저장한 도구도구 추천
FOXAI SEARCH

전체 검색

뉴스 기사 490

training-alignment

Adaptive Scheduling Pipeline for Multi-Instance Asynchronous Reinforcement Learning

arXiv:2611.08864v1 Announce Type: new Abstract: We present the Adaptive Scheduling Pipeline (ASP), an offline rollout scheduler for multi-instance asynchronous reinforcement learning (RL) of large language models (LLMs). ASP jointly reduces three sources of rollout underutilization - padding bubbles from intra-minibatch length heterogeneity, skewness bubbles from long-tail outputs, and sorting bubbles from fluctuations in training-queue occupancy induced by length-sorted dispatch - through a single offline preprocessing pass. This enables efficient multi-instance rollout that overcomes the early scaling plateau of single-instance inference while smoothing the arrival of generated samples at the trainer. We integrate the scheduler into an asynchronous architecture that decouples rollout from training while keeping policy lag observable and bounded. Complete ASP improves end-to-end throughput by up to 3.22 times and reduces iteration time by up to 76.8% relative to MindSpeed-RL, while reducing iteration time by up to 62.5% relative to verl. Within ASP, scheduling reduces mean iteration time by 10.9%. ASP also scales more robustly with additional generation resources than these baselines and shows no visible convergence degradation in the measured bounded-lag regime. ASP has since been used in an internal telecom-domain RL post-training workflow, reducing the reported training period from more than three months to about two months and producing a domain-enhanced model for knowledge answering and tool calling.

번역 서비스가 일시 중지되어 원문을 표시합니다.
arXiv cs.DC
reasoning-methods

Aspect-wise Pairwise Tournaments for Test-Time Selection

arXiv:2611.08640v1 Announce Type: new Abstract: Scaling test-time compute has proven highly effective at improving LLM reasoning. A popular method is best-of-n sampling, where a model generates many candidate solutions and one is selected. This motivates the question: how can we select the best candidate? When humans make choices, we find it more natural to compare two options than judge each in isolation. Additionally, when the options become complex, we like to judge the options one aspect at a time, such as comparing two job candidates on experience, communication, etc. In the test-time setting, candidate solutions should be judged in the same way. Here we introduce Aspect-wise Pairwise Tournaments (APT), which verifies by comparison and compares by aspect. Each matchup of a single elimination tournament is decided by a vote over multiple off-the-shelf aspect verifiers, each expressing a pairwise preference along its own aspect. Across five domains and three generator LLMs, APT improves over binary multi-aspect approval, self-consistency and reward model verification, with the largest gains on open-ended tasks. We find that both components matter; neither binary approval over the same aspects nor a generic pairwise judge at a matched budget matches APT. Because aspects are specified in natural language, APT also lets users steer selection toward properties they care about, better aligning with user preferences without needing to regenerate solutions. No component is trained, so APT transfers to new domains by simply rewriting prompts.

번역 서비스가 일시 중지되어 원문을 표시합니다.
arXiv cs.AI
coding-engineering

Claude Fable 5.2 now available on AI Gateway

Claude Fable 5.2, designed for open-ended technical tasks, is now available on AI Gateway with improved coding, reasoning, and vision capabilities. It operates as an orchestrator or advisor, supports model fallbacks for cybersecurity classifier rejections, requires data retention (no Zero Data Retention), and can be accessed via the AI SDK, APIs, or integrated tools like Claude Code and Cursor. AI Gateway offers unified API access, cost tracking, failover options, and pricing transparency without markup.

번역 서비스가 일시 중지되어 원문을 표시합니다.
Vercel
ai-chips

DataHelix: A CPU-GPU Co-Scheduling Data Plane for Multimodal Model Training

arXiv:2611.09411v1 Announce Type: new Abstract: Multimodal model training breaks the loose coupling between data loading and model execution assumed by conventional training pipelines. Large per-sample variation in decoding cost, token/latent count, and tensor size causes both CPU preprocessing time and GPU compute time to vary substantially across data-parallel ranks (training processes). In synchronous training, these two sources of heterogeneity interact tightly: GPU execution determines how long the input pipeline has to prepare the next batch, while delayed preprocessing in any rank in turn stalls all GPUs at synchronization. As a result, optimizing CPU throughput or GPU load balance in isolation cannot eliminate pipeline bubbles. We present DataHelix, a metadata-first CPU-GPU co-scheduling data plane for multimodal training. DataHelix jointly reasons about preprocessing demand and GPU execution progress, treating each GPU compute window as the readiness budget for the next batch and coordinating preprocessing across ranks to minimize deadline misses rather than CPU makespan. To avoid expensive data movement, placement is decided while samples remain lightweight descriptors, before decoding and collation amplify them into large tensors. Descriptor-level token-latent-aware batching performs coarse pre-balancing, while cost-gated collective rebalancing corrects residual skew only when its expected benefit exceeds transfer cost. Across 11 video and multimodal workloads on 8 Ascend 910B accelerators, DataHelix reduces mean deadline-miss rate by 67.8% and mean p95 input-wait latency by 87.0%, while achieving 2.72x geometric-mean throughput over rank-local DataLoaders and 1.82x over a Ray actor pool under the same CPU budget.

번역 서비스가 일시 중지되어 원문을 표시합니다.
arXiv cs.DC
model-evals

Decision Context: Persistent Behavioral Learning for AI Agents

arXiv:2611.08644v1 Announce Type: new Abstract: LLM agents do not learn from corrections across sessions. A user fixes the same mistake at the start of every conversation. Existing agent memory systems primarily handle facts, such as preferences, entities, and conversational history, rather than procedural knowledge. We present Forge Decision Context (forge_dc), a system that captures behavioral signals from human corrections and self-assessed tool-call failures, compiles them into imperative RULE directives, and injects them at the decision points where they apply. We evaluate forge_dc alongside four predecessor approaches on DC-Bench, a new benchmark of 71 scenarios across nine behavioral dimensions, built from real executed agent traces in a deterministic mocked-tool environment. forge_dc reaches 95% compliance, up from 55% with naive retrieval, while injecting approximately 10 times fewer tokens per query. Across the five-stage progression, the strongest systems represent behavioral memory as explicit directives and deliver it at the decision points where it applies. The results identify representation, activation, and placement as central design choices for persistent behavioral learning.

번역 서비스가 일시 중지되어 원문을 표시합니다.
arXiv cs.AI
coding-engineering

Decision models take 26% of requests; Anthropic now wins new teams 2:1

In November 2026, decision models accounted for 26% of all AI Gateway requests and 0.53% of total spend, with adoption highest among high-volume teams—70% of those processing over 10 billion tokens monthly used a decision model. Voice model usage rose 2.6 times since July, reaching 11% of paid teams. OpenAI and Anthropic now serve nearly equal shares of paid teams, but Anthropic gained more new teams (2:1), while OpenAI’s share of gateway spend increased by two-thirds since December. Over half of language-model requests went to models less than two months old, reflecting rapid model turnover with a median half-life of six weeks. In October, Anthropic’s Opus 5.5 became the top model by spend, and image and video model usage reached record levels, with OpenAI’s GPT-Image-2.5 Sunburst leading image spend and ByteDance’s Seedance 2.5 leading video spend.

번역 서비스가 일시 중지되어 원문을 표시합니다.
Vercel
copyright-privacy

Embedding Consistent Recurrent Neural Networks in Finite Element Simulations for Path-Dependent Damage Prediction

arXiv:2611.08494v1 Announce Type: new Abstract: Capturing sheet-metal localisation physics with a trained surrogate inside a finite element (FEM) solver requires the surrogate to deliver consistent predictions irrespective of the solver's strain-increment count, which is difficult to control in non-linear explicit codes. We address this by re-implementing a trained recurrent neural network (RNN) damage criterion in Fortran and embed it as a live, increment-by-increment fracture criterion inside an Abaqus/Explicit user material subroutine (VUMAT), advancing the network state alongside the solver. Two architectures are compared: a SimpleRNN and the proposed Consistent RNN (ConsRNN), whose transition function is designed so that predictions converge as the strain path is refined rather than drifting with the increment count. Both are trained on bilinear strain paths and evaluated under varying temporal discretisations and nonlinear histories. SimpleRNN predictions drift as the increment count increases, which disqualifies it for embedding; ConsRNN converges under refinement, at a substantial fixed-resolution accuracy cost relative to SimpleRNN on multilinear paths. Deployed at structural scale, the embedded ConsRNN surrogate governs element deletion in a clamped steel plate and reaches the same peak force and peak displacement as an established two-parameter fracture criterion. This pre-peak agreement is expected by construction, since no element is deleted before the peak and the two runs therefore share an identical constitutive path; the responses diverge only on the post-peak softening branch. The results establish discretisation consistency as the decisive property for embedding recurrent surrogates in explicit FEM solvers, and report the first embedding of a customised, non-standard-transition recurrent cell at structural scale.

번역 서비스가 일시 중지되어 원문을 표시합니다.
arXiv cs.LG
model-evals

Evolutionary Adaptive Metric Learning: A Multi-Objective Framework for Learning Local Distance Functions in Nearest Neighbor Classification

arXiv:2611.08549v1 Announce Type: new Abstract: Nearest neighbour classifiers inherit the geometry of their distance function. Fixed distances and globally learned metrics apply the same geometry throughout the input space, although relevant feature directions may vary across regions. We propose EAML (Evolutionary Adaptive Metric Learning), a framework that learns instance-dependent positive-definite metrics through a low-dimensional parameterization. Each local metric combines an input-dependent diagonal scaling, generated by softmax interpolation between anchor points, with a global low-rank correlation component. A real-coded elitist genetic algorithm optimizes the parameters using a scalar objective combining leave-one-out prediction error, robustness, metric smoothness, and regularization. We establish positive definiteness, metric equivalence, existence of an optimal solution, and Lipschitz continuity of the metric field. We further investigate prediction stability, generalization bounds, and evolutionary convergence under explicit assumptions. Experiments on synthetic data with region-dependent feature relevance show that EAML can recover local metric variations and outperform global metrics. Evaluation on thirteen public datasets using nested 5-fold cross-validation reveals a more nuanced picture: EAML improves performance on selected datasets but does not significantly outperform global metric baselines overall. Additional experiments show that global diagonal metrics can be preferable under nuisance features and test-time noise. These findings identify the conditions under which evolutionary local metric learning is beneficial and highlight its practical limitations.

번역 서비스가 일시 중지되어 원문을 표시합니다.
arXiv cs.LG