LIVE · refreshes every 20 min
updated Oct 18, 09:03 UTC
110010.art(ificial intelligence)
AI products, research, and launches — clustered and ranked, not just re-blogged.
110010.art / archive

Research

Sep 24
24d ago 4DGS-JEPA: a Gaussian-native joint-embedding predictive architecture for causal multi-horizon prediction over dynamic Gaussian scenes arXiv cs.AI → story
24d ago An accurate and interpretable hypergraph neural network for GBM survival prediction arXiv cs.AI → story
24d ago Automated extraction of RoPA records using hybrid retrieval and locally deployed LLMs arXiv cs.CL → story
24d ago Benchmark dataset classifies interpretive canons at the sentence level using decisions of the German Federal Constitutional Court arXiv cs.CL → story
24d ago Building socio-affective artificial intelligence for interactive multi-agent simulations arXiv cs.AI → story
24d ago Can one adapted model do it all? Fine-tuning strategy selection for customer support LLMs arXiv cs.CL → story
24d ago Caveat: Towards robust computer-use agents in incentive-misaligned environments arXiv cs.AI → story
24d ago ChipMEM: Verification-grounded memory for EDA agents arXiv cs.LG → story
24d ago Clarification is not correction: LLMs fail to let go arXiv cs.AI → story
24d ago COMED: The Missing Middle Between Routing and Collaboration in Multi-LLM Inference arXiv cs.CL → story
24d ago Count evidence, not sentences: tempered evidence fusion of LLM judgments for long-text value measurement arXiv cs.CL → story
24d ago CRISP: scalable importance-stratified coresets for imbalanced tabular learning arXiv cs.LG → story
24d ago Distilling sequential computation in transformer language models arXiv cs.CL → story
24d ago Do We Need Complex Topology Control? Distinct-Peer Random Routing Improves Cost-Efficiency in Sparse Multi-Agent Debate arXiv cs.AI → story
24d ago DRSR: learning set-level deletion risk for efficient long-horizon agents arXiv cs.AI → story
24d ago EduBehaviors: Assertion-based schemas for auditable coding of educational dialogues arXiv cs.CL → story
24d ago Enhancing small language models for power outage report generation via minimum risk training arXiv cs.AI → story
24d ago EnSIMem: entity-structured long-term memory for agents arXiv cs.AI → story
24d ago Escaping Python dependency hell: a hybrid replay-and-repair pipeline for Python dependency resolution arXiv cs.AI → story
24d ago Experts rise where LLMs disagree; using cross-model disagreement to target expert effort in LLM codebook revision for large-scale annotation arXiv cs.CL → story
24d ago From decorative to load-bearing: task difficulty shapes the causal role of chain-of-thought arXiv cs.AI → story
24d ago From offline proxies to online decisions: a layered engagement evaluation framework for conversational AI arXiv cs.AI → story
24d ago GeoRVQ uses decoder-aware geometry for residual-token prediction in physiological signals arXiv cs.LG → story
24d ago Group Relative Policy Optimization improves credit assignment for decomposed subtasks in language-model agents arXiv cs.AI → story
Sep 23
24d ago Offloaded inference for real-world physical AI robotics improves task success and efficiency by moving AI inference beyond the robot, according to Microsoft Research. Microsoft Research → story
24d ago MIT appoints David Siegel SM ’86, PhD ’91 as Innovation Fellow to explore AI-assisted scientific discovery MIT News (AI) → story
25d ago Prompt breadth and rollout refresh interact in on-policy distillation; study varies prompt bank and response-generating policy snapshots to evaluate impact on performance arXiv cs.CL → story
25d ago ReliMap decomposes LLM-based human behavior simulation into three layers to assess reliability arXiv cs.CL → story
25d ago Retrieved-span training improves query-focused meeting summarization on QMSum; fine-tuning a Fusion-in-Decoder model with 2,000-word retrieved spans recovers ROUGE-1 loss and outperforms a 1.2B system on test. arXiv cs.CL → story
25d ago Same quantity, different answer: numerical representation invariance in language models arXiv cs.CL → story
25d ago Self-cleaning and captured anyway: one measured primitive for error in a store an agent writes to itself, and what a falling score actually measures arXiv cs.CL → story
25d ago Semi-supervised federated ASR with online pseudo-labels and server update stabilization for reduced divergence arXiv cs.LG → story
25d ago Spatiotemporal Kronecker covariance neural networks arXiv cs.LG → story
25d ago Stable unsupervised continual chunking with Sheaf SyncMap arXiv cs.LG → story
25d ago Terminal shrinkage averaging reveals a schedule-estimator interaction in LLM pretraining arXiv cs.LG → story
25d ago Topological signal processing with unoriented operators arXiv cs.LG → story
25d ago Training a language model end-to-end in Rust; author reports training alone with no PyTorch or Python and documents five defects in Candle, Burn backends for 2026, in a measured failure taxonomy arXiv cs.CL → story
25d ago Trains but doesn't learn: a post-training delivery benchmark for LLM agents as forward-deployed engineers arXiv cs.LG → story
25d ago Ufakzeka-1: a 151M-parameter Turkish language model pretrained from scratch and instruction-tuned for chat; paper documents building and evaluation costs, tokenizer, and benchmarking arXiv cs.CL → story
25d ago What does 99% accuracy measure? A reproducible audit of shortcut learning in a widely used fake news corpus arXiv cs.CL → story

Older items: monthly archive →