ml infra

RSS

31
WeVisDoc: From Coverage to Capability for Robust End-to-End Document Parsing arxiv.org
WeVisDoc introduces a two‑stage data‑centric pipeline for end‑to‑end document parsing. Stage I expands coverage using heterogeneous data and structure‑preserving degradations. Stage II probes the Stage I model with a held‑out set, clusters residual errors, and directs targeted data creation and token‑budget reallocation. The 4‑billion‑parameter model reaches 95.38 Overall on OmniDocBench v1.6 and…
Hugging Face Daily Papersarxiv.orgHao Yu, Kang Liu, Linnan Zhao5d1 minpaper
32
The Other Half of the Memory Wall: Serving 35B MoEs from SSD with Trained Routing Prediction arxiv.org
The paper presents Edge0, a streaming MoE inference engine that predicts the next layer's routing one token ahead, allowing expert weights to be fetched from SSD while compute proceeds. This enables a 35 B‑parameter MoE to run on a single 24 GB machine at ~20 tokens/s using only ~3 GiB of active memory and with near‑teacher accuracy.
Hugging Face Daily Papersarxiv.orgYu Lin, Yiming Wang, Runyuan Cai6d1 minpaper
33
Vibe Coding is the new Internet Dating?joecmarshall.com
A speculative opinion piece comparing the rise of AI‑assisted "vibe coding" to the evolution of internet dating, arguing that as AI becomes the default coding partner the term will become obsolete.
Hacker News front pagejoecmarshall.comJoe Marshall7d2 minHN6299
34
RiskChainBench: A Benchmark for Obfuscated Platform Message Restoration and Evidence-Grounded Web Investigation arxiv.org
RiskChainBench is a new benchmark that pairs synthetic obfuscated message restoration inputs with human‑labeled local web environments, requiring models to both decode malicious instructions and investigate the linked site. Across ten models, restoration accuracy varies widely and web‑agent failures dominate the error budget.
Hugging Face Daily Papersarxiv.orgZhuoXin Liu, Zhiming Ma, Ying Zhang7d1 minpaper
35
Reflect, Revise, Reuse: Training-Free Skill Evolution for GUI Agents arxiv.org
EvoSkill‑GUI lets GUI agents revise their procedural skills on‑the‑fly without extra training by using a reflect‑revise‑reuse loop that edits skill packages during execution. The approach yields up to +16.2% improvement on MobileWorld and similar gains on AndroidWorld and OSWorld, and the evolved skills transfer to related tasks.
Hugging Face Daily Papersarxiv.orgBofan Chen, Boxuan Zhang, Fei Tang7d1 minpaper
36
In-Context Robot Learning with VLM Agents arxiv.org
GPT‑Policy is a framework that lets a large vision‑language model (e.g. GPT‑6 Astra) perform in‑context robot learning: a context compiler extracts visual transitions from demos, the VLM proposes tool actions, and a constrained controller verifies and executes them. Real‑robot experiments show that raw video demos improve success rates even without explicit action labels, and that providing align…
Hugging Face Daily Papersarxiv.orgDongzhou Cheng, Taoran Yi, Ye Fang6d1 minpaper
37
How do Traffic Signals Work (2019)practical.engineering
A high‑level overview of how traffic signals are designed and operated: basic phases, timing rules, actuated detection, coordination between adjacent lights, and emerging adaptive‑control systems that use centralized data and ML. The piece stays at the level of civil‑engineering concepts and does not dive into implementation details relevant to software engineers.
Hacker News front pagepractical.engineeringby Grady Hillhouse, Grady Hillhouse13h9 minHN5636
38
OmniVBench: A Benchmark and Large-Scale Dataset for Omni Reference-to-Video Generation arxiv.org
OmniVBench is a new benchmark and the Omni‑R2V Dataset, offering 7 task families, 18 fine‑grained reference‑to‑video generation tasks and a factor‑grounded evaluation checklist of over 12 k items. The dataset provides 340 k industrial‑grade video samples and pipelines for constructing reference‑target pairs, exposing large performance gaps in current R2V models.
Hugging Face Daily Papersarxiv.orgWenxue Li, Peiyan Guan, Haoyang Jiang4d2 minpaper
39
TypeSafe AI's Jev now available on AI Gateway vercel.com
Vercel AI Gateway now offers Jev, a probabilistic decision model that returns typed choices, scores, and booleans instead of raw text. TypeSafe AI reports it runs up to 193× faster and 445× cheaper than standard LLMs, exposed via the experimental evaluate API in AI SDK 7.
Vercelvercel.comRohan Taneja6d2 minrelease
40
Your Agent Aced the Task. Will It Do It Again?huggingface.co
The post introduces the Consistency Analyzer, a cheap black‑box diagnostic that flags flip‑prone decision steps in LLM agent traces, and shows how feeding the resulting consistency guidelines back into ALTK‑Evolve halves the gap between mean success and all‑run success (Pass⁵) on the AppWorld benchmark without hurting average accuracy.
Hugging Facehuggingface.coEvelyn Duesterwald, Lilian Ngweta, Vatche Isahagian, Jayaram Radhakrishnan, Vinod Muthusamy, Gaodan Fang, Ashwath Vaithinathan Aravindan, Punleuk Oum, G Thomas, Merve Unuvar7d8 minHN21
41
Retention-Constrained Post-Training Quantization of Cellpose-SAM for Stem Cell Microscopy arxiv.org
The paper proposes a retention‑constrained post‑training quantization benchmark for Cellpose‑SAM, showing that weight‑only W8A16 and a mixed W4/W8 scheme keep instance F1 scores while cutting model size 6.76×, whereas ternary quantization fails on most images.
Hugging Face Daily Papersarxiv.orgSebastián A. Cruz Romero5d1 minpaper
42
ModularRSI: Modular and Generalizable Recursive Harness Self-Improvement arxiv.org
ModularRSI proposes a contrastive, modular approach to evolve agent harnesses by independently updating five functional modules and aggregating evidence across 2,000 benchmark‑disjoint tasks. Experiments on TB2.0 and SWE‑Bench Verified show consistent gains on unseen tasks and cross‑model transfer.
Hugging Face Daily Papersarxiv.orgSiwei Wu, Jincheng Ren, Yizhi Li8d1 minpaper
43
FRAUDSkill: Structured Frozen-Weight Skill Optimization for Audio Anti-Fraud Detection arxiv.org
FRAUDSkill is a framework that keeps a pretrained audio‑language model frozen and learns an external layer of skill programs, routing policies, and decision rules to meet a structured anti‑fraud detection protocol. On the TeleAntiFraud benchmark it reaches 73.5% Macro‑F1 (≈32% improvement) while cutting invalid predictions to 1.94%.
Hugging Face Daily Papersarxiv.orgChengxian Hu, Zhiming Ma, Mingjun Pan6d1 minpaper
44
Apple detectives solved mystery of ancient tree and rewrote the history of fruit scientificamerican.com
Geneticist Cameron Peace used DNA fingerprinting to identify a centuries‑old apple tree in Maine that turned out to be a lost progenitor of many modern U.S. apple varieties, highlighting the importance of preserving genetic diversity for future breeding.
Hacker News front pagescientificamerican.comMaryn McKenna6d19 minHN6919
45
tokenizers v1: encode, decode and scaling, measured huggingface.co
Hugging Face has released `tokenizers` v1, a major performance update that achieves 3-30x faster encoding than v0.23 while maintaining identical output and API compatibility. Key optimizations include a SIMD-accelerated splitter, a thread-local word cache, and an allocation-free BPE merge loop, ensuring tokenization doesn't bottleneck ML workflows.
Hugging Facehuggingface.coArthur Zucker, Simon Brandeis, Luc Georges, Lysandre1d10 min
46
Open-weight models take 56% of token volume, Astra doubles Fable 5.1 spend vercel.com
Vercel’s September AI Gateway Production Index shows open‑weight models processing 56% of token volume (up from 7% in Dec 2025) while accounting for only 14% of spend. Token price fell 23.2% month‑over‑month. Anthropic’s Opus 5 captured 22.5% of spend, overtaking Fable 5 which dropped to 4.9%. OpenAI’s new GPT‑6 Astra grabbed ~7.7% of total gateway spend in its first 12 days, more than double Ant…
Vercelvercel.comAmelia Charles5d6 minHN2
47
PhysStream: Streaming Physics-Grounded Video Generation with Structured Scene Memory and Fine-Grained Motion Control arxiv.org
PhysStream introduces a two‑stage autoregressive video generator that uses online‑derived positional and tracking maps (structured scene memory) and sparse velocity‑increment signals to enable fine‑grained, physics‑grounded control of multi‑object tabletop scenes. It cuts motion distribution error by 33 % and trajectory error by 12 % versus strong baselines, and wins 85 % of human preference test…
Hugging Face Daily Papersarxiv.orgChuhao Chen, Peter Wonka, Chaoyang Wang7d1 minpaper
48
Shared Selective Persistent Memory for Agentic LLM Systems apple.com
Apple proposes a memory architecture for agentic LLMs that selectively persists reusable context (specs, schemas, configs, constraints) across sessions and users. Shared workspaces with role‑based access and a zero‑token data‑refresh mechanism cut token usage by 97×, reduce task time by 14×, and raise task‑completion rates to 96% versus 71%‑79% for baselines.
Apple Machine Learning Researchapple.com6d1 minpaper
49
Mind2Dialogue: Training Human-Aware Language Models by Simulating User Mental States arxiv.org
Mind2Dialogue introduces a psychology‑guided simulator that generates user mental states and uses privileged distillation to train LLM assistants on Oracle responses that know those states. The resulting models improve personalization and theory‑of‑mind metrics by up to 41 percentage points versus standard instruction‑tuned baselines.
Hugging Face Daily Papersarxiv.orgZixuan Wang, Yufan Zhou, Jinzhou Tang8d2 minpaper
50
Your AI coding agent evaluation is only as good as its sandbox microsoft.com
Evaluating AI coding agents requires a robust sandbox to prevent agents from retrieving answers from the environment, which can invalidate tests of internal knowledge. A correct answer doesn't guarantee a valid measurement if the agent accessed information it shouldn't have, highlighting the need to define sandboxes by information boundaries rather than just tool restrictions. Always review agent…
Microsoft for Developersmicrosoft.comWaldek Mastykarz6d5 min
51
AI Infra Summit: NVIDIA Vera Rubin and DSX Platform Advancements Showcase Energy Efficiencies of Optimizing Tokens Per Watt for AI Factories nvidia.com
NVIDIA showcased new power‑aware AI‑factory features at the AI Infra Summit. DSX MaxLPS and Flex improve token‑per‑watt efficiency by up to 40%, while the Vera Rubin NVL72 platform delivers up to 30× higher throughput per megawatt on agentic workloads.
Nvidianvidia.com7d7 min
52
Presentation: Context Engineering at LinkedIn: How We Built an Organizational Context Layer for AI Agents with MCP infoq.com
LinkedIn built a “Contextual Agent Playbooks and Tools” layer that wraps internal services (code search, docs, feature flags, etc.) behind the open‑source Model Context Protocol (MCP). By feeding LLM‑powered coding agents the exact internal artifacts they need, the agents can diagnose incidents, generate PRs, and update incident tickets in minutes, delivering a reported 20 % productivity gain wit…
InfoQinfoq.comAjay Prakash3d28 mintalk
53
Decoy Direction Optimization: A Post-Hoc Defense Against LLM Abliteration arxiv.org
Decoy Direction Optimization (DDO) is a post‑hoc weight‑editing defense for open‑weight LLMs that injects a high‑magnitude nonlinear decoy into MLP neurons, corrupting contrastive estimators used by Refusal Feature Ablation (RFA) attacks. The paper proves a spectral bound on the effect, evaluates DDO on six model families (including Llama‑3‑8B‑Instruct), and shows <10 % attack success rate (ASR)…
Hugging Face Daily Papersarxiv.orgAashiq Muhamed, Mona T. Diab, Virginia Smith8d1 minpaper
54
OmniHarness: Harnessing Generalizable Visual Generation via Symbolic Policy Learning arxiv.org
OmniHarness introduces a symbolic‑policy framework that extracts reusable procedural knowledge from multimodal LLM‑driven visual generation runs. By decoupling task logic from instance inputs, the system can instantiate, adapt, and compose policies for new visual tasks, using intermediate verification for on‑the‑fly refinement while keeping the underlying model frozen. Self‑directed practice task…
Hugging Face Daily Papersarxiv.orgXu Xu, Jinxiu Liu, Zhangbo Qiao9d1 minpaper
55
Article: Your Next DSL Author Is a Language Model infoq.com
Typed Domain Grounding (TDG) embeds a DSL inside a mainstream language the LLM already knows (e.g., Kotlin) and uses the host compiler as an oracle. The author describes five building blocks—embedding, choosing a host language with high training‑data frequency, compiler‑driven type safety, a generate‑compile‑repair loop, and an on‑demand teaching tool—and shows measured results from kUML, a Kotli…
InfoQinfoq.comIrakli Betchvaia6d18 min
56
Glyph: A Multi-Strategy Agentic System for Column Description and Sensitivity-Ontology Tagging of Enterprise Data Catalogs apple.com
Glyph is a production system that uses coordinated LLM agents and a fine‑tuned MiniLM encoder to automatically generate column descriptions and assign ontology tags in enterprise data catalogs. It combines code‑grounded retrieval, regex, and contrastive vector search, achieving NDCG@10 0.92 and MAP@100 0.90, and provides auditable provenance for each tag.
Apple Machine Learning Researchapple.com6d1 minpaper
57
Beyond Top-k Skill Retrieval: Diversity-Aware Skill Routing for LLM Agents arxiv.org
Beyond Top‑k Skill Retrieval: Diversity‑Aware Skill Routing (DSR) applies a Determinantal Point Process with a query‑residual diversity kernel to rerank skill candidates, balancing relevance and redundancy. On the SkillRouter benchmark it raises recall and full‑coverage, especially for multi‑skill queries, showing that skill routing benefits from set‑selection rather than independent ranking.
Hugging Face Daily Papersarxiv.orgWang Wei, Tiankai Yang, Samyadeep Basu17d1 minpaper
58
University of Manchester Uses NVIDIA Earth-2 to Forecast Air Pollution Across the UK nvidia.com
University of Manchester used NVIDIA Earth‑2 CorrDiff and StormCast generative models to downscale UK‑wide air‑pollution simulations. Training on the Isambard‑AI supercomputer (5,448 GH200 chips, 21 EFLOPS) took two days on an eight‑GPU node, producing a 2‑3 km resolution model. Inference runs on a desktop‑class DGX Spark, enabling rapid scenario forecasting and potential real‑time health alerts.…
Nvidianvidia.comIsha Salian6d4 min
59
From Megawatts to Tokens: How NVIDIA Maximizes AI Factory Production nvidia.com
NVIDIA’s DSX platform lets AI data‑centers shift workloads in response to grid signals, squeezing ~24% more token throughput (4 M→5 M tps) and ~23% better performance‑per‑watt on a fixed megawatt budget. The first production demo used Emerald AI’s Conductor to drop a 4 MW load to 3 MW in under a minute without interrupting high‑priority jobs. DSX MaxLPS reallocates headroom across HGX B200 server…
Nvidianvidia.comVishal Ganeriwala7d5 min
60
MilleMiglia: A realistic instance generator for middle-mile logistics research.google
MilleMiglia is an open‑source C++ generator that creates realistic, privacy‑preserving middle‑mile logistics instances (space‑time graphs with fixed schedules, throughput limits, and synchronization constraints). It uses data‑driven spatial, demand, and rotation distributions, serializes with protobuf, and ships small files for small‑toy to continent‑scale problems, enabling both exact/heuristic…
Google Researchresearch.google3d7 min