Add to Chrome — Free ExtensionToday's Papers
AutoDesign: Meta-Harness Optimization for Long-Horizon Agentic Design
AutoDesign optimizes meta-harnesses for long-horizon agentic design, excelling in academic poster generation.
OmniScientist: An Omni-Modal Omni-Discipline AI Scientist
OmniScientist is an omni-modal AI scientist that conducts multidisciplinary research directly from raw, heterogeneous evidence, completing full workflows.
Localised Horizons and Holographic Thermodynamics: Supercooling in the 1/D Expansion
This paper uses a 1/D expansion in holography to predict universal features of thermal confinement transitions and supercooling.
V-RAE: Rethinking Video Latent Spaces for Generation
V-RAE leverages frozen vision foundation models to create semantically rich video latent spaces, significantly improving video generation and prediction.
HumanTracker: Towards Comprehensive and Human-Aligned Motion Tracking Benchmark
This paper presents HumanTracker, a new benchmark and HumanScore, a preference-aligned metric, to improve humanoid motion tracking evaluation.
Defensive Boosting for Online Probabilistic Forecasting
The Defensive Booster algorithm provides simultaneous strong guarantees for online probabilistic forecasting against adaptive adversaries, outperforming prior methods in both accuracy and efficiency.
PlayWorld: Benchmarking World Models with Agent Players over Long-Horizon Objectives
PlayWorld introduces a new benchmark using agent players to evaluate world models on long-horizon interactive objectives, revealing current models' limitations.
Exponential Convex Calibration Dimension for the Multi-Label Jaccard Measure
Exact calibration for the multi-label Jaccard score requires exponential prediction dimensions, though polynomial approximations are possible.
QuoteBench: How Matched Scores Can Hide Command-Path Failures
QuoteBench reveals how LLM coding agent performance is skewed by command-path failures, advocating for detailed evaluation beyond matched scores.
Alaya-EVOKE: From Linear-Scaling Supervision to Endless World
Evoke is a world model that uses external memory and a redesigned teacher to enable long-horizon, open-ended interactive generation with bounded context.
LittleLearner: Language Models Under Pedagogically Controlled Knowledge Exposure
LittleLearner is an LLM trained on a curated elementary school curriculum (LittleCurriculum) to study knowledge acquisition under controlled exposure.
Effective field theory of quasi-hydrodynamics from kinetic theory
This paper develops a rigorous effective field theory for linear quasi-hydrodynamics, showing Israel-Stewart-like dynamics universally describe slow relaxation modes.
Browse by Category
Browse by author →Browse by Topic
All topics →📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.
The URL swap trick
Reading a paper on arxiv.org? Just change the domain to arxivtldr.org in your address bar:
Paste a link
Enter any arXiv URL or paper ID
AI summarizes
Get a TLDR, key bullets, and why it matters
Read & share
Get key insights in seconds, share with colleagues