跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+1112新增论文
+875其中 AI 相关
+133已有论文修订
+94新增 AI 导读

arXiv 最近完成扫描:2026-11-19 19:10 · 全站最近入库:2026-11-19 19:09各信源进度 →

统计口径与数据新鲜度

统计 2026-11-18 19:47 — 2026-11-19 19:47(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-19 19:10最近发现新增版本:2026-11-19 19:09库内论文最新更新时间:2026-11-19 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

27,982 已收录论文23,249 AI 相关8,559 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 20,663 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-19 19:07(北京时间)
版本更新 · v3arXiv:2602.03864

Have Large Language Models Enhanced the Way Civil & Environmental Engineers Write? A Quantitative Analysis of Scholarly Communication over 25 Years

Morgan D. Sanger, Brett W. Maurer

原文摘要

Large language models (LLMs) have rapidly emerged in civil and environmental engineering (CEE) research, education, and practice as tools for project ideation, execution, and communication. However, a gap remains in understanding how prevalent LLM adoption is across CEE scholarship and whether it measurably alters research prose. This study applies a vocabulary-based frequency-shift methodology to 149,452 abstracts published by the American Society of Civil Engineers between 2000 and 2025. Deviations from long-term vocabulary trends suggest that 15.3% and 26.9% of 2024 and 2025 abstracts, respectively, are likely LLM-influenced. Prior to the introduction of LLMs in 2022, CEE publications exhibit long-term trends toward longer abstracts, greater use of segmenting punctuation, higher required reading levels, and use of active, first-person verb constructions. Beginning around 2023, however, the frequencies of many stylistic marker words (e.g., enhance) sharply depart from their historical trajectories, a shift consistent with the emergence of LLMs. Abstracts classified as likely LLM-influenced also depart from historical semantic trends, exhibiting increased lexical diversity, comma use, and complexity, with reduced passive voice and hedging language, producing prose that is more segmented, complex, and confident. The artificial intelligence contribution of this study is the use of natural language processing to detect population-level linguistic signals of LLM-generated text, estimating the prevalence of LLM use and its influence on CEE writing. Together, these novel findings provide the first large-scale, data-driven assessment of how CEE scholarly communication is changing alongside LLM adoption.

cs.DL

新论文 · v1arXiv:2611.21319

Wavelet-Energy-Driven Dynamic Patching for Time Series Forecasting

Syed Hafiz Bin Salim Alsree, Stephen Roberts, Stefan Zohren

原文摘要

Patch-based transformers tokenise a time series into fixed-size, uniform windows, which commits the model to a single temporal resolution even though real signals are multi-scale. The discrete wavelet transform (DWT) is a natural alternative because it decomposes a signal across scales. A growing body of work makes the wavelet filter learnable, on the assumption that a more flexible wavelet forecasts better. That assumption has rarely been tested in isolation, because a single reported improvement usually credits three things at once. These are the use of any multi-resolution decomposition, the choice of wavelet and training the filter. We introduce a forecaster that turns a learnable-DWT per-scale energy map into variable-length patches by merging adjacent windows that share a dominant scale. Three findings follow. First, most of the improvement is there before any filter is learned. A frozen wavelet beats a fixed-patch baseline on eight of the nine datasets by 1.5% to 7.8%, and on four of them it recovers all of the gain. Traffic is the one exception, where it is 1.2% worse. Second, learning the filter gives a further improvement on only some datasets. It helps on three of the nine datasets and makes Traffic 2.7% worse. Third, having variable-length patches helps, but exactly where their boundaries fall does not. A fixed-size uniform grid, even with its patch size tuned per task, is worse than our variable-length patches on 27 of 36 settings. Recomputing the boundaries from the learned filter at every step, rather than caching them once, never helps. The improvement in wavelet-based forecasting comes from the multi-resolution structure and the variable-length patches it induces, not from making the wavelet learnable. In practice, start from any valid wavelet, cache the patch boundaries once and treat training the filter as a per-dataset decision rather than a default.

cs.LG

新论文 · v1arXiv:2611.21705

Disambiguating Moral Circuits and Moral Features in Language Models: Moral content, normative judgment or cue words? Evidence from 14 models

Davide Liga, Liuwen Yu, Matteo Magnini

原文摘要

Interpretability methods can locate the attention heads and sparse autoencoder (SAE) features that change the answer a language model gives to a moral prompt, e.g., whether "Stealing medicine for a dying child is morally" continues with "justified" or "wrong". Such a component could respond to moral content, to the judgment of conduct by any norm, or to the word "morally" in the prompt. We built 227 minimal pairs of short phrases, validated by five annotators, whose two versions differ in a moral standard, in another kind of norm (prudence, etiquette, aesthetics or evidence) or in a descriptive quality, and showed them to 14 open-weight models (Llama, Gemma and Qwen, 1B to 14B, base and instruction-tuned), with and without the word "morally". In every model, the heads that change the moral answer belong to an evaluation circuit the models use for every kind of normative judgment we tested: these heads carry the same share of the effect for moral as for other normative pairs (pooled ratio 1.00, 97.5% confidence interval [0.92, 1.07], with the word; 0.90, [0.84, 0.96], without it), and the models need them as much for both. The models need these heads also when nothing in the prompt asks for a judgment, and on 400 scenarios of the ETHICS benchmark, where removing them takes away 37 to 39% of the difference between acceptable and unacceptable scenarios, against 1 to 3% for ten random heads. The SAE features we located respond mostly to the word "morally" rather than to what the phrase is about. The models represent the category of a pair (the standard the annotators assigned to it): it can be read from their hidden states, and, in the eight models we tested, removing the direction that carries it weakens the moral verdict by a few percent, and more than the other verdicts in seven of them. We release the pairs, the ratings, the code, the trained SAE dictionaries and the result files.

cs.CL · cs.AI · cs.LG

新论文 · v1arXiv:2611.21312

PointZip: Statistically Lossless Compression for Climate Data with Pointwise Error Bounds

Jiayong Li, Pratyai Mazumder, Yuejiang Yu, Langwen Huang, Alexandru Calotoiu, Torsten Hoefler

原文摘要

High-resolution climate simulations generate data volumes that increasingly strain storage, I/O, and data dissemination. Error-bounded lossy compressors can reduce these costs, but their tolerances are usually user-chosen and difficult to standardize across heterogeneous climate datasets. We address this limitation with statistically lossless compression, where the pointwise compression error is bounded by the pointwise data uncertainty. We propose PointZip, a two-stage compressor that combines a rate-distortion-efficient base layer with error-bound-aware residual coding to enforce pointwise guarantees. On outputs from the widely used climate modeling framework ICON, PointZip makes 1km simulations practical by reducing storage bandwidth requirements by 72x. On the ERA5 climate dataset, PointZip reaches 69.3x on the most representative variables, reducing the commonly used subset to 4 TB. In downstream evaluations, PointZip preserves data spectra, extreme events, and air parcel trajectories, while increasing neural weather forecast RMSE by less than 1%, with negligible impact on the Fractions Skill Score (FSS).

cs.CE

新论文 · v1arXiv:2611.21310

Can AI Mature a Model's Geometry? Training-Free Weight Baking in LLMs

Amin Golnari, Saeid Sanei

原文摘要

Can an AI make a trained language model more mature, in the sense of a better organized latent space, by looking only at the geometry of that space? An analyst model reads a fixed-size digest of topological descriptors of the layers of a frozen target and proposes a transformation, which closed-form ridge regression writes into the feed-forward weights without gradient training. Using lightweight models, 126 runs show that category purity rose on unseen words and sentences, benchmark accuracy stayed within sampling error, and the analyst matched a category-aware heuristic with gentler edits. In a five-model round robin approach, 16 pairs raised purity and four left the target unchanged. The results provide a baseline for future work on LLM-guided, training-free model editing without using raw weights.

cs.LG · cs.AI

新论文 · v1arXiv:2611.21309

Safe Physical Interaction of All-Electric Heavy-Duty Robotic Manipulators via Control Barrier Functions and Variable Impedance Control

Amir Hossein Barjini, Mohammad Bahari, Jouni Mattila

原文摘要

All-electric heavy-duty robotic manipulators are increasingly expected to operate close to tunnel walls, equipment, and other structures while retaining the force capability required for large-payload tasks. Accurate motion control alone does not address the resulting safety and interaction requirements. This paper develops a safety-aware physical-interaction architecture that combines variable impedance control (VIC) with control barrier functions (CBFs) on top of an established virtual decomposition control (VDC) framework for electromechanical linear actuator (EMLA)-driven manipulators. VIC continuously adapts the task-space stiffness and damping according to predicted proximity and contact risk, whereas a joint-velocity CBF filter enforces Cartesian workspace boundaries and the physical stroke limits induced by the closed-chain lift and tilt mechanisms. To account for nonideal realization of the filtered velocity by the inner VDC/EMLA dynamics, the CBF inequalities are tightened using a measured bound on the joint-velocity tracking error. Four configurations, namely VDC, VIC+VDC, CBF+VDC, and VIC+CBF+VDC, are evaluated under workspace, actuator-limit, and physical-interaction scenarios. VIC reduces the peak and RMS environment-interaction forces by 92.59% and 96.33%, respectively; both robust CBF configurations maintain positive workspace and EMLA safety margins, while VIC+CBF+VDC reduces the CBF intervention metric by 10.30% compared with CBF+VDC.

cs.RO

版本更新 · v5arXiv:2605.08541

Tokens-per-Parameter Coverage Is Critical for Robust LLM Scaling Law Extrapolation

Joshua Shay Kricheli, Alexander Lawrence Reid, Soumajyoti Sarkar, Venkata Gandikota, Paulo Shakarian

原文摘要

Neural scaling laws approximate a language model's loss as a power-law function of parameter count $N$ and token count $D$. Following Chinchilla-style compute-optimal training, many studies fit scaling laws from runs at one or a few fixed tokens-per-parameter (TPP) ratios $k$, with $D = kN$. We show that this collinear design, combined with the empirically common near-equality of the exponents governing $N$ and $D$, induces an inherent ill-conditioning in the Gauss-Newton least-squares problem: the condition number of the design grows as the inverse square of the gap between the $N$- and $D$-exponents. The scale coefficients become practically unidentifiable, with confidence intervals inflating by orders of magnitude, yielding a "sloppy" model whose extrapolations degrade sharply off the training ray. We prove this for four scaling-law formalisms and derive a closed-form TPP-diversity threshold that characterizes well-conditioned estimation to leading order. Empirically, in controlled experiments with 5M-77M-parameter models, non-collinear designs outperform collinear ones on held-out data with a 97.3% win rate across four laws and five corpora, and 98.7% in bfloat16. We further show the degeneracy is rooted in Jacobian geometry and is not an artifact of the loss function: any smooth objective whose Gauss-Newton weights are bounded above and away from zero inherits the same ill-conditioning.

cs.LG

新论文 · v1arXiv:2611.21307

Programmable Semantic Interventions: Learning Causal Carriers from Behavioral Programs

Paolo Muratore, Mackenzie Weygandt Mathis

原文摘要

Causal abstraction explains a frozen network by aligning its activations with the variables of a simpler, interpretable model, and checking that interventions on the two agree. Finding such an alignment currently requires a counterfactual dataset: for every pair of prompts, the answer the network ought to give once the variable has been changed. In practice, those labels are produced by an executable algorithm for the behavior, written before studying the model supposed to implement it. However, the criterion these methods optimize only leverages the answer an intervention should produce. Here, we develop an approach that lets the investigator define that relationship directly: the answers the variable can produce, and which of them an intervention should yield. We call this interface a semantic program and introduce Programmable Semantic Interventions (PSI): a low-rank activation subspace learned so that interchanging it changes the model's answer as the program declares, with the expected answers read from the frozen model itself. On the MIB benchmark PSI performs on par with supervised alignment at matched data and training budget, without curated pairs, and it needs no ground-truth labels; while on natural science questions with gold labels withheld, it tracks the model's own answers better than a subspace trained on gold labels. The declaration also fixes a coordinate system with known semantics, enabling donor-free steering: jointly training PSI for both localization and control realizes 94.8% of held-out requests for answers the investigator chooses, without the need for donor prompts.

cs.LG · cs.AI

新论文 · v1arXiv:2611.21305

High-probability Uniform Mixing via Dobrushin Contraction and Accelerated Hamiltonian Monte Carlo

Stefan Oberdörster

原文摘要

Hamiltonian Monte Carlo, since its invention and subsequent rise to widespread use, has been motivated by its ability to accelerate convergence to equilibrium compared to diffusive methods, both in terms of dependence on condition number and dimension. In the present work, we establish this acceleration simultaneously in both aspects for critically damped Metropolis-adjusted kinetic Langevin with persistent uniforms. More precisely, for target measures $μ(dx)\propto e^{-U(x)}dx$ with $m$-strongly convex, $L$-gradient Lipschitz, and $L_H$-Frobenius Hessian Lipschitz potentials, we show that the method mixes to $\varepsilon$-accuracy in total variation using $\mathcal O\bigl((κ^{1/2}+L_H^{1/2}m^{-3/4})\,d^{1/4}\bigr)$ gradient evaluations up to logarithmic factors, both uniformly over starting points in domains of high equilibrium probability and from warm starts. The proof incorporates ideas from previous studies by Bou-Rabee and Chen et al. that show either aspect of acceleration separately, into novel coupling technology. This new coupling technology for general Markov chains extends earlier frameworks aimed at capturing fast local equilibration in the presence of slow global mixing. Specifically, via a weak-$(1,1)$ maximal inequality that follows from Hopf's maximal ergodic theorem, we uncover a stability mechanism in high-probability domains with respect to equilibrium. This enables straightforward localization to naturally arising domains of fast local mixing, and seamlessly fits broad local frameworks via Dobrushin contraction, Wasserstein contraction and local Harris' theorems. We develop the Dobrushin formulation.

math.PR · cs.LG · stat.CO · stat.ML

版本更新 · v2arXiv:2607.23798

Interpolation via Generalized Splitting

Lutz Straßburger

原文摘要

We propose a new proof theoretical method for proving Lyndon interpolation. Our proof does not use the sequent calculus but is based on a generalization of the splitting lemma in deep inference. We then formulate the interpolation theorem as a decomposition of a derivation into an up-fragment and a down-fragment. This can be seen as (i) a strengthening of the standard formulation of the interpolation theorem, and (ii) a generalization of the cut elimination theorem. We demonstrate the flexibility of our approach by applying it to linear logic, classical logic, and modal logics. For this, we also introduce novel cut-free proof systems for several modal logics in deep inference.

cs.LO

新论文 · v1arXiv:2611.21304

Fix, Don't Retake: Physics-Simulated Training for Correcting Patient Positioning Errors in Dental Panoramic Radiographs

Sema Helali, Lina Abu Nada, Faleh Tamimi, Rafat Damseh

原文摘要

Patient-positioning errors can distort panoramic radiographs and lead to repeat examinations. This study investigated artificial intelligence for correcting these errors in existing images. A physics-based simulator generated paired correctly positioned and mispositioned panoramics from cone-beam computed tomography volumes, with controlled individual and combined positioning errors. Five approaches were compared: U-Net, pix2pix, conditional diffusion, an error-conditioned model, and a mixture of error-specific specialists. Evaluation used image similarity and geometric metrics across three test sets. All approaches improved mean image similarity and reduced geometric errors, with U-Net and the two error-aware approaches outperforming pix2pix and conditional diffusion.

cs.CV · cs.AI

新论文 · v1arXiv:2611.21302

Déjà-Events: Multi-view 3D Geometry with Event Cameras via Sparsity-Aware Adaptive Gating

Arda Cinar Demirtas, Pablo Varela Bernal, Marc Pollefeys, Sunghwan Hong

原文摘要

Event cameras keep capturing structure where conventional cameras blur or saturate, which makes them attractive for 3D perception in fast or poorly lit scenes. Event datasets with dense 3D ground truth, however, are far smaller and less diverse than the image collections behind feed-forward 3D reconstruction models. We present Déjà-Events, an event-only model that predicts depth, camera pose and pointmaps from multiple views in a single forward pass. Rather than relying on scarce 3D labels, we distil it from Déjà View, a feed-forward RGB model that refines its tokens by repeatedly applying one shared transformer block. On paired RGB and event recordings, the event student learns to reproduce the frozen teacher's final tokens, which requires no depth or pose ground truth and trains only about 1% of its parameters. Because events are dense at edges and moving texture but nearly absent in textureless regions, we further introduce sparsity-aware adaptive gating, which adapts the shared block's gates to events and adjusts them per patch by event density, so that accuracy holds as refinement steps are added. Trained only on simulated data, Déjà-Events achieves the lowest pointmap error among the event-based methods we evaluate, on EventScape and zero-shot on the real MVSEC and DSEC datasets. On DSEC it halves the error of the strongest baseline (0.15 vs. 0.30 AbsRel) with about a tenth of its parameters. It also has the lowest depth error on EventScape and DSEC and the lowest trajectory error on EventScape.

cs.CV · cs.RO

新论文 · v1arXiv:2611.21300

LATTE-WordSeg-Lite: Multi-Granularity Thai Word Segmentation on a CPU, with a Reproducible Benchmark

Thodsaporn Chay-intr, Kobkrit Viriyayudhakorn, Thanaruk Theeramunkong

原文摘要

Large language models read Thai through subword tokens, without a word segmenter. Yet Thai treebanks, speech metrics and morphological resources are built on words, often at several granularities, and word boundaries resolve ambiguities that subword tokens leave open. LATTE (Lattice ATTentive Encoding) reads a text at several granularities at once, through a lattice of dictionary words, but it is a large GPU model. We distil it into LATTE-WordSeg-Lite, a CPU segmenter of 1.17M parameters that keeps the lattice as features. It outputs compounds, words and morphemes in one pass, and learns only from annotated data and text whose licences allow commercial use. To compare it with other tools, published Thai scores do not suffice, as they come from different files, splits and scorers. We therefore rebuild seven test sets from pinned sources and score twelve systems under one protocol. With the unit that fits each test set, LATTE-WordSeg-Lite scores above DeepCut on five sets, four of them from corpora that its teachers learnt from. Called line by line on one CPU thread, it runs 23 times as fast as DeepCut. We release the model, the benchmark scripts and all predictions.

cs.CL

新论文 · v1arXiv:2611.21301

Latent Bridge: Asking for Help Without Words

Qichao Wang, Tingyu Qu, Pengcheng Wu, Yiran Zhong, Feida Zhu, Zhiqi Shen, Peilin Zhao, Steven Hoi

原文摘要

Text-based model collaboration requires a large model to generate intermediate responses before a smaller model can use its guidance, adding large-model decoding cost to inference. We introduce POND (Prefill-Only, No-Decode), a method that uses a frozen large sender for prefill and a small receiver for all autoregressive decoding. A trainable bridge maps the sender's hidden states to additional attention keys and values, allowing the receiver to use these representations throughout generation. POND also supports periodic review, in which the sender processes the receiver's partial response and updates its guidance through additional prefill passes while preserving the receiver's existing KV cache. Across reasoning and coding benchmarks, POND improves receiver performance with senders from different model families. In particular, GLM-5.3 provides gains comparable to those of Qwen3.8-Max when paired with a Qwen receiver, supporting the effectiveness of representation transfer between independently trained model families. POND further demonstrates its ability to generalize across domains. When trained on math and coding tasks, it transfers directly to general knowledge tasks without additional domain-specific training. These results demonstrate that large frontier model prefill representations can improve small-model generation without requiring large-model autoregressive decoding.

cs.CL

版本更新 · v2arXiv:2611.11461

Sharp Rates for Optimal Recovery of Bandlimited Functions from Localized Oversampling

Wentao Huang, Haizhang Zhang

原文摘要

It is well-known that one can recover a function $f$ bandlimited to $[-δ,δ]$ ($0<δ<π$) from its finite oversampling data $f(j),-n+1\le j\le n$, with an approximation error exponentially decaying to zero as $n$ tends to infinity. Various regularized Shannon sampling formulas have been proposed to achieve such exponential approximation. A recent sinh regularized formula achieved so far the best rate $$ O(\frac1{\sqrt{n}}\exp(-(π-δ)n)). $$ We establish that the optimal recovery algorithm in the worst-case scenario achieves an approximation error $ρ_{n,δ}$ satisfying $$ \frac{c_*}{\sqrt{n}}(\tan\fracδ4)^{2n+1/2}\le ρ_{n,δ}\le \frac{C_*}{\sqrt{n}}(\tan\fracδ4)^{2n+1/2}, 0<δ\leqπ, n\geq1, $$ where $c_*,C_*$ are explicit positive constants independent of $n$ and $δ$. This result identifies the best possible exponential rate to reconstruct bandlimited functions from two-sided finite oversampling data. In other words, $$ ρ_{n,δ} \asymp \frac{\sqrt{\tan\fracδ4}}{\sqrt{n}}\exp(-c_δ(π-δ)n), $$ where $c_δ$ is strictly decreasing on $(0,π]$ and $$ \lim_{δ\toπ^-}c_δ=1,\ \lim_{δ\to 0^+}c_δ=+\infty. $$ For one-sided sample data $f(j)$, $-n+1\le j\le 0$, we also obtain the sharp approximation order of optimally recovering function values on $[0,1]$: $$ ρ_{n,δ}^-\asymp \sqrt{δ/π}\sin^n(δ/2),\ 0<δ<π,\ n\ge1, $$ which improves the celebrated result of H. J. Landau on extrapolation of bandlimited functions.

cs.IT · math.CA

新论文 · v1arXiv:2611.21298

Agent-Native by Design: A Programming Language for Agentic Applications

Jayanaka Dantanarayana, Savini Kashmira, Xiaoyu Ma, Malitha, Malitha Prabhashana, Lingjia Tang, Jason Mars

原文摘要

Building agentic applications requires combining large language model (LLM) operations, workflow structure and application state, and model-guided routing. With SDKs and agent frameworks, developers connect these components through supporting code that prepares model inputs, propagates state, and dispatches model-selected steps. Workflow changes can require coordinated updates to this code, increasing development and testing effort. Splitting execution logic between application code and framework internals also limits compiler visibility into control flow and state dependencies, hindering automatic optimization. We introduce Agent-Native Programming, which unifies these three components with shared execution semantics. We realize this approach in Jac, a Python-superset language, through "by llm()" for typed model operations and tool use, Object-Spatial Programming for workflow structure and state, and "visit ... by llm()" for model-guided routing. The routing construct derives the model's choices from executable workflow steps and schedules the selected steps. This explicit connection between model decisions and program execution gives the compiler and runtime a common basis for dependency analysis and automatic optimization. Across five benchmarks and five baselines, Jac achieves a 1.8-8.7x reduction in AI-integration code relative to the OpenAI SDK, with competitive observed task quality and lower coding-agent development effort. Two case studies demonstrate speculative execution and service partitioning through configuration alone, using program analysis and runtime checks to preserve workflow behavior under stated conditions.

cs.PL · cs.AI · cs.SE

版本更新 · v3arXiv:2510.13818

denet, A lightweight command-line tool for process monitoring in benchmarking and beyond

Ben Carrillo, Izaskun Mallona

原文摘要

Summary: denet is a lightweight process monitoring tool providing real-time resource profiling of running processes. It reports CPU, memory, disk I/O, network activity, and thread usage, including recursive child monitoring, with adaptive sampling rates. denet exposes two interfaces: a CLI with colorized output and a Python API. Output format is JSONL. Each record carries performance metrics and process metadata (PID and the executed command). Structured output makes denet suitable for benchmarking and tuning data-intensive pipelines in bioinformatics and beyond. We provide CLI and API examples, including a bioinformatics workflow in Snakemake, showcasing denet's diagnostic properties. Availability and implementation: denet is open-source software released under the GPLv3 terms, and maintained at https://github.com/btraven00/denet. It is implemented in Rust, with Python bindings provided via maturin, and installs from Cargo (cargo install denet), PyPI (pip install denet) or conda-forge. Most functionality does not require administrative privileges, enabling use on cloud platforms, HPC clusters, and standard Linux workstations.

cs.PF

版本更新 · v2arXiv:2610.20824

Wolof-Adapted Industry-Ready Open-Weight Language Models

Gwénolé Lecorvé, Delphine Charlet, Pierre Adam, Floriane Behanzin, Lionel Delphin-Poulat, Johannes Heinecke, Géraldine Damnati, Pelin Kaytmaz 等 18 位作者

原文摘要

Wolof is spoken by 23M people, mostly in Senegal, but remains weakly supported by open-weight language models, while API models raise cost, governance, and sovereignty constraints for local deployments. In this paper, we present 0.6B-14B open-weight models adapted to Wolof through instruction tuning, preference alignment, and model merging. After presenting the pipeline and data, extensive experiments on 63~public and in-house tasks show good performance on Wolof understanding and generation while remaining effective on other languages and cross-lingual skills, with the 8B and 14B models being particularly competitive with substantially larger commercial systems. The models are currently being evaluated in internal customer-care pilots and are publicly available at https://huggingface.co/collections/Orange/wolof-lm .

cs.CL · cs.AI

新论文 · v1arXiv:2611.21295

Deep Learning for Physicists: From Linear Regression to Reasoning Models

Masanori Hanada

原文摘要

These lecture notes are based on an introductory course on neural networks and deep learning that the author taught in the School of Mathematical Sciences at Queen Mary University of London in 2023/24 and 2024/25. The author, however, is a physicist, and it shows: wherever possible, ideas are explained with physical intuition rather than heavy formalism. Readers with a background in physics or mathematics who want a self-contained, introductory route into modern machine learning may therefore find these notes particularly friendly. The notes start from maximum likelihood, least squares, linear regression and (stochastic) gradient descent. They then cover neural networks, backpropagation, the vanishing gradient problem, batch and layer normalization, overfitting and regularization, adversarial examples, and the kernel method. The second half turns to natural language processing: word embeddings, recurrent networks and LSTMs, the attention mechanism and Transformers, and large language models, including scaling laws, the history of GPT, instruction tuning, reinforcement learning from human feedback, hallucination, alignment faking, chain-of-thought reasoning and test-time scaling. The emphasis on large language models reflects requests from the students. The notes are aimed at third- and fourth-year students in the UK system. Because the course was also taken by students from other departments, very little is assumed: even basic material such as matrix multiplication is explained from scratch, so readers familiar with it can safely skip such parts. No originality is claimed; the aim is a coherent, intuitive path through the material. Developments after spring 2025 are not covered, but the basic ideas emphasized here are unlikely to go out of date soon.

cs.LG · hep-lat · hep-th · physics.comp-ph

版本更新 · v4arXiv:2303.16880

Storage and Learning phase transitions in the Random-Features Hopfield Model

Matteo Negri, Clarissa Lauditi, Gabriele Perugini, Carlo Lucibello, Enrico M. Malatesta

原文摘要

The Hopfield model is a paradigmatic model of neural networks that has been analyzed for many decades in the statistical physics, neuroscience, and machine learning communities. Inspired by the manifold hypothesis in machine learning, we propose and investigate a generalization of the standard setting that we name Random-Features Hopfield Model. Here $P$ binary patterns of length $N$ are generated by applying to Gaussian vectors sampled in a latent space of dimension $D$ a random projection followed by a non-linearity. Using the replica method from statistical physics, we derive the phase diagram of the model in the limit $P,N,D\to\infty$ with fixed ratios $α=P/N$ and $α_D=D/N$. Besides the usual retrieval phase, where the patterns can be dynamically recovered from some initial corruption, we uncover a new phase where the features characterizing the projection can be recovered instead. We call this phenomena the learning phase transition, as the features are not explicitly given to the model but rather are inferred from the patterns in an unsupervised fashion.

cond-mat.dis-nn · cs.LG

新论文 · v1arXiv:2611.21291

Benchmarking Large Language Models for Linguistic Annotation

Ondřej Tichý, Anna Marklová, Barbora Bulantová, Magdalena Titlbachová, Jiří Milička

原文摘要

Assessing large language models for linguistic annotation requires task-specific evaluation under documented conditions. We present dhAIBench, a Python benchmarking and annotation tool with a web interface for exploring and sharing results. It records prompts, configurations and predictions, enabling comparisons of annotation quality, confidence calibration, repeatability and estimated cost. We demonstrate the framework and the associated methodology through 293 runs of 48 models on nine linguistic annotation tasks spanning morphology, syntax, semantics, discourse-pragmatics, spelling normalization and learner-language correction. Models achieve high agreement with reference annotations on several tasks, including those on non-standardized, historical varieties as well as requiring contextual interpretation. The results support evaluating models against specific linguistic tasks, rather than relying on general rankings. To this end, we make the current and future benchmarks publicly available at dhaibench.silent3.ff.cuni.cz.

cs.CL · cs.AI

新论文 · v1arXiv:2611.21288

GNoF: A Low-SM-Overhead GPU-Native NVMe-over-Fabrics Completion Path for Data-Intensive GPU Workloads

Lei Tian, Wenhao Gu

原文摘要

GPUs are the main compute engine for data-intensive workloads whose datasets exceed device memory and reside on remote flash. The conventional host CPU orchestrates every NVMe-over-Fabrics (NVMe-oF) transaction and stages data through host memory, costing high CPU usage, repeated CPU-GPU synchronization, and redundant data movement. GPU-initiated storage removes the CPU from the I/O path, but existing designs leave two problems: synchronous lock-step completion pins streaming multiprocessors (SMs) in spin-polling, wasteful over a 10-100 microsecond RDMA round trip, and current GPU-native remote systems focus on target-side firmware while using per-command doorbells and periodic CQ polling. We present GNoF, a GPU-native NVMe-oF/RDMA data path where GPU threads build command capsules, post send work-queue elements, ring doorbells, and reap completions entirely on the device; the CPU only handles one-time connection setup, the admin queue, and a keep-alive thread. Three mechanisms target long fabric round trips: a doorbell mover that coalesces a burst's doorbells into one MMIO ring; a batched CQ poller with a lock-free fast peek and monotonic completion counters that cannot lose a wakeup; and a set-associative, reference-pinned GPU page cache with split fetch/wait pipelining and orphan-fill takeover. GNoF needs no target or firmware change and runs against an unmodified SPDK target. On a 200 Gb/s InfiniBand fabric, GNoF reaches 1.77 million 4 KB IOPS per client at 0.37% host CPU and about 1% SM occupancy, versus 40% CPU for kernel nvme-rdma and about 100% SM occupancy for lock-step polling. Against three existing baselines (kernel nvme-rdma, BaM over a BlueField-3 DPU, and BaM on a local SSD) under identical workloads, we show 10-22x gains on selective column I/O, CPU-burn immunity, and the regime where bulk staging is faster.

cs.DC · cs.AR

新论文 · v1arXiv:2611.21287

Connecting Extended Modal Dependence Logics

Johan van Benthem, Chenwei Shi, Wenfei Ouyang

原文摘要

We consider three natural extensions of the basic decidable dependence logic LFD, adding, respectively, conditional dependence atoms, independence atoms, and identity atoms between variables. We show that these systems are all connected to first-order logic with identity by SAT reductions and in several cases also faithful translations. In particular, this implies that all these extended modal-style dependence logics are undecidable. This solves a number of problems left open in the original publication [1] introducing the LFD framework. The definitions and proofs presented here were found by ChatGPT and checked and streamlined as needed by the authors. At the end, we provide some discussion of what new things we learnt about dependence logics and about the methodology of using AI in logical research.

cs.LO

新论文 · v1arXiv:2611.21285

The Price of Not Planning: Parallel Sampling Depth Peaks at the Reconstruction Threshold

Chuling Wen, Weijie Liang, Jian Lu

原文摘要

Masked diffusion models sample the tokens of one round independently. Planned decoders first generate a coarse plan, then fill in the tokens in parallel. We ask when a plan saves rounds, how many it saves, and how fine it must be. We study this on broadcast trees: the leaves are the tokens, the internal nodes are the plan, and the edge correlation $θ$ sets how strongly tokens depend on the plan. Errors are measured by the forward Kullback-Leibler divergence $\varepsilon$, and the lower bounds hold for every deterministic value-adaptive rule, also with learned conditionals. We show that a decoder without the plan needs of order $\min\{n,(n/\varepsilon)^α\}$ rounds below the reconstruction threshold, where $α=\log b/(4\log(1/θ))$ rises to $1/2$ at the threshold, and that the exponent stays below $1/2$ above it (binary channel: $0.02\leθ\le0.98$). With the plan, $\log_b n+1$ rounds suffice. In the binary model below the threshold, a plan with units of $m$ tokens needs of order $\log_b(n/m)+\min\{m,(n/\varepsilon)^α\}$ rounds, so it gives logarithmic depth exactly when $m$ is logarithmic. A plan therefore helps most near the threshold, and only if its units are small. Exact divergence computations in the Gaussian model with up to $2^{20}$ leaves confirm these exponents below the threshold.

cs.LG

新论文 · v1arXiv:2611.21283

PLECTA: Overlap-aware instance segmentation of dense strand networks from binary masks

Oday Allan, Igor Ostanin

原文摘要

Many materials and tissues, including carbon-nanotube sheets, collagen and paper, comprise overlapping strands whose arrangement influences their effective properties. Measuring this geometry from micrographs requires individual strands to be traced through crossings and signal gaps, which merge distinct strands or split one strand into disconnected fragments. We present PLECTA (from Latin plecta, a braid), a deterministic method that segments individual strands using only a binary image whose foreground pixels mark strand centrelines. At each junction, PLECTA selects which strand segments continue into one another while allowing segments to terminate. It formulates this decision as an exact maximum-weight partial-matching problem, with an explicit penalty for unpaired segments and no restriction on the number of segments entering a junction. The output permits crossing pixels to belong to multiple strands. On synthetic scenes, PLECTA resolved strand identities more accurately than three released alternatives at every density tested. The same ranking held on an independently published generator run at its own settings. A separate stage uses the corresponding greyscale micrograph to estimate which strand lies on top at each crossing. The complete workflow is demonstrated on scanning electron micrographs of carbon nanotube sheets, from raw images to individual strand instances with estimated crossing order, without manual annotations for training or as pipeline inputs.

cs.CV · cond-mat.soft

新论文 · v1arXiv:2611.21702

Watermark Evidence under Open-Weight LLM Modification: Exact Bounds for Low-Rank Output Watermarks

Prashant Shekhar, Dumindu Samaraweera, Caroline Howard

原文摘要

Open-weight language models can continue generating after their released weights are modified, while a watermark verifier still scores outputs against the original base and marked models. We study how such modifications change the evidence available to an OpenStamp-style verifier. At each shared prefix, we represent the modification as a residual in the next-token logits. We derive an exact finite-vocabulary lower bound on the expected log-likelihood-ratio evidence under a residual-span constraint, along with a sharp span-only deviation bound. We also characterize the first-order evidence shift through a covariance term with an explicit second-order remainder, and determine the residual span needed to reach a target evidence level. For sequence-level detection, we derive an exact conditional exponential-moment bound and combine it with a supermartingale argument and a distribution-free order-statistic tolerance bound. This yields a high-confidence upper bound on miss probability after attack selection on calibration data and evaluation on independent certification continuations. In experiments with OpenStamp on Llama-2-7B and Mistral-7B-v0.3, INT8 quantization and 30% pruning retain test true-positive rates of 85.6-94.8%, while 500-step LoRA fine-tuning lowers them to 28.8% and 11.6%. The covariance prediction tracks the measured mean LoRA evidence drift on both models, with no remainder-bound violations across 160,000 test prefixes. Additionally, we observe zero false positives among 500 base-model continuations per model. Through these theoretical and empirical results, the current research provides a better understanding of watermark robustness to model-editing/tuning attacks.

cs.CR · cs.AI · cs.CL

版本更新 · v3arXiv:2608.14400

Ten simple rules for non-visual, reproducible and accessible bioinformatics

Jacqueline G. Kientsch, Stephan C. F. Neuhauss, Izaskun Mallona

原文摘要

Bioinformatics workflows rely heavily on visual representations. Quality-control plots, cell embeddings, heatmaps, genome-browser tracks, and interactive dashboards are not merely illustrations, but instruments for making analytical decisions. For blind and low-vision researchers who use screen readers, braille displays, or audio-based interfaces, these create a barrier: the evidence used to justify an analysis is often encoded in visual form, while the underlying decision remains undocumented. We argue that non-visual accessibility and computational reproducibility are closely aligned, as they both require analyses to be transparent and to record why decisions were made. We present ten simple rules for non-visual bioinformatics, covering plots as decision records, cautious use of AI-generated figure descriptions, accessible computing environments, text-first literate programming, structured data and metadata, compact object summaries, accessible publication formats, collaboration practices, shared community infrastructure, and accessibility as part of FAIR research. The intended audience is computational biologists and developers. Using single-cell RNA-seq as a running example, we show that the accessible equivalent of a plot is a structured decision record. That is, a plot companion that goes beyond storing the underlying data by also stating the purpose of the analysis and the resulting quantitative evidence and uncertainty. We argue that treating accessibility in this way makes bioinformatics more inclusive and also more transparent and auditable.

q-bio.GN · cs.HC

版本更新 · v2arXiv:2611.09387

A Bridge from Paste to Place: Physics-Guided One-Step Image Composition via Object-Scene Interaction

Hongzhen Li, Youwei Pan, Yingfang Zhu, Sen Zhu, Fengjie Zhu, Leilei Cao

原文摘要

Realistic image composition makes a pasted object appear naturally present in a scene while preserving the pasted layout and fine foreground and background details. Reference-based insertion methods preserve overall object identity but can alter pose and fine details, often at substantial computational cost. Foreground harmonization and shadow generation address individual aspects of this task, but neither alone ensures natural integration. An efficient, unified solution remains an open challenge, and large-scale, high-quality datasets and benchmarks for this task are lacking. To address these challenges, we introduce Paste2Place, an end-to-end, physics-guided flow-based model that learns a direct latent bridge from pasted inputs to realistic composites, guided by image-space corrections analytically derived from predicted physical components of object and scene appearance changes. To support training and evaluation, we introduce P2P-Data with 86K manually reviewed synthetic and real-scene training pairs, including dense physical annotations from our Blender rendering and validation pipeline for the synthetic subset. P2P-Bench combines paired compositions and manually pasted inputs from collected images into 1.5K examples for evaluating composition under user-chosen placements. Extensive qualitative and quantitative experiments demonstrate that Paste2Place achieves substantial advantages in image composition quality and inference speed across diverse scenarios.

cs.CV

版本更新 · v2arXiv:2602.05846

Optimal scaling laws in learning hierarchical multi-index models

Leonardo Defilippis, Florent Krzakala, Bruno Loureiro, Antoine Maillard

原文摘要

In this work, we provide a sharp theory of scaling laws for two-layer neural networks trained on a class of hierarchical multi-index targets, in a genuinely representation-limited regime. We derive exact information-theoretic scaling laws for subspace recovery and prediction error, revealing how the hierarchical features of the target are sequentially learned through a cascade of phase transitions. We further show that these optimal rates are achieved by a simple, target-agnostic spectral estimator, which can be interpreted as the small learning-rate limit of gradient descent on the first-layer weights. Once an adapted representation is identified, the readout can be learned statistically optimally, using an efficient procedure. As a consequence, we provide a unified and rigorous explanation of scaling laws, plateau phenomena, and spectral structure in shallow neural networks trained on such hierarchical targets.

stat.ML · cs.LG

版本更新 · v2arXiv:2611.14783

A Multi-Prover Interactive Protocol for Certification of Resolution Refutations

Philipp Czerner, Adrian Krauss, Josia Pietsch, Javier Esparza

原文摘要

A resolution refutation of a SAT-formula $\varphi$ is an independently checkable certificate proving that $\varphi$ is unsatisfiable. However, the refutations of $\varphi$ can be exponentially larger than $\varphi$. So, a verifier needs worst-case exponential time in the size of $\varphi$ to check their correctness. We present ResCertify, an interactive probabilistic protocol in which the verifier interacts with multiple independent, untrusted provers (MIP) to check the refutation with very low probability of error. The verifier runs in linear time in the size of $\varphi$. The provers require time $\mathcal{O}(n \cdot |Φ|)$, where $Φ$ is the resolution refutation and $n$ is the number of variables of $\varphi$. If they can be trusted to act as a multilinear oracle on certain queries, a single prover suffices, and the runtime can be improved to $\mathcal{O}(|Φ|)$. Finally, while the focus of our work is theoretical, we also report on a proof-of-concept implementation of ResCertify.

cs.LO