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

+0新增论文
+0其中 AI 相关
+0已有论文修订
+93新增 AI 导读

arXiv 最近完成扫描：2026-11-16 01:34 · 全站最近入库：2026-11-14 15:09各信源进度 →

统计口径与数据新鲜度

统计 2026-11-15 06:16 — 2026-11-16 06:16（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：2026-11-16 01:34最近发现新增版本：2026-11-14 15:09库内论文最新更新时间：2026-11-13 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。

23,348 已收录论文19,580 AI 相关6,729 有修订112 HF 社区发现

浏览论文库

研究日报 / 周报 →按信源浏览 →

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 23,348 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-16 01:33（北京时间）
版本更新 · v3arXiv:2606.15015

NEXUS: Neural Energy Fields for Physically Consistent Contact-Rich 3D Object Dynamics

Qizhen Ying, Guangming Wang, Yangchen Pan, Victor Adrian Prisacariu, Brian Sheil, Yixiong Jing

原文摘要

Physics-grounded video generation requires controllable 3D object dynamics that remain physically consistent under contact, deformation, and external forcing. Existing trajectory-based methods often model isolated physical effects, making it difficult to compose conservative and non-conservative dynamics in contact-rich 3D scenes. We present NEXUS, a neural energy-field framework for contact-rich 3D object dynamics. NEXUS represents each object as a structural graph and constructs dynamic object-object and object-environment contact graphs. Inspired by Hamiltonian Neural Networks, NEXUS formulates motion through scalar energy and dissipation terms rather than directly predicting states or accelerations. Conservative effects, including gravity and elastic deformation, are composed as additive energy terms, while non-conservative effects such as damping and impact-induced energy loss are modeled with learned Rayleigh-style dissipation. Forces are derived by differentiating the energy and dissipation functions and rolled out with a multi-substep semi-implicit integrator. Across controlled trajectory benchmarks, NEXUS improves long-horizon accuracy over representative learned and physics-structured dynamics baselines under varying mechanical properties and physical-effect compositions. We further show that NEXUS trajectories provide effective guidance for contact-rich video generation, improving physical plausibility while maintaining competitive visual quality.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2603.10055

Training Language Models via Neural Cellular Automata

Dan Lee, Seungwook Han, Akarsh Kumar, Pulkit Agrawal

原文摘要

Pre-training is crucial for large language models (LLMs), as it is when most representations and capabilities are acquired. However, natural language pre-training has problems: high-quality text is finite, contains human biases, and entangles knowledge with reasoning. This raises a fundamental question: is natural language the only path to intelligence? We propose using neural cellular automata (NCA) to generate synthetic, non-linguistic data for pre-pre-training LLMs--training on synthetic-then-natural language. NCA data exhibits rich spatiotemporal structure and statistics resembling natural language while being controllable and cheap to generate at scale. We find that pre-pre-training on only 164M NCA tokens improves downstream language modeling by up to 6% and accelerates convergence by up to 1.6x in 400M-3B models trained with Chinchilla optimal data budgets. Surprisingly, this even outperforms pre-pre-training on 1.6B tokens of high-quality natural language data with more compute. Investigating what drives transfer, we find that attention layers are the most transferable, and that optimal NCA complexity varies by domain: code benefits from simpler dynamics, while math and web text favor more complex ones. These results enable systematic tuning of the synthetic distribution to target domains. More broadly, our work opens a path toward more efficient models with fully synthetic pre-training.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.03951

Evolving LLM-Generated Features for Interpretable Classification

Jack Butler, Zainab Afolabi, Nikita Kozodoi

原文摘要

Large language models (LLMs) are increasingly used as classifiers, yet they operate as opaque systems whose decisions are difficult to interpret, which complicates their use in regulated domains such as credit scoring or medical diagnosis. We propose an evolutionary framework that iteratively discovers natural language feature definitions (rubrics) for interpretable classification. An LLM generates candidate binary features, evaluates each sample against them, and the resulting vectors can be used to train a transparent classifier such as logistic regression. The feature set evolves over multiple iterations guided by classification errors, per-class activation rates, and feature ablation scores. We evaluate across three benchmarks, including a credit risk dataset representative of regulated domains, comparing single-shot LLM rubrics, evolved rubrics, and direct zero-shot LLM classification. Evolved features improve over single-shot rubrics by +2.9 pp on average and outperform zero-shot LLM classification on two of three tasks, while providing fully auditable decision logic. On the credit risk task, the zero-shot LLM performs at chance (50.7%) with a strong bias toward a single class, whereas evolved features achieve balanced, interpretable predictions. Crucially, this failure is invisible in aggregate accuracy and surfaces only under per-class auditing. Analysis reveals that evolution is most effective when label boundaries cannot be inferred from category names alone or when the LLM lacks reliable domain-specific reasoning.

cs.LG · stat.ML

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.29621

CENDRe: Concept Extraction with Natural Domain Representations

Antonia Holzapfel, Andres Felipe Posada Moreno, Sebastian Trimpe

原文摘要

Convolutional neural networks (CNNs) are widely used for time-series classification, but their deployment in critical contexts requires understanding the patterns across signal domains that drive their predictions (e.g., temporal or spectral). Concept extraction (CE) methods identify such patterns by analyzing representations within the models' latent space. However, existing time-series CE methods have three limitations: they operate only in the time domain and cannot locate features in other domains, they depend on a predefined number of concepts that biases them, and they produce localizations misaligned with the regions the model uses. We address these limitations with CENDRe, a concept extraction method for CNNs. For concept discovery, we propose a procedure that clusters per-timestep latent representations in two stages, with a silhouette-guided aggregation that selects the number of concepts automatically. For localization, we define a presence score that contrasts the latent representations with their prototypes, and use its gradients to produce masks that concentrate on the regions driving the concept. These gradients, propagated through a differentiable invertible mapping of the input such as a Fourier transform, yield localizations for the same concepts in different domains. Finally, each concept receives a relevance score that quantifies its contribution to each class. On synthetic benchmarks, CENDRe achieves representation correctness comparable to state-of-the-art CE methods and higher importance correctness, while producing visualizations in different domains. On real bearing-fault data, CENDRe extracts the frequency bands driving the model's predictions, located in regions commonly inspected for fault diagnosis, producing evidence to assess the model that time-domain CE methods cannot.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.06486

Model Merging Does Not Require Uniform LoRA Ranks

Zirong Xie, Zainab Afolabi, Nikita Kozodoi, Jack Butler

原文摘要

Model merging combines separately fine-tuned experts into a single model without retraining. Standard practice trains every expert at the same LoRA rank, which couples the rank choice across experts even though each domain has its own best rank. We show this coupling is a property of the merge method, not of merging. We train 36 experts on three domains (mathematics, instruction following, multilingual) at four ranks and three model scales, then merge with six methods at uniform ranks and at each domain's own best rank. Giving experts different ranks costs averaging and utility-based selection nothing, and gains them up to 1.8 pp of multi-domain accuracy, while the conflict-resolution methods are inconsistent and can lose up to 4 pp. The split has a simple cause. Merging keeps only a fixed share of the pooled spectral components, and methods that choose them by size discard many that would have reduced loss. They also tend to discard the same expert: the instruction adapter has the smallest singular values, so it is dropped almost entirely and IFEval falls by 16 to 25 pp. Choosing components by their effect on loss instead keeps all three experts, and leaves accuracy nearly flat across an 8$\times$ rank range (0.7 pp, against 6.7 pp for TIES at 4B scale). Rank can therefore be chosen per expert whenever the merge operator does not select by magnitude.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.13062

Anchoring Clinical Events in Time: UID-Preserving Multimodal Reconstruction and Source-Grounded Adjudication

Sayantan Kumar, Nicolas Grimaldi, Jack Cummins, Jeremy C. Weiss

原文摘要

Interpreting a patient's hospital course depends on when symptoms developed, tests were performed, and treatments began. Discharge summaries often leave this chronology unclear, while structured electronic health record (EHR) tables timestamp only some of these events. We reconstruct clinical timelines by retrieving structured evidence for narrative events and jointly revising their estimated times. Each extracted occurrence carries a persistent unique identifier (UID) linked to its source span, keeping repeated mentions distinct throughout reconstruction. We also introduce GAVEL, a large language model (LLM) judge that checks disagreements between two timelines against the note and structured record and cites evidence for its verdicts. We evaluated six open-weight models on 40 clinician-annotated discharge summaries. Our pipeline recovered more reference events than LLM-Timeline and TKW2 when each comparison used the same backbone. For GLM-5.2, adding structured evidence improved timestamp agreement with essentially unchanged event recovery; ablations showed that links to timestamped source rows contributed to this improvement. Results varied across the other models, with some timing gains accompanied by fewer recovered events. Both GAVEL judges, GLM-5.2 and MiMo-V2.6-Pro-RL, favored multimodal over text-only GLM-5.2 timelines, while neither clearly distinguished the DeepSeek V3.2 variants. Manual review upheld most sampled GAVEL findings. The source links allow reconstructed times to be checked, and GAVEL identifies disagreements for human review.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.36457

Memory Consolidation Flattens the Temporal Shape of User Facts

Sugam Panthi, Muhaiminul Yeamin, Siyan Luo, Rabab Abdelfattah

原文摘要

LLM agents with long-term memory store conversations as short notes that later models read. A note can turn progressive "I am working at Corvida" into simple-present "The user works at Corvida," dropping the cue that the job may be temporary and that the stored fact may go stale. We call this aspectual flattening and measure it with LAPSE, a benchmark of matched user statements that differ only in temporal form. Three writer models flattened the progressive statement but kept its simple-present match in 210 of 381 pairs, never the reverse. This one-way loss held in all 11 models tested and in the agent memory systems mem0, Graphiti, and Letta. In exploratory tests, reader models judged a fact less likely to still hold when only its stored verb was progressive. In a planned test, the reader GLM-5.2 acted without asking the user on 64% of "works" notes but 40% of "is working" notes. Gemini 3 Flash showed a smaller gap, one reader added later showed the reverse, and others showed none. In a further exploratory test, free-text notes erased this difference and notes that kept the user's words restored it. Agent memory systems should keep the user's wording as evidence for later readers.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15737

Beyond the Diff - VulSpot: A Vulnerability Dataset Where No Vulnerability Goes Unlocalized

Dyna Soumhane Ouchebara, Stéphane Dupont

原文摘要

A vulnerability detector is only as trustworthy as the benchmark that scores it, and today's real-world datasets tend to flatter detectors. They are often built once and never refreshed, so they quickly miss new vulnerabilities, their labels are either noisy or heavily filtered, and their random splits let models train on near-copies of test code and on future vulnerabilities. Above all, when a fix only adds code, such as a missing NULL check, they do not point at where the vulnerability was, preventing detectors from learning to localize some of the most common C/C++ weaknesses. We present VulSpot, a living benchmark for detecting, classifying, and localizing vulnerabilities, with over 500k functions in 21 languages from 24k registered CVEs, including over 15k vulnerable C/C++ functions. By linking the code added by a patch to the existing statement it protects, VulSpot locates the vulnerability in 99.3\% of the functions that previous diff-based datasets used to leave without a location. Rather than discarding uncertain labels, it tells users how reliable each one is, and it comes with leak-free splits, complementary code context, and hierarchical CWE labels. An automated pipeline refreshes VulSpot every month, so that each release brings vulnerabilities no model has seen yet.

cs.CR · cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15735

Where Is Uncertainty Observed? A Mechanistic Understanding of Uncertainty Quantification in LLMs

Xiaoyu Liang, Virginia Ceccatelli, Aliai Eusebi, Tobias Braun, Nils Loose, Carlo Malagnino, Lorenzo Cavallaro

原文摘要

Knowing when a large language model's (LLM's) response can be trusted is critical for its safe use. Uncertainty quantification (UQ) can help LLMs recognize if their responses may be unreliable. Existing taxonomies organize UQ methods by uncertainty source or estimation mechanism. They leave unresolved the question of where uncertainty is observed in the model's computation. The answer matters because the observation point determines whether an estimator has access to enough information to recover a given target, and how well it can recover it. Drawing on mechanistic interpretability, we define four mechanistic observation lenses: the structural lens observes component updates, the representational lens observes hidden states, the distributional lens observes token probabilities, and the semantical lens observes distributions over meanings. We relate these lenses through the residual stream and order them by the information they carry, which gives a principle for selecting UQ methods for different targets. We survey existing methods through the four lenses. Furthermore, we distinguish evaluation targets, criteria, and metrics to clarify different UQ evaluation protocols. Finally, we identify open problems and research directions to connect uncertainty estimation to the internal computation and downstream use of LLMs.

cs.CL · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.13498

Evaluating Compositional Generalization in Transformers: The Role of Composition Equivalence and Module Coverage

Purva Pruthi, Andrew Yuan, Alexander D'Amour, David Jensen

原文摘要

Compositional generalization is an important form of out-of-distribution generalization, defined as the ability to train on some combinations of modules (e.g., functions such as sort, filter) and then generalize to unseen combinations. Despite substantial empirical work evaluating compositional generalization in transformers, findings have been inconsistent because evaluations were often conducted on a limited set of train-test splits, defined in an ad hoc manner. Through a systematic evaluation, we identify two key aspects of the train-test distribution that strongly affect performance: composition equivalence and module coverage. We show that the apparent performance of direct models (trained only on final outputs) results primarily from exploiting composition equivalences, meaning different sequences of modules that reduce to identical end-to-end functions. When benchmarks eliminate these equivalences, performance drops significantly, showing their inability to generalize to compositions of known modules that produce novel end-to-end functions. We further demonstrate that performance degrades monotonically as module coverage shifts increase (when modules appear in different contexts between train and test), affecting both direct and step-by-step models (trained on intermediate and final outputs). These findings have concrete implications for benchmark design: benchmarks that include compositional equivalences can lead to significant overestimation of the generalization abilities of direct models, and benchmarks that include both module coverage violations and compositional equivalences can exacerbate shortcut learning. As a case study, we demonstrate the role of compositional equivalence on the generalization performance of the pre-trained model Llama3-8B on the code-output prediction task in the coding benchmark LiveCodeBench.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15732

Right Advice, Wrong Reasons: How Decision and Reasoning Alignment Shape Reliance on AI Advisors

Eunkyu Park, Vasudha Varadarajan, Wesley Hanwen Deng, Anisha Reddy, Gunhee Kim, Motahhare Eslami, Jana Schaich Borg, Maarten Sap

原文摘要

When people seek advice from AI on social situations, the system may recommend a decision they fully agree with but justify it through reasoning they disagree with. Conversely, it may offer reasoning they agree with, and yet reach a decision they reject. We distinguish decision alignment, whether an advisor recommends the option favored by a user's measured values, from reason alignment, whether its stated reasons align with the user's values. In this work, we ask how these two forms of alignment shape people's preferences for AI advisors and their willingness to use them in different roles. We conducted a within-subjects study (N=121) in which participants evaluated three AI advisors varying in decision and reason alignment across eight everyday moral dilemmas. Participants then chose which advisors they would consult, allow to draft messages for review, or allow to act on their behalf in new tasks. When seeking advice, participants most often preferred an advisor that reasoned through their prioritized values while recommending the opposing decision (40%). However, when participants had to pick advisors for new tasks, their preferences shifted toward advisors that also recommended decisions aligned with their values, rather than those that only shared their reasons. Also, as participants considered tasks that gave AI advisors greater autonomy, they became more likely to opt out of using any advisor, from 12% when consulting to 38% when allowing an advisor to act on their behalf. Our findings suggest that decision and reason alignment shape advisor preferences differently across contexts of AI use, and that alignment with a user's values does not necessarily imply willingness to delegate.

cs.HC

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2608.15247

Improved Metric Distortion Bounds for Deterministic Weighted-Tournament Voting Rules

Jianan Lin, Chenhao Wang

原文摘要

We study deterministic weighted-tournament voting rules, also known as C2 rules, which use only the fraction of voters who prefer one candidate to another for each pair of candidates. Voters and candidates lie in a common but unknown metric space, and voters rank candidates by distance. The goal is to select a candidate minimizing the sum of distances to the voters, and distortion measures the worst-case approximation ratio relative to an optimal candidate. Charikar et al.~[EC 2025] introduced the Unblanketed Set rule with distortion at most $3.9312$ and proved a lower bound of $3.1128$ for deterministic C2 rules. We extend their rule to the Path-Unblanketed Set rule, which runs in polynomial time and achieves distortion at most $1+2\sqrt{2}\approx3.8284$. The rule uses paths between candidates to combine pairwise comparisons without increasing the approximation loss with the path length. With computer assistance, we prove a better upper bound of $3.3346$ for five or six candidates and a general lower bound of $3.1828$ using a construction with 17 candidates.

cs.GT

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.27378

PanDent: Toward Comprehensive Tooth-Level Structure-Language Consistency in Dental Radiology

Xiaohan Li, Xinyu Liu, Chang Liu, Kuofeng Hung, Sum Wing Au Yeung, Jun Liu, Yixuan Yuan, Hui Chen

原文摘要

Accurate evaluation of multimodal large language models in dental panoramic radiography is limited by the lack of fine-grained, clinically reliable benchmarks that reflect expert interpretation. This work introduces PanDent, a large-scale benchmark built on expert-validated tooth-level annotations. It comprises 9,554 panoramic radiographs with comprehensive structured findings produced by experienced dentists and further validated by an oral and maxillofacial radiologist. Clinician-guided reports are constructed from these findings using clinician-defined reporting logic, establishing explicit correspondence between structured clinical evidence and free-text descriptions. The dataset supports complementary evaluation of report-level textual agreement and tooth-level clinical consistency, with 9,049 public-source cases for supervised training and an independent 505-case in-house benchmark. Experiments cover proprietary, general-domain open-source, and medical-specific models. Current models generate fluent reports but exhibit substantial errors in fine-grained localization and tooth-level diagnosis. Supervised fine-tuning on PanDent markedly improves report alignment and diagnostic accuracy, bringing predictions closer to expert dental interpretation. These results establish PanDent as a rigorous benchmark and training resource for clinically grounded dental multimodal reasoning. Dataset sample: https://huggingface.co/datasets/Desperado1103/Pandent_Sample; code: https://github.com/HKUDentistry/PanDent. Full data will be released upon publication under source redistribution terms.

cs.CV · cs.MM

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15729

DISCERN: Preserve at Write, Resolve at Read for Persistent LLM-Agent Memory

Chirag Ahuja, Vladislav Tanov, Ramcharan Kakarla, Jonathan Taws

原文摘要

Enterprise agents accumulate memory across dozens of sessions in which facts contradict, get superseded, and blur as similarly named entities collide in embedding space. The field treats this as a retrieval problem; we show it is not. On ENT-310, a synthetic enterprise sales-agent memory benchmark of 310 queries (210 answerable, 100 unanswerable), a vanilla Retrieval-Augmented Generation (RAG) reader over a memory store already matches strong memory systems on the recall pillars. Retention and currency are near-saturated, roughly 0.93 to 0.98 for every competent system, so recall is essentially solved. What remains unsolved is resolution: which entity a query refers to, which time-state is current, and whether to commit an answer at all or decline when evidence is insufficient. We propose DISCERN (DISambiguation and Cross-session Episodic memoRy Navigation), an agent-memory framework built on a single rule: preserve at write, resolve at read. At write time DISCERN builds a strictly additive bi-temporal provenance graph that never merges confusable entities and never drops superseded values. At read time a zero-LLM tool surface drives an agentic loop that navigates this graph and commits exactly one of four gated terminal actions. A four-judge LLM-as-a-Judge council scores two axes: answer correctness on answerable questions, and the answer-versus-abstain decision as a classification problem. On ENT-310, DISCERN is the best-calibrated system on abstention (F1 0.924 versus at most 0.802 across eight baselines) and leads macro answer correctness under equal pillar weighting (0.861 versus 0.806 for the strongest baseline), a lead driven by entity resolution (0.800, where no baseline exceeds 0.40). Applied with zero tuning to the full LoCoMo corpus, the same pipeline again leads macro answer correctness (0.663 versus 0.638) and stays among the best-calibrated on false-premise abstention.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.12140

Visible Tails: Characterizing Adaptation to Unknown Moments in Heavy-Tailed Bandits

Tuan Ngoc Do, Hai Duc Nguyen

原文摘要

Heavy-tailed bandit learners must tell a poor arm from an arm whose favorable rewards have not yet appeared. We characterize this obstruction through a scale of one-sided visibility conditions on the optimal reward law: the mean hidden in a rare event is at most a constant multiple of a retained central moment. A quantile-trimmed UCB index that is given the visibility constants attains arm-specific moment rates at every exponent, with no lower bound on the exponent. Conversely, every strategy with a fixed exploration budget has polynomial regret on each law that violates the corresponding condition. For the variance benchmark we determine the order of both budget constants in the visibility level. The frontier depends on the largest exponent the benchmark promises, and the resulting classes are strictly nested. The same index has logarithmic expected regret when rewards are integrable, with a coefficient that can be exponential in the inverse gap.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15722

Reliable Near-Field Source Enumeration and Positioning Informed by Two-Stage MUSIC

Jiaying Li, Haifeng Wen, Changsheng You, Yuanwei Liu, Hong Xing

原文摘要

Accurate and reliable near-field multi-source localization is essential for location-aware processing in future wireless systems. However, coherent multipath propagation complicates the direct recovery of line-of-sight (LoS) source positions, while existing estimation methods generally lack explicit control of the risk of source underestimation and positioning uncertainty. In this paper, we investigate reliable source enumeration and positioning in near-field coherent multipath environments. Specifically, we develop a conformal risk control (CRC) based source enumeration scheme that calibrates the penalty coefficient of an eigenvalue-based criterion to control the expected miss-detection ratio at a prescribed level. We further propose MUSIC-Net, a model-based deep learning framework informed by two-stage MUltiple SIgnal Classification (MUSIC), which learns the LoS-related subspace and surrogate distances through subspace-distance-based training, thereby enabling direct source positioning without estimating non-line-of-sight (NLoS) path parameters or performing subsequent path association. To quantify the uncertainty of the estimated positions, we develop a split conformal prediction (SCP) based calibration method that constructs angle-range predicted regions with finite-sample simultaneous coverage guarantees. Numerical results show that the proposed framework effectively controls the risk of source underestimation, while providing improved positioning accuracy with tighter predicted regions than benchmarks given the same target coverage levels.

eess.SP · cs.IT · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.06504

NaijaFacts: A Framework for Disentangling Language from Knowledge in Facts about Nigeria

Zainab Afolabi, Nikita Kozodoi, Jack Butler, Babs Khalidson, Sabastine Dekaa Francis, Christiana I. Ezeilo, Olosunde Akinlolu Adesey

原文摘要

Multilingual benchmarks for African languages mostly translate generic questions into local languages, measuring linguistic competence but not whether models know facts about Africa. We introduce NaijaFacts, a benchmark of factual and statistical questions about Nigeria, authored by native speakers and grounded in authoritative official statistics, with every question available in English, Nigerian Pidgin, Yorùbá, Igbo, and Hausa. Answering similar questions for the UK and US yields a country by language matrix that disentangles a language effect (the model knows the fact but degrades in a low-resource language) from a knowledge effect (the fact is absent in every language). Across eight frontier and Nigeria-adapted models, the knowledge effect dominates: English-query accuracy is 49-50% for the US/UK but 20% for Nigeria. The language effect is smaller and tracks web representation, degrading from English to Yorùbá (9 points on Nigeria, up to 24 on US/UK questions). We find scale does not help: the strongest models reach 66-79% on the UK and the US but none exceed 31% on Nigerian statistics. Extending to all 54 African countries with our best performing frontier model shows the same profile continent-wide, 96% on culture and 98% on politics against 26% on statistics, so the gap is not specific to Nigeria.

cs.CL · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15720

Terms of Care: Designing for Participation in Data Governance by Disability Communities

Madiha Zahrah Choksi, Reeda Shimaz Huda, Rita Faia Marques, Anja Thieme, Cecily Morrison

原文摘要

Improving disability representation in AI-generated media requires community-specific datasets, yet openly sharing such data can expose communities to misuse, decontextualization, and exploitation. We examine how disability communities can participate in dataset governance through context-sensitive licensing. Working with three disability advocacy organizations, we conducted semi-structured interviews, an educational webinar and modular license activity to promote community participation in data stewardship. Our findings show that governance in these high-context communities relies on organizational memory (especially situated harms); both formal authority and informal relational practices; communication; and resulting trust. The modular license activity made organizational data choices more accessible and values-led, allowing communities to adapt Creative Commons clauses to express permitted uses, sector-specific restrictions, safeguarding requirements, and commitments to respectful representation. Yet, concerns remain about monitoring and enforcing appropriate data use. We contribute (1) an empirical account of governance tensions, (2) a practical modular approach, and (3) design implications for AI data platforms that support relational, context-sensitive governance beyond standardized, one-size-fits-all frameworks.

cs.HC · cs.AI · cs.DL

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2604.17693

COSAC: Counterfactual Credit Assignment in Sequential Cooperative Teams

Shripad V. Deshmukh, Jayakumar Subramanian, Raghavendra Addanki, Nikos Vlassis

原文摘要

In cooperative teams where agents act in a fixed order and share a single team-level reward (multi-agent language systems, sequential robotic tasks), per-agent credit assignment is under-determined. Critic-based approaches scale poorly as the number of agents grows owing to the costly maintenance of joint/factored critic(s), whereas the existing critic-free alternatives have other issues: common credit across agents that couples every agent's signal to teammate noise, importance-sampling corrections for upstream-update staleness that incur variance exponential in team size, or per-agent counterfactual replay that isolates each agent's effect at the price of extra environment or reward calls. We propose COSAC, a critic-free policy gradient framework for sequential cooperative teams. COSAC fits an additive per-agent decomposition of the team reward by a single ridge regression on the rollout batch (giving each agent a learning signal decoupled from teammate noise), and computes each agent's counterfactual advantage from fictitious continuations of the current policy (policy forward passes that replace both importance-sampling reweighting and per-agent environment replay, at no extra environment or reward cost). The estimator instantiates the Sequential Aristocrat Utility (SeqAU), our extension of Wolpert and Tumer's (2001) aristocrat utility to sequential teams. We derive bias and conditional variance bounds for SeqAU credits: the direct-effect bound has no explicit team-size factor, while the indirect effect's reward-noise bound scales with the number of downstream agents. Sequential bandits test the analysis across team sizes up to K=16. In sequential multi-agent LLM fine-tuning, COSAC reduces training regret by 17% on consensus-based science reasoning relative to the best critic-free multi-agent baseline and by 61% on Python coding relative to MA-GRPO.

cs.LG · cs.AI · cs.MA

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15718

Orbit: An Evidence-Grounded Benchmark for Evaluating LLMs on IT Operations Reasoning

Sathwik Narkedimilli, Rakshit Verma, Rahul Soni, Ankit Kushwaha, Balaji Ganesan, Kaustabha Ray, Sunil Saumya, Pratibha Moogi

原文摘要

LLMs now support IT operations (ITOps), but benchmarks rarely show why a model fails. We present ORBIT, a benchmark of ITOps questions in three formats, namely multiple-choice, open-ended, and troubleshooting, with each question accompanied by a reference answer and a recorded source. An audit found template defects in about a tenth of the generated items, so we drop them and rebuild the troubleshooting format from real practitioner threads with accepted answers. Under one independent judge, five LLMs answer most multiple-choice items correctly, between a third and just over half of the open-ended items, and a minority of the troubleshooting items, even under a more lenient root-cause rubric. When models judged their own answers, they accepted many answers to defective items, so we use one independent judge for every model. Word overlap with the question finds the correct option for more than half of the multiple-choice items. ORBIT is a diagnostic aid for selecting and supervising assistants rather than a deployment test.

cs.LG · cs.SE

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.11163

LILA: Data-Free Structured Pruning of LLMs via Latent Spectral Geometry

Sankar Behera, Dhruv Singh, Anshika Agnihotri, Raj Kumar Choudhary, Satyadev Ahlawat, Yamuna Prasad

原文摘要

Structured pruning of large language models (LLMs) has universally required calibration data, gradient computation, or large auxiliary policy networks at pruning time. LILA (\emph{Latent-Informed Layer Analysis}) demonstrates that this dependency is avoidable: each FFN weight matrix is factorized with Non-negative Matrix Factorization (NMF) to obtain a parts-based spectral representation, and neuron importance is scored via the Kolmogorov--Smirnov (KS) distance between singular value distributions of the full and neuron-ablated NMF matrix, a closed-form criterion requiring no calibration corpus, forward pass, or gradient computation. Two data-free matrix variants (Abs, Split) generate pruning masks from weight spectral geometry alone; a calibration-guided variant (Act) is provided as an internal ablation. Evaluation spans 7 model families (up to 70B). Without fine-tuning, data-free LILA surpasses PruneNet (45M-parameter RL policy) by up to 2.12~pp in zero-shot accuracy on LLaMA-2-7B and outperforms WikiText-2-calibrated SliceGPT by up to 6.1~pp across all sparsity levels, confirming data-free performance is not an artifact of omitting activation statistics. After one LoRA RFT epoch (matched WikiText-2 protocol), LILA Act surpasses SliceGPT+RFT by 6.39~pp on average using 98.4\% fewer calibration samples (16 vs.\ 1,024) across LLaMA-2-7B and Phi-2. NTK analysis confirms a $17{\times}$ reduction in functional distortion versus random pruning, providing theoretical grounding for the spectral criterion.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v6arXiv:2009.03996

Combining Determinism and Indeterminism

Michael Stephen Fiske

原文摘要

Our goal is to construct mathematical operations that combine indeterminism measured from quantum randomness with computational determinism so that non-mechanistic behavior is preserved in the computation. Formally, some results about operations applied to computably enumerable (c.e.) and bi-immune sets are proven here, where the objective is for the operations to preserve bi-immunity. While developing rearrangement operations on the natural numbers $\mathbb{N}$, we discovered that the bi-immune rearrangements generate an uncountable subgroup of the infinite symmetric group (Sym$(\mathbb{N})$) on the natural numbers. This new uncountable subgroup is called the bi-immune symmetric group. The bi-immune symmetric group contains the finitary symmetric group on the natural numbers, and consequently is highly transitive. Furthermore, the bi-immune symmetric group is dense in Sym$(\mathbb{N})$ with respect to the pointwise convergence topology. Moreover, the bi-immune symmetric group contains permutations that are a single infinite cycle on all of $(\mathbb{N})$, even though every one of its generators has only finite orbits. Using an invariant called the {\it crossing number}, we also prove that the bi-immune symmetric group is a meager subset and a proper subgroup of Sym$(\mathbb{N})$.

math.LO · cs.CC · cs.CL · math.GR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15717

Outside the Purview of the Reward: Capability Drift in Multilingual Post-Training

Sutanshu Saksena Raj, Purusharth Malik, Pasunuti Prasanjith, Daksh Maan, Parijat Chatterjee, Avinash Benki

原文摘要

Post-training can improve rewarded behavior while weakening capabilities users still need. We study two GRPO branches from a shared supervised checkpoint in a multilingual mixture-of-experts model. The released branch gains 36.07 points on MATH but loses 19.32 F1 points on XQuAD-Indic, while XQuAD-English improves under the same scoring rule. An independently trained sibling adds Indic instruction following and language-aware shaping. It recovers the regressing Indic task scores and preserves MATH, but does not restore script compliance and trades off competition mathematics and code. In a later round, judge training is followed by a sharp extraction decline from the Indic branch; reward-free distillation from its earlier checkpoint restores most of the lost F1. A MATH score decomposition associates most of the measured gain with increased boxed-answer emission; accuracy among boxed answers changes little. These results show why benchmark scores should be read alongside the response forms they reward.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15716

A Reliability Study of Explainable AI for Freight Mode Classification

Ali Alawi Balfaqih, Mansur M. Arief

原文摘要

Machine learning models are increasingly used to support freight transportation decisions, but their predictions are rarely accompanied by evidence that the associated explanations can be trusted. In this study we investigate explainable machine learning for classifying shipments as Air/Air Charter or Truck/Ocean using the PEPFAR Supply Chain Management System Delivery History dataset, with 5,964 shipment records and 47 input features after cleaning, where freight cost and pack price are excluded as near-proxies of the transportation mode. We compare three Multi-Layer Perceptron (MLP) architectures, Logistic Regression, Random Forest, and gradient boosting with imbalance-aware metrics and paired significance tests across one hundred random replications. We then evaluate the explanations themselves, namely the stability of SHAP feature rankings across retraining and across models, the faithfulness of SHAP and LIME attributions under a feature-deletion test, and the geographic structure of the destination effects, which we map for all 38 destination countries. The results suggest that the tree ensembles significantly outperform the neural and linear models (86.8% against 83.3-85.0% mean testing accuracy) while their leading features remain the most stable across retraining. Removing the five features ranked most important by SHAP or LIME changes the predicted probabilities roughly four times as much as removing randomly chosen features. The destination effects form a coherent geographic pattern that simple port access does not explain. However, models with similar accuracy disagree substantially on which features matter most, and thus explanation stability and faithfulness should be assessed before such explanations are used to support freight transport decisions.

cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2507.21135

Quantum Geometry of Data

Alexander G. Abanov, Luca Candelori, Harold C. Steinacker, Martin T. Wells, Jerome R. Busemeyer, Cameron J. Hogan, Vahagn Kirakosyan, Nicola Marzari 等 12 位作者

原文摘要

We demonstrate how Quantum Cognition Machine Learning (QCML) encodes data as quantum geometry. In QCML, features of the data are represented by learned Hermitian matrices, and data points are mapped to states in Hilbert space. The quantum geometry description endows the dataset with rich geometric and topological structure---including intrinsic dimension, quantum metric, and Berry curvature---derived directly from the data. QCML captures global properties of data through a compact operator representation. At fixed Hilbert-space dimension, the number of operator parameters grows only linearly with the number of features. The present work provides the first full exposition of QCML as a framework for matrix geometry, establishing its mathematical foundation and demonstrating how operator-level tools from quantum geometry can be directly applied to data. We introduce the matrix Laplacian and its eigenmaps as a geometry-preserving alternative to graph-based embeddings, and we show how topological invariants such as Chern numbers can be extracted from data for the first time. An important practical outcome is the efficiency of the representation: the relevant geometric structure of high-dimensional datasets can be captured with Hilbert spaces of size as small as $N \sim 10$. We illustrate these results on a number of synthetic and real-world examples, including the Wisconsin Breast Cancer dataset, where QCML uncovers geometric separability between benign and malignant samples and yields an interpretable decomposition of feature relevance. The emphasis of this work is on a new geometric representation of data rather than on algorithmic benchmarking, establishing a foundation for extracting global geometric and topological structure from high-dimensional datasets.

cs.LG · quant-ph · stat.ML

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15711

Beyond Worst-Case Online Algorithms via Input Subsampling

Siddhartha Banerjee, Tianle Jiang, Kamesh Munagala

原文摘要

We study whether classical competitive ratios survive input subsampling, when both the online algorithm and the offline optimum are evaluated on the same subsampled input. We consider the canonical problems of prophet inequalities and ski-rental. In these settings, for any single subsampling rate, the classical worst-case ratios remain unchanged. We ask a stronger question: can one adversarial input remain worst-case at two distinct rates $p>q$? For prophet inequalities, in the model where the policy knows the sampling rate and learns item survival online, every instance has competitive ratio strictly below $2$ at one of the two rates, and we determine the tight simultaneous guarantee. For ski rental, in the fluid model where sampling scales the horizon, the answer depends on $q/p$. Against adversarial distributions over horizons, one unrestricted randomized algorithm achieves a strict simultaneous improvement over $e/(e-1)$ for every distribution exactly when $q/p$ is below approximately $0.34$. Requiring the algorithm to buy by the breakeven time lowers this threshold to approximately $0.31$, showing that unlike classical ski-rental algorithms, buying later can provably help in this setting. Our results therefore reveal a new form of beyond-worst-case fragility under input subsampling: instances that are worst-case at one rate need not remain hard at two rates simultaneously.

cs.DS · cs.GT

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15710

Curiosity-0826-4B: Generative Medical Event Sequence Modeling Across Pediatric and Adult Populations

Mani Nagarathnam, Luke Benham, Paul Blazek, John Cashy, Kevin Chung, John Muenning, Hank Voeller, Shane Waxler 等 17 位作者

原文摘要

We introduce Curiosity-0826-4B, an updated version of the generative medical event foundation model trained on Epic Cosmos that simulates patient timelines to produce zero-shot clinical predictions. Curiosity-0826-4B incorporates a broader patient population and richer clinical event representations, while generally matching or exceeding Curiosity-0226-1B, the largest model in the previous Curiosity-0226 series, on evaluated tasks supported by both models. Laboratory value prediction improved, with MAE 18--46% lower than that of Curiosity-0226-1B. We evaluated prediction of disease events across pediatric cohorts aged 2--17 years. AUROCs for new asthma diagnosis (0.742--0.785) and asthma exacerbation (0.829--0.852) were comparable to those in the adult reference cohort, while AUROCs for sickle cell crisis (0.820--0.847) were lower than in adults (0.915). Explicit representation of documented death enabled new mortality prediction tasks among adults across hospice, inpatient, and ambulatory settings (AUROC 0.821--0.942). Scaling-law analysis guided compute-optimal initial pretraining followed by continued pretraining with context extension. The resulting Curiosity-0826-4B model has more parameters (4.3B versus 1B) and a longer context window (16K versus 8K). It processes 56% fewer training tokens and requires 2.2 times the training floating point operations (FLOPs) compared with its predecessor. By extending a common zero-shot prediction strategy to pediatric patients and additional outcomes, Curiosity-0826-4B broadens opportunities to develop clinical and operational applications without training separate task-specific models. Further validation should assess local calibration, performance across patient subgroups and event frequencies, timeline completion rates, and generalization to new settings, alongside prospective evaluation of clinical and operational utility.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15707

Efficient Knowledge Transfer in Federated Bayesian Optimization through Neural Network Surrogates

Alexander Gräfe, Max van Gemmeren, Paul Brunzema, Sebastian Trimpe

原文摘要

Federated Bayesian Optimization (FedBO) has emerged as a powerful paradigm for enhancing Bayesian Optimization (BO) in distributed, privacy-sensitive settings, enabling agents to accelerate the optimization of their local objectives by sharing knowledge about their similarities without exposing raw data. However, existing methods rely on restrictive assumptions, such as near-identical objective functions or shared Gaussian process hyperparameters that are often violated in heterogeneous real-world applications. We present GLOBAFED, a novel framework that enables efficient knowledge transfer across different but related objectives without compromising privacy. At its core is a probabilistic neural network surrogate with a shared feature-extracting backbone and agent-specific heads equipped with varia- tional Bayesian last layers to provide uncertainty estimates. During optimization, agents use federated learning on the shared backbone to collaboratively build a common latent feature representation, while agent-specific heads adapt to their individual objectives. Our empirical results demonstrate that GLOBAFED significantly accelerates convergence and reaches better final performance compared to state-of-the-art FedBO baselines, and enables agents to warm-start their optimization from previously learned global features.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.15705

Characterizing Predictive Multiplicity in Ranking and Its Fairness Implications

Ainhize Barrainkua, Mikel Malagón

原文摘要

Prediction problems often admit competing models that achieve comparable performance yet produce conflicting outputs, a phenomenon known as predictive multiplicity. While this challenge has been extensively studied in classification tasks, its implications remain largely unexplored in ranking frameworks, which are critical for high-stakes societal decisions such as the allocation of welfare resources and social assistance. This paper investigates the phenomenon of predictive multiplicity in score-based ranking problems. We introduce formal definitions to capture multiplicity in ranking settings, supported by foundational theoretical guarantees. Furthermore, we examine the dual relationship between predictive multiplicity and algorithmic fairness in ranking problems: we analyze how model multiplicity can act as a potential source of unfairness, while demonstrating how exploring this decision space provides a principled avenue for discovering less discriminatory rankings. Finally, we complement our theoretical contributions with insightful empirical evaluations on both synthetic and real-world datasets, validating our analytical findings and illustrating their practical implications. Taken together, this work highlights the real-world relevance of predictive multiplicity in ranking and sheds light on the potential of leveraging this effect to achieve fairer outcomes.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.05479

Privilege and confidentiality in generative AI workflows

Václav Janeček, Thomas Melham

原文摘要

Generative AI (GenAI) systems store and process client data in three distinct ways: in the model's parameters through training and memorisation, in the context window during a live session, and in knowledge databases for retrieval-augmented generation (RAG). Each mode creates different and often counter-intuitive risks to confidentiality and legal professional privilege, and each calls for specific governance responses. Drawing on the first English and American decisions to address privilege and generative AI, UK and Munir v Secretary of State for the Home Department and United States v Heppner, on the orthodox privilege authorities against which those decisions must be read, and on recent computer science research, we explain the three modes of data storage and processing in terms accessible to practitioners and analyse the legal consequences of each. We then situate the analysis within the regulatory framework governing solicitors in England and Wales and within the ordinary principles of professional negligence, arguing that the standard of effective information governance (and with it the benchmark against which negligence and misconduct will be measured) is changing. Although we write primarily for SRA-regulated practitioners, our data-governance analysis is framed to extend to any jurisdiction in which the protection of privilege or professional secrecy depends on demonstrable confidentiality. The ultimate aim of this article is to help legal services professionals understand salient data leakage risks in GenAI systems and thereby facilitate a more responsible deployment of GenAI on client data and other sensitive material.

cs.CR · cs.AI

查看论文
arXiv ↗PDF ↗