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

+479新增论文
+361其中 AI 相关
+63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:2026-11-16 13:44各信源进度 →

统计口径与数据新鲜度

统计 2026-11-15 18:42 — 2026-11-16 18:42(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:2026-11-16 13:44库内论文最新更新时间:2026-11-14 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。

23,827 已收录论文19,941 AI 相关6,921 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 23,827 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)
新论文 · v1arXiv:2611.12767

李群上轨迹跟踪的带认证环带可达性分析

Certified Zonotopic Reachability Analysis for Trajectory Tracking on Lie Groups

Yanliang Huang, Peng Xie, Zhen Zhang, Zhuoqi Zeng, Amr Alanwar

AI 速读 基于摘要

针对李群上轨迹跟踪的误差动力学,用环带(zonotope)传播线性部分并闭式包络逆右雅可比对扰动的影响,得到连续时间下可证明包含全部可达跟踪误差的紧致管道。

eess.SY · cs.RO

查看原文摘要

Autonomous vehicles track planned trajectories close to obstacles. Unknown disturbances make them deviate from these trajectories. With a bound on the tracking error that disturbances can cause, a planned trajectory can be certified as safe before execution. Existing set-based methods bound the tracking error by one invariant ellipsoid for the whole trajectory or by intervals, which are both conservative. For aerial and space vehicles, the tracking error evolves on a matrix Lie group with nonlinear dynamics. However, for group-affine systems, the error dynamics are linear in Lie algebra coordinates when the vehicle receives the inputs of the reference. The control input and the disturbance enter through the inverse right Jacobian, a matrix that depends on the error. We propagate the linear part of the error dynamics with zonotopes and enclose the effect of the inverse right Jacobian on the disturbance in closed form. The resulting tubes are tight and provably contain all reachable tracking errors in continuous time. When some error components have no input, the inverse right Jacobian leaves unmatched nonlinear terms in the closed loop. None of the tested methods certifies such a closed loop. Hybrid coordinates make these terms matched, so that our controller cancels every nonlinear term that does not contain the disturbance. Our tubes give tighter bounds than the ellipsoid method in simulations of a planar aircraft and a fully actuated multirotor and in real flights of a quadrotor, where the tube contains every measured attitude error. Under larger disturbances and the same feedback gain, the ellipsoid method yields no valid ellipsoid, whereas our tubes remain valid.

新论文 · v1arXiv:2611.12763

PORT:通过稳健的 token 级重加权进行偏好优化

PORT: Preference Optimization via Robust Token-Level Reweighting

Ding Zhu, Xiukun Wei, Tian Xie, Zhihui Zhu, Xueru Zhang, Mohammad Mahdi Khalili

AI 速读 基于摘要

PORT 提出一种 token 级重加权框架,利用 token 对数概率的经验累积分布函数(CDF)抑制噪声偏好数据中的损坏 token,以提升大语言模型对齐的稳健性。

cs.CL · cs.AI

查看原文摘要

Preference optimization has become a central approach for aligning large language models with human values, but its effectiveness depends heavily on the quality of preference annotations. In practice, preference data is often noisy due to annotation errors and ambiguity. Existing robust preference optimization methods primarily operate at the sequence level, implicitly treating entire responses as uniformly correct or incorrect. However, rejected responses may still contain informative reasoning steps, while preferred responses can include subtle errors. In this work, we propose \textbf{P}reference \textbf{O}ptimization via \textbf{R}obust \textbf{T}oken-level reweighting \textbf{(PORT)}, a fine-grained framework for robust alignment under noisy preferences. PORT performs token-level reweighting using the empirical cumulative distribution function (CDF) of token log-probabilities, yielding an efficient forward-pass-only proxy to gradient-norm penalization that selectively suppresses corrupted tokens without additional backward-pass computation. We provide theoretical analysis showing that PORT reduces the gradient bias from noisy preference labels by minimizing an upper bound on the bias term. Extensive experiments across diverse noise settings demonstrate that PORT consistently improves robustness and outperforms existing sequence-level baselines. The code is available at https://github.com/osu-srml/PORT

新论文 · v1arXiv:2611.12761

连续时间随机过程的霍尔逻辑

A Hoare Logic for Continuous-Time Stochastic Processes

Kei Imada, Justin Hsu

AI 速读 基于摘要

作者提出用程序验证的思路处理连续时间随机过程:为概率命令式语言 pWhile^dt 定义连续语义,并给出基于循环不变式与递减函数的霍尔式逻辑。

cs.LO · cs.PL · math.PR

查看原文摘要

Programming language researchers have developed a wide array of tools to reason about probabilistic programs, but it remains unclear how to apply these proof techniques to continuous-time stochastic processes. To bridge this gap, we propose a verification method for properties of continuous-time stochastic processes by reasoning about them as standard programs. To demonstrate our approach, we define a continuous semantics for a probabilistic imperative language $\mathrm{pWhile}^{\mathrm{dt}}$ by taking the Kuratowski limit of the standard semantics, which collects limit points of the standard program semantics as the time-step approaches zero. Then, we present a Hoare-style logic that uses loop invariants and decreasing functions to verify properties of continuous programs. To establish soundness, we identify a broad class of assertions that are preserved from the standard semantics to the continuous semantics. To demonstrate our approach, we verify properties of ten continuous-time stochastic processes modeled in $\mathrm{pWhile}^{\mathrm{dt}}$, including variants of Brownian motion and a stochastic hybrid system.

新论文 · v1arXiv:2611.12760

DeepChronos:面向多步攻击检测的时间神经符号推理

DeepChronos: Temporal Neuro-Symbolic Reasoning for Multi-Step Attack Detection

Sofia Burkow, Andrea Agiollo, Enkeleda Bardhi, Fernando Kuipers, Kaitai Liang, Georgios Smaragdakis

AI 速读 基于摘要

提出神经符号网络入侵检测系统 DeepChronos,用显式攻击阶段规则与时间先验约束检测决策,从“先检测后关联”转向“按逻辑进程检测”。

cs.CR · cs.AI · cs.NI

查看原文摘要

Modern cyber threats commonly unfold as multi-step attacks comprising temporally and causally related phases. Current defenses commonly follow a detect first and correlate later paradigm. Intrusion detection systems classify events independently, while downstream systems reconstruct attack progression from the resulting alerts or audit records. In this context, learning-based detectors learn attack-step dependencies implicitly as statistical regularities in the data and cannot guarantee that their predictions follow the real attack progression. Conversely, correlation systems represent attack relationships explicitly, but apply this knowledge only after the underlying attack observations have been detected. We present DEEPCHRONOS, a neuro-symbolic network intrusion detection system that shifts the paradigm from detect first and correlate later to detection through logical progression. DEEPCHRONOS combines learned flow-level evidence with explicit attack-phase rules, temporal prerequisites, and previously inferred phases, allowing the evolving attack history to inform and constrain each detection decision. Evaluations on DARPA 2000 and AIT-LDS V2 show that DEEPCHRONOS improves macro F1 over purely neural baselines by up to 11.3%, eliminates causal violations under strict temporal constraints, and maintains millisecond-scale inference latency. These results demonstrate that integrating explicit attack progression knowledge into detection produces more accurate and temporally coherent multi-step attack predictions.

新论文 · v1arXiv:2611.12757

简洁无限时域Lyapunov与LQR值问题是BQP完全的

Succinct Infinite-Horizon Lyapunov and LQR Value Problems Are BQP-Complete

Mohamed Serry, Jun Liu

AI 速读 基于摘要

作者证明:以布尔电路简洁表示的无限时域Lyapunov值问题(SLV)是BQP难的,且更一般的简洁LQR值问题(SLQRV)有多项式时间量子算法,因此二者均为BQP完全。

quant-ph · cs.CC · eess.SY · math.OC

查看原文摘要

Fundamental tasks in linear systems analysis and control require evaluating infinite-horizon output energies for autonomous systems and optimal value functions for controlled systems. These quantities are characterized by Lyapunov and algebraic Riccati equations, respectively. In this work, we study promise problems for their evaluation when the system matrices are sparse and represented succinctly by Boolean circuits specifying their nonzero entries, with input length polynomial in the logarithm of the system dimension. For dissipative autonomous systems with coordinate-vector initial states, we define the succinct Lyapunov value problem (SLV). Its extension to linear-quadratic optimal control is the succinct LQR value problem (SLQRV). We prove that SLV is $\BQP$-hard: every problem in bounded-error quantum polynomial time ($\BQP$) reduces to it in polynomial time. We then establish that the more general SLQRV problem admits a quantum algorithm running in time polynomial in the succinct input length. Since SLV is a special case of SLQRV, both problems are $\BQP$-complete. Consequently, unless bounded-error probabilistic polynomial time coincides with $\BQP$, no classical randomized algorithm can solve either problem on all promised instances in time polynomial in the succinct input length.

新论文 · v1arXiv:2611.12755

从格兰杰因果到动力系统中的因果结构

From Granger causality to causal structure in dynamical systems

Víctor Elvira

AI 速读 基于摘要

综述观测空间格兰杰分析的最新进展,并探讨在何种条件下潜变量动力模型中的有向结构可被因果解读。

stat.ML · cs.LG · stat.CO

查看原文摘要

The idea of testing whether the past of one signal improves the prediction of another can be traced back to Wiener. It was Granger who formalized the concept. Granger causality has become one of the most widely used tools to infer directed interactions between the components of a system. The name however can be misleading: it detects predictive relations among the measured signals, not causal relations among the mechanisms that generate them. This article starts by reviewing recent advances in observation-space Granger analysis. It then follows a path to directed structure in latent dynamical models, and asks under which conditions that structure can be read causally.

版本更新 · v2arXiv:2610.40309

面向视觉语言模型零样本对抗鲁棒性的文本响应感知引导

Textual Response-Aware Guidance for Zero-Shot Adversarial Robustness of Vision-Language Models

Yubo Cui, Xianchao Guan, Zheng Zhang

AI 速读 基于摘要

提出文本响应感知引导(TRAG),利用预训练文本编码器对输入嵌入局部变化的响应来约束视觉表示漂移,以提升视觉语言模型的零样本对抗鲁棒性。

cs.CV

查看原文摘要

Vision-language models (VLMs) exhibit strong zero-shot generalization but remain vulnerable to adversarial attacks. Existing adversarial fine-tuning methods often use pre-trained model guidance to preserve image--text alignment while improving robustness. Our analysis of visual representation drift reveals complementary limitations of two representative forms of guidance. Matching visual features constrains all directions equally, providing broad coverage without semantic emphasis. Matching prediction distributions prioritizes semantically relevant directions, but remains limited by the candidate texts and prediction probabilities. These limitations can hinder preservation of pre-trained image--text relationships and degrade zero-shot performance. We therefore propose Textual Response-Aware Guidance (TRAG), which constrains visual representation drift using the pre-trained text encoder's responses to local variations in input embeddings. TRAG constructs anchor-level constraints from the response directions and magnitudes around semantic anchors, then aggregates them using pre-trained image--anchor relevance, yielding an image-adaptive constraint. Extensive experiments across architectures and tasks demonstrate state-of-the-art clean and robust performance. Further analysis shows that TRAG helps preserve pre-trained image--text relationships.

版本更新 · v2arXiv:2606.07487

用深度多智能体强化学习大规模建模观点动力学

Modelling Opinion Dynamics at Scale with Deep MARL

Lukas Seier, Brandon Kaplowitz, Sebastian Towers, Richard Bailey, Jakob Foerster

AI 速读 基于摘要

作者提出一个可扩展到1000个智能体的GPU加速共识与求真博弈,用深度MARL学习观点行为,并在Bluesky网络子集上做初步验证。

cs.MA · cs.GT · cs.SI

查看原文摘要

Modelling opinion dynamics typically relies on hand-crafted local interaction rules to study emergent macroscopic phenomena such as consensus and polarisation. In contrast, multi-agent reinforcement learning (MARL) enables agents to learn such behaviours directly by optimising simple rewards. To explore the potential of MARL for opinion dynamics, we introduce a GPU-accelerated consensus and truth-finding game that scales to populations of up to 1000 agents, comparable to many real-world social sub-networks. To prevent unrealistic conventions, we extend other-play to general-sum social interactions. We next validate our model on a subset of the Bluesky network by recovering agent importance structures from graph topology alone via a learned attention layer, finding that highly conforming populations most closely match human data. In large social media networks, such high levels of conformity significantly reduce collective accuracy and promote dishonest agents that lie to fit in. By contrast, small hunter-gatherer networks are less affected; here, conformity can even improve collective agreement. This suggests a mismatch between evolved human conformity heuristics and modern social media environments as a potential contributor to misinformation.

新论文 · v1arXiv:2611.12754

滑向平衡:双无人机运输中滑动载荷对负载分配与扰动抑制的益处

Sliding into Balance: Benefits of a Sliding Payload for Load Distribution and Disturbance Rejection in Dual-UAV Transportation

Jorge Vicente-Martinez, Francisco José Mañas-Álvarez, Edgar Ramirez-Laboreo

AI 速读 基于摘要

该论文分析双无人机通过单根绳索悬挂可滑动载荷的协同运输方案,并与固定载荷方案在平衡、扰动和阵风下的表现进行对比。

cs.RO · eess.SY

查看原文摘要

This paper analyzes the cooperative transportation of a payload suspended from two drones in which, instead of being attached to two independent ropes, the payload is allowed to slide along a single rope connecting both UAVs. First, a planar nondimensional dynamical model of the sliding motion with viscous friction is presented. Then, considering stationary drones, this configuration is compared with a fixed payload with the same total rope length, both at equilibrium and under disturbances. The analytical results show that the sliding payload equalizes the tension in both rope segments for any relative position of the drones, so that no rope loses tension at equilibrium; it doubles the region reachable by the payload and, except for low friction combined with a large drone separation, it enlarges the set of admissible disturbances. Finally, a dynamic simulation also shows that under realistic gusty wind profiles, the average tension imbalance between the drones is reduced with respect to the fixed-payload solution.

新论文 · v1arXiv:2611.12750

叙事鸿沟:LLM能否帮助我们跨越语言导航多元叙事?

The Narrative Gap: Can LLMs Help Us Navigate Diverse Narratives Across Languages?

Nikhil Sharma, Kelly Marchisio, Kenton Murray, Ziang Xiao

AI 速读 基于摘要

论文提出六项多元多语言信息获取能力并构建NarrativeBench基准,评估14个LLM后发现其难以可靠地帮助用户导航跨语言竞争性叙事。

cs.CL · cs.AI · cs.HC · cs.IR

查看原文摘要

Large language models (LLMs) increasingly mediate how people seek and consume information, including accounts that differ across languages and regions, especially on contested topics. Supporting diverse information seeking requires models to help users navigate competing narratives across languages and information streams. We define six capabilities necessary for diverse multilingual information seeking: three foundational capabilities for narrative inclusion, narrative exclusion, and missing narrative identification; and three higher-level capabilities for presenting diverse narratives, preserving narrative coverage under confirmatory queries, and maintaining information parity across languages. To guide the development of those information seeking capabilities, we introduce \textsc{NarrativeBench}, a benchmark comprising 1,333 news articles in 23 publication languages across 23 contested topics and 460 Reddit-grounded queries. Evaluating 14 LLMs across 24 query languages, we find that current models lack the capabilities needed to reliably navigate diverse narratives, limiting users' control over the perspectives they receive and leaving substantial gaps in exposure to competing accounts even when diversity is explicitly requested. Information access also differs across query framings and languages, risking echo chambers, linguistic filter bubbles, and information disparities that weaken the common ground needed for democratic discourse. \textsc{NarrativeBench} provides a foundation for developing information systems that give users across languages and regions agency to navigate diverse narratives.

版本更新 · v2arXiv:2610.37166

LiveSearch:将问题作为可执行验证器代码,用于专业来源金融搜索智能体的实时基准

LiveSearch: Questions as Executable Verifier Code for a Live Benchmark of Professional-Source Financial Search Agents

Sinuo Wang, Wentao Zhao, Jiacheng Lu, Rui Sun, Beidi Luan, Qiming Shi, Jing Li, Daxin Jiang 等 9 位作者

AI 速读 基于摘要

LiveSearch 把金融问题存成可执行的验证器程序,在评测时调用专业金融数据服务动态生成金标准答案,从而构建一个答案会随市场变化而更新的实时基准。

cs.AI

查看原文摘要

In finance, the answer that matters is usually the current one: prices, rates, and disclosures hold only as of a given day, so retrieving up-to-date facts is a core capability of a financial agent. Yet most financial benchmarks overlook timeliness. They fix each answer at annotation time, so the gold goes stale as markets move and, once leaked into training data, rewards recall rather than search and analysis. We introduce LiveSearch, a live benchmark that stores each question as executable verifier code: a typed program of calls to a professional financial data service, followed by a declarative extractor. The question is frozen, while its gold answer is determined by executing the program against professional financial data at evaluation time, and a daily job re-executes it with no human or language model in the loop. A memorized answer is thus likely outdated, and a model must reach the current one through tools and analysis. Construction is equally automatic: a language model only verbalizes programs enumerated from real tool responses and never chooses an answer. Under general web search, six contemporary tool-calling models reach 70.9% to 76.7% answer accuracy. Over three consecutive days, 70% of served questions change their gold value, and even under tolerance-aware matching, a literally reused earlier snapshot fails on a majority of questions within a single day. Per-question tracking uncovers dynamic shifts in correctness masked by aggregate accuracy.

新论文 · v1arXiv:2611.12747

团约束用于精确因果结构学习

Clique Constraints for Exact Causal Structure Learning

Xiaoyu He, Pavel Rytíř, Petr Ryšavý, Jakub Mareček

AI 速读 基于摘要

在基于评分的精确因果结构学习中引入骨架最大团上界约束,并系统评估其对分支切割搜索的计算权衡。

cs.LG · math.OC · stat.ML

查看原文摘要

Exact score-based causal structure learning supports explicit structural constraints and optimisation certificates, but its branch-and-cut search can be computationally demanding. We augment exact causal structure learning methods with an upper bound on the maximum clique size of the learned undirected skeleton, enforced at integer candidates through lazy cycle and clique constraints. On static linear-Gaussian Erdős-Rényi and scale-free DAGs, we evaluate callback order, alternative clique-cut policies, incumbent repair, fractional separation, and Big-M bounds. We report causal-recovery metrics jointly with optimality gaps, primal and dual bounds, node throughput, MIPSOL events, and callback cost. Two variants receive particular attention: B3 selects the most violated minimal (k+1)-clique inequality at each callback, whereas C3 repairs an infeasible integer candidate by deleting violating edges and refitting affected regression columns using QR-based least squares. The results do not indicate that clique constraints uniformly accelerate optimisation. However, their benefit depends on graph density, sample size, and cut-management policy, while large collections of correlated constraints can reduce node throughput. Tighter valid Big-M bounds reduced the fixed-budget optimality gap in most tested settings, without necessarily changing SHD or F1 at a fixed time limit. These findings characterize the computational trade-offs of clique-constrained exact causal learning, and identify selective separation and incumbent repair as promising directions.

新论文 · v1arXiv:2611.12746

EQP⊆IP 不相对化

$\mathsf{EQP}\subseteq\mathsf{IP}$ Does Not Relativize

Sinan Oral, Supartha Podder

AI 速读 基于摘要

作者构造递归预言机 O 与一元语言 L,使 L 属于 EQP^O 但不属于 IP^O∪PH^O,从而把 BQP 与 IP 的预言机分离强化到精确量子计算。

quant-ph · cs.CC

查看原文摘要

We construct a recursive oracle $O$ and a unary language $L\in\mathsf{EQP}^O\setminus(\mathsf{IP}^O\cup\mathsf{PH}^O)$, strengthening the $\mathsf{BQP}$ versus $\mathsf{IP}$ oracle separation of Bouland et al. to exact quantum computation. The separation is based on a variant of the Forrelation problem in which the normalized correlation of one Boolean function with the Fourier transform of another is promised to be exactly $+1/2$ or $-1/2$. The main technical contribution is a repair method that constructs distributions supported on these exact levels while preserving the indistinguishability needed for the $\mathsf{IP}$ and $\mathsf{PH}$ lower bounds. A uniform polynomial-time quantum algorithm distinguishes the two cases with certainty using only three oracle queries.

版本更新 · v4arXiv:2311.11091

深度张量网络

Deep Tensor Network

Yifan Zhang

AI 速读 基于摘要

提出由查询、键、值收缩构成的序列算子族,利用结合律以 O(d²+dd_v) 的逐 token 工作量精确计算 Gram 型混合,而无需构造 n×n 的 token 矩阵。

cs.LG · cs.AI · cs.CV · quant-ph

查看原文摘要

The Deep Tensor Network is a family of sequence operators built from contractions of queries, keys, and values. Tensor Attention mixes tokens through the Gram operators $Q(K^\top K)Q^\top$ and $K(Q^\top Q)K^\top$; Tensor Interaction mixes channels through operators derived from $C=Q^\top K$. Associativity lets us evaluate these products exactly without forming $n\times n$ token matrices. For query/key width $d$ and value width $d_v$, the Gram constructions use $O(d^2+dd_v)$ work and summary state per causal token, or $O(d^2)$ when $d_v=O(d)$. We derive prefix recurrences and distinguish trace, diagonal, and row normalization, including the conditions under which the weights are nonnegative and row-stochastic. The unmasked token operators are positive semidefinite with rank at most $d$, but their entries can be negative. We also give exact evaluations and costs for Hadamard and matrix-exponential variants.

新论文 · v1arXiv:2611.13059

A Multi-Word Tokenizer for 204 Languages That Needs Fewer Tokens Than GPT-4o's, Including on English

Johin Johny, Ankur Ghosh, Shervin Sam

原文摘要

Languages of India need 1.6 to 14 times as many tokens as English on GPT-4o's tokenizer for the same content, which makes them more expensive and slower to serve. We train tokenizers from scratch on human-written text in about 200 languages. Havon v3 is a byte-level BPE tokenizer with 262,144 tokens and a pre-tokenizer that keeps combining marks with their letters and chunks scripts written without spaces. On FLORES-200 devtest it needs fewer tokens than GPT-4o's o200k_base on all 25 languages of India (median 0.74 of GPT-4o's count) and on 172 of 204 languages, but 7% more on English. Havon v4 adds a second, multi-word training stage (SuperBPE) to v3. At exactly GPT-4o's vocabulary size of 200,019 tokens, v4 needs 6.1% fewer tokens than GPT-4o on English and fewer on 161 of 204 languages and all 25 languages of India (median 0.74); at the same size, v3 needs 9.4% more on English and wins on 133. With 262,144 tokens, v4 needs 9.6% fewer tokens on English and fewer on 182 of 204 languages, and with 524,288 tokens it needs fewer on 203 of 204. The gains are largest for English, Hindi and the Romance languages; scripts without spaces, such as Khmer, Burmese and Tibetan, gain little and lose subword merges. Havon v5 replaces v4's Python-only code sample with permissively licensed code in 15 languages: on code it needs 1.03 to 1.22 times GPT-4o's tokens, against 1.11 to 1.29 for v4, while keeping all 25 languages of India (median 0.74, all significant under a sentence-level bootstrap) and English (0.94). These are compression results only: we have not yet evaluated a model trained with v4 or v5.

cs.CL · cs.AI

新论文 · v1arXiv:2611.13058

The Indic Token Tax in 2026: Measuring Tokenizer. Disparity and Its Cost for the Languages of India

Johin Johny Arimbur, Ankur Ashit Ghosh, Shervin Sam

原文摘要

Commercial language model APIs bill per token, so a tokenizer that needs more tokens for the same content in one language makes that language more expensive, slower, and able to fit less text in a fixed context window. We measure this disparity for nine public tokenizers, including those of GPT-4, GPT-4o, Qwen2.5, DeepSeek-V3, Mistral-Nemo and four Sarvam models, on all 25 languages of India available in FLORES-200 (2,009 professionally translated sentences aligned with English) and on 13,732 human-translated Hinglish-English social media pairs from PHINC. On GPT-4o's tokenizer the same content needs between 1.57 times (Hindi) and 5.01 times (Odia) the tokens of English, and 13.7 times for Santali in Ol Chiki script; GPT-4's tokenizer needs up to 12.8 times. Sarvam-30B/105B is the only third-party tokenizer that stays below 2.6 times on every language (median 1.53). In an illustrative scenario using official list prices accessed on 5 November 2026 and stated assumptions, a support bot handling 100,000 chats a month on GPT-4o would cost about INR 56,640 in English and INR 102,829 in Marathi; these are estimates, not measured production costs. Against an openly licensed Hindi and Marathi morphological segmentation set, no tokenizer places more than 35% of its word-internal boundaries on morpheme boundaries (boundary F1 at most 0.48). In a companion paper we present a tokenizer that reduces this tax. Code, pinned artifacts and raw token counts are available at https://github.com/HavonAI/indic-token-tax.

cs.CL · cs.AI

新论文 · v1arXiv:2611.12740

Computing EF1 Allocations under Budget Constraints in Polynomial Time

Tingwei Hu, Lili Mei, Guochuan Zhang

原文摘要

We study the fair allocation of indivisible items to agents with budget constraints and additive valuations. Each agent compares its own bundle with the subsets of another bundle that fit its budget. An allocation is envy-free up to one item (EF1) if, for every agent and every nonempty subset under comparison, removing an item of maximum value to that agent eliminates its envy. We require EF1 both between agents and towards the charity, which contains all unallocated items. EF1 allocations exist, but computing them in polynomial time has remained open even for homogeneous valuations. We resolve this problem with a polynomial-time algorithm for EF1 allocations under heterogeneous valuations. We also prove that detecting an EF1 violation is NP-complete. Equivalently, checking whether a given allocation is EF1 is coNP-complete. Finally, we give polynomial-time algorithms for envy-freeness up to any item (EFX) both when each agent's valuation is proportional to item size and when each agent has a fixed number of distinct positive item values. As an application, for homogeneous valuations and any $\varepsilon>0$, we obtain an EF1 allocation with at least a $(1/2-\varepsilon)$ fraction of the maximum social welfare in time polynomial in the input length and $1/\varepsilon$.

cs.GT · cs.DS

新论文 · v1arXiv:2611.12739

Plant Orbiter: 3D Plant Reconstruction and Generation with Video Diffusion Orbit Cameras and 3D Gaussian Splatting

Zane K. J. Hartley, Yong En Kok, Lewis A. G. Stuart, Andrew P. French, Michael P. Pound

原文摘要

Modern digital phenotyping pipelines require the precise measurement of a range of 3D morphological plant traits. The accurate, high throughput acquisition of these traits helps breeders and researchers accelerate the development of new varieties, techniques and products. Recent advances in 3D technologies have enabled a rapid transition from 2D to 3D pipelines, but high throughput acquisition of digital 3D twins of plants remains a significant challenge due to complex morphology, thin structures and occlusions. In this paper we present Plant Orbiter, a new approach to generating 3D plants that leverages recent advances in video diffusion to produce high quality 3D Gaussian Splats of plants from a single image. We show that our method more accurately recreates plant morphology than other state-of-the-art methods, achieving the highest 3D F-score against all baselines, while being more practical for high throughput pipelines due to easier alignment and reduced cost of camera placement. Combined with a 2D diffusion model, we show that our method also achieves state-of-the-art results in the generation of 3D plant models from a text prompt.

cs.CV

新论文 · v1arXiv:2611.12738

Quantum Robustness via Sandwiched Rényi Divergences

Luc Rey-Bellet

原文摘要

If a quantum state is known only approximately, how much can the expectation of an observable change? We answer with the worst-case expectation over all states within a sandwiched Rényi divergence budget of a reference state, for every order $α\in [1/2,\infty]$. Convex duality gives this value as a one-dimensional optimization over a sandwiched Schatten norm, running from a fidelity risk at $α= 1/2$, through the entropic risk at $α= 1$, to a quantum conditional value-at-risk at $α= \infty$. The main result is a classicalization theorem: the quantum problem coincides with the classical one for a probability measure $ω_α$ on the spectral interval of the observable, constructed from Heinävaara's tracial joint spectral measure and given by Stahl's theorem at $α= 1$. The Hoeffding, Bernstein and Markov--Krein bounds of classical probability thereby apply to the quantum problem, and yield Pinsker-type inequalities with Bernstein-type refinements. The measure $ω_α$ is the Born measure only at $α= 1/2$; for $α> 1/2$, noncommutativity of the observable with the state strictly lowers its variance below the Born variance, and with it every variance-sensitive bound: a quantum deflation.

quant-ph · cs.IT · math-ph

版本更新 · v2arXiv:2501.10884

Smoothed Analysis of Nash Equilibrium and Fixed Point Computation, and Efficiency of the Homotopy Method

Idan Attias, Yuval Dagan, Constantinos Daskalakis, Rui Yao, Manolis Zampetakis

原文摘要

We study, under a smoothed analysis perspective, the complexity of three PPAD-complete problems: computing Nash equilibria, computing fixed points, and solving variational inequalities. We consider a general perturbation model that bounds the probability density of the quantities evaluated by our algorithms. As special cases, it captures worst-case games perturbed by small random per-action costs, and worst-case maps perturbed by a small random affine function. Under these models, we show that classical algorithmic paradigms, cover-based search and homotopy path-following provably outperform the worst-case state-of-the-art. For games with a constant number of players and $n$ strategies per player, a recursive cover-based algorithm computes an $\varepsilon$-Nash equilibrium using $\log(1/\varepsilon)\cdot 2^{O(n)}$ expected payoff-vector computations, compared to the worst-case runtime tradeoffs of $\log(1/\varepsilon)\cdot n^{O(n)}$, $(1/\varepsilon)^{O(n)}$, $2^{O(\log^2 n/\varepsilon^2)}$, and $2^{O(n)}\mathrm{poly}(\log(1/\varepsilon))$. For fixed points and variational inequalities of bounded and Lipschitz maps on the $n$-dimensional $\ell_2$ unit ball, the same approach runs in expected time $\log(1/\varepsilon)\cdot 2^{O(n)}$, compared to the worst-case $(1/\varepsilon)^{O(n)}$; the $2^{Ω(n)}$ factor is necessary by prior work [Rubinstein, 2017]. For general smooth maps, we further analyze the classical predictor--corrector homotopy method, for which only qualitative convergence guarantees were previously known, showing that under perturbations the homotopy path is well conditioned and has length $2^{O(n)}$, so that the method finds an $\varepsilon$-approximate solution in time $2^{O(n)}\cdot\mathrm{poly}(1/p)/\varepsilon$ with probability at least $1-p$.

cs.GT · cs.DS · cs.LG

新论文 · v1arXiv:2611.12732

$Σ$-LeWM: Learning World Models via S-Information for Robot Control

Xinrui Zu, Michael Przystupa, Zhao Yang, Shujian Yu, Kevin Sebastian Luck

原文摘要

World models have increasingly been applied as a model-based control method for robotics tasks such as manipulation and locomotion. While early world model architectures relied on decoders to prevent mode collapse and learn a predictable latent space, recent works on world models, such as Joint Embedding Predictive Architecture or LeWorldModel, increasingly aim to learn embeddings without a grounding decoder, utilising only a latent forward dynamics predictor and a stabilising entropy term. In this paper, we set out to demonstrate that the further introduction of two additional latent losses to LeWorldModel, a backward dynamics predictor and an inverse dynamics action predictor, enforces further structure on the latent embedding, leading to further gains in both performance and planning efficiency. These are critical properties for improving the applicability of such methods to real-time model-based control of real-world robots. We demonstrate these gains on standard world model benchmarks and two simulated robot manipulation tasks.

cs.LG · cs.AI · cs.RO

新论文 · v1arXiv:2611.12731

NIFT-SQL: Customer-Supervised Table Scoring for Enterprise Text-to-SQL Retrieval

Arjun Rao, Nishant Gopinath

原文摘要

Enterprise text-to-SQL starts with finding the tables a question needs, which is hard when a schema spans several databases and hundreds of cryptic tables. We ask whether models trained separately on each customer's example questions, reading only table names, column names, and declared keys, improve table retrieval. NIFT-SQL (Necessity-Informed Fusion for Text-to-SQL) fuses hybrid search over a retriever fine-tuned on the customer's questions with a fine-tuned 560M cross-encoder that scores every table for whether the query needs it, through a logistic regression fitted on held-out questions. On BEAVER, LiveSQLBench, and LiveSQLBench-Large (244 to 971 tables; four splits, three trainings), NIFT-SQL places every reference table in the top 15 for 69.2%, 89.5%, and 85.6% of test questions, against 30.4%, 41.0%, and 21.8% for hybrid search and 61.1%, 81.7%, and 70.4% for the fine-tuned retriever with the same labels. With off-the-shelf search, the cross-encoder without fine-tuning is 21 points lower. A zero-shot 8B reranker needs no labels and, routed by database, shows no resolved complete@15 difference on two collections; adding its scores as two fusion features gives a high-accuracy mode that reaches 72.4%, 92.5%, and 88.6%, ahead of the validation-selected zero-shot reranker variant on two collections, with no resolved difference on the third. Rank fusion of the two trained parts, grouped by database, shows no resolved difference from the learned fusion. Scored once on a newer BEAVER release, NIFT-SQL beats search and leads the label-matched retriever at top 8 on the synthetic test, but trails it at top 15 on real questions.

cs.IR

新论文 · v1arXiv:2611.12729

KiteNorm: Variance Regularisation for Stable and Scalable Post-LN Transformers

Leon A. Trochelmann, Sajad Movahedi, Shiwei Liu, Antonio Orvieto

原文摘要

Post-LayerNorm Transformers have seen limited practical adoption due to enduring difficulties in scaling them to large depths. While prior research has focused on stabilising Post-LayerNorm by improving conditioning at initialisation, stability often deteriorates when models are trained with large learning rates, forcing additional compromises. In the decoder-only settings we study, existing Post-LayerNorm methods fail to outperform strong Pre-LayerNorm baselines. We propose KiteNorm, a post-norm training method designed to break this trend by constraining hidden-state variance. This approach is motivated by a bound on the LayerNorm Jacobian suggesting that variance growth contributes to gradient vanishing, alongside empirical evidence linking it to rank collapse. KiteNorm combines residual scaling and variance regularisation to maintain stability throughout training, supported by scalarised double-normalisation that provides separately learnable scales for the skip and residual branches. Across decoder-only Transformers up to 1.3B parameters, KiteNorm improves training stability and outperforms leading baselines at their optimal learning rates, with scaling laws favouring KiteNorm across depth, width, batch size, and training budget. Ablations show that each component is necessary for the full stability and performance gains.

cs.LG

版本更新 · v3arXiv:2606.04120

SaliMory: Orchestrating Cognitive Memory for Conversational Agents

Kai Zhang, Xinyuan Zhang, Hongda Jiang, Shiun-Zu Kuo, Hyokun Yun, Ejaz Ahmed, Shereen Oraby, Ziyun Li 等 14 位作者

原文摘要

Recall and personalization, enabled by memory, are pivotal to a lifelong conversational agent. However, simply expanding context windows with raw retrieval degrades reasoning quality, while training memory agents via standard reinforcement learning creates a severe credit assignment bottleneck in a multi stage pipeline. To solve this, we introduce SaliMory, a framework that learns to manage and utilize a cognitively structured memory spanning user facts, preferences, and working memory. By introducing novel stage wise process rewards and reward decomposed contrastive refinement, SaliMory provides isolated supervision for distinct memory operations (selective filtering, consolidation, and cue driven recall) end to end. Across four datasets, SALIMORY records the highest Good Personalization rate at 39.8% and the best QA accuracy at 72.9%, while consistently producing the least vague and most factual memories, keeping vagueness under 20% and factuality above 89%.

cs.CL · cs.AI

新论文 · v1arXiv:2611.12728

Safe Reinforcement Learning for Multi-Turn Clinical Agents: When Dense Rewards Beat Outcome Rewards

Brendan P. Murphy, Agustina Saenz

原文摘要

Reinforcement learning with verifiable rewards is increasingly used to improve language model reasoning, yet its effects on sequential decision procedures remain poorly understood. We study this problem in clinical multi-turn agents that must decide what information to request, when to stop gathering evidence, and which action to take. We introduce three controlled environments: medication refill triage, acute diagnostic work-up, and dosing titration, each with hidden clinical states, query costs, exact adaptive oracles, and deterministic outcome verifiers. Across these tasks, outcome-only group relative policy optimization often improves aggregate reward while producing undesirable procedures, including premature commitment, one-question collapse, and failures on rare high-risk cases. In diagnostics, assigning dense credit to questions that reduce uncertainty improves held-out value and moderate-case support across three training seeds relative to outcome-only reinforcement learning, without the high-risk seed collapse observed under outcome-only training. Separating exploration from exploitation further increases high-risk support at its best checkpoint, but at the cost of excessive information gathering on low-acuity cases. Additional experiments show that the optimal strength of process credit is task-dependent; that after a supervised warmup that already gathers carefully, reinforcement learning raises aggregate reward by trimming questions while degrading high-risk performance; and that reinforcement learning does not reliably repair rare behavior gaps inherited from biased supervised data. Verifiable final outcomes are therefore insufficient for training reliable multi-turn agents: reward design must account for the procedure used to reach an answer, and evaluation must distinguish adaptive information gathering from reward-exploiting shortcuts.

cs.LG · cs.AI · cs.CL

版本更新 · v2arXiv:2610.39062

HEARSAYBENCH: Can AI Navigate from Abstract Human Rights to Lived Lives?

Ava Iranmanesh, Sobhan Lotfi, Ali Iranmanesh, Liwei Jiang

原文摘要

As large language models (LLMs) reach users worldwide, personal counseling has emerged as one of their most consequential applications, sometimes in life-critical situations. Yet broader access does not mean equal support. The text these models learn from disproportionately reflects people with the freedom to act, embedding an implicit assumption that users can leave, report, seek help, or appeal. Existing benchmarks largely inherit this assumption, overlooking users whose agency is constrained by structural conditions such as war, statelessness, and displacement. Meanwhile, models are increasingly shaped by normative frameworks such as Constitutional AI, often grounded in ideals like human rights, even as many users live where those rights offer little protection or recourse. We introduce HEARSAYBENCH, a human-authored dataset of 400 test cases with in-region human verification grounded in scenarios where structural barriers stand between a person and the exercise of their rights. Built from United Nations and other human rights reports, the scenarios span over 80 regions and cover social, environmental, and personal constraints. Drawing on Nobel laureate Amartya Sen's Capabilities Approach, HEARSAYBENCH tests whether models can distinguish what a person is formally entitled to do from what they are actually able to do. Across 14 proprietary and open-weight models, every model infers a user's situation more accurately than it advises on it. Stronger models narrow this gap, but their recommendations remain unsafe to follow. Models may understand lives shaped by constraint, yet that knowledge remains hearsay if it cannot translate into advice people can act on. Those people are not alignment's long tail; they are among the users for whom good guidance matters most. HEARSAYBENCH makes this failure measurable, a step toward systems that serve the world as it is rather than as assumed.

cs.CL

版本更新 · v2arXiv:2605.15721

Contexting as Recommendation: Evolutionary Collaborative Filtering for Context Engineering

Jiachen Zhu, Zhuoying Ou, Congmin Zheng, Yuxiang Chen, Zeyu Zheng, Rong Shan, Lingyu Yang, Lionel Z. Wang 等 12 位作者

原文摘要

Large Language Models (LLMs) are highly sensitive to their input contexts, motivating the development of automated context engineering. However, existing methods predominantly treat this as a global search problem, seeking a single context strategy that maximizes average performance across a dataset. This restrictive assumption overlooks the fact that different inputs often require distinct guidance, leaving substantial instance-level performance gains untapped. In this paper, we propose a paradigm shift by formulating context engineering as a recommendation problem. We introduce \textbf{Neural Collaborative Context Engineering (NCCE)}, a framework that transitions optimization from a static global search to dynamic, instance-wise routing. NCCE first bootstraps a diverse catalog of anchor contexts and then employs a novel \textbf{Context-CF Co-Evolution} mechanism. This stage establishes a synergistic feedback loop: a lightweight Neural Collaborative Filtering (NCF) model learns instance-context preferences to guide the generation of specialized context variants, while the newly evaluated contexts continuously refine the NCF model's understanding of latent preferences. At inference time, the trained NCF model acts as a context router, dynamically assigning the most suitable context strategy to each unseen instance. Theoretical Proofs and comprehensive experiments demonstrate that by matching individual inputs with their optimal contexts, NCCE significantly improves task accuracy, highlighting the critical importance of personalization in LLM context engineering.

cs.CL

版本更新 · v5arXiv:2501.18404

Graded String Diagrams with Applications in Probabilistic Reasoning

Ralph Sarkis, Fabio Zanasi

原文摘要

We introduce string diagrams for graded symmetric monoidal categories. Our approach includes a definition of graded monoidal theory and the corresponding freely generated syntactic category. The para construction on monoidal actegories is a motivating example for our framework, and we show how an axiomatic presentation of the resulting category $\mathbf{Para}(\mathbb{G},\mathbf{C})$ may be modularly obtained from presentations for $\mathbb{G}$ and $\mathbf{C}$. We demonstrate three different applications of our diagrammatic language under the same theme: probabilistic reasoning. First, we show how to axiomatise a variant of the graded category $\mathbf{ImP}$, recently introduced by Liell-Cock and Staton to model imprecise probability. This culminates in a representation, as string diagrams with grading wires, of programs with primitives for nondeterministic and probabilistic choices and conditioning. Second, we exploit our modular presentation result to provide a new proof for a presentation of $\mathbf{Gauss}$, the category of Gaussian stochastic maps. Third, we use grading wires to represent causal interventions on Bayesian networks.

math.CT · cs.LO

新论文 · v1arXiv:2611.12727

OptiCover: A Scalable and Multi-Surface Framework for Cold-Start Visual Optimization under Semantic Constraints

Guillaume Guy, Mihajlo Grbovic, Joe Mills, Selen Uguroglu, Liwei He, Pratiksha Kadam, Nalini Kartha, Qian Aria Li

原文摘要

Selecting an effective cover photo is a critical yet understudied problem in visual merchandising for e-commerce marketplaces. On Airbnb, hosts upload many photos depicting different rooms, amenities, and points of interest, but must manually choose a single static cover photo that is reused across heterogeneous product surfaces. This creates substantial challenges: over 90\% of listing photos have no engagement labels, leading to a severe cold-start regime; guest preferences vary across surfaces and search intents, and visually captivating photos may be semantically misaligned with the underlying listing. At Airbnb scale, this corresponds to hundreds of millions of photos and hundreds of thousands of new uploads per day. We propose \emph{OptiCover}, a scalable framework for cold-start visual optimization under semantic constraints and multi-surface engagement signals. The framework combines (i) a photo candidate generator that identifies candidate photos through various proxies, (ii) lightweight vision heads applied to frozen CLIP embeddings for large-scale semantic scoring, (iii) a cross-modal grounding validator based on a vision-language model (VLM) that enforces semantic representativeness, and (iv) an Exploration Logging Module (ELM) that orchestrates the collection of randomized engagement data across heterogeneous product surfaces and contextual search intents. We evaluate OptiCover through extensive offline studies and multiple online experiments across HomePage, and Search surfaces. OptiCover is now deployed in production and performs tens of millions of photo swaps daily. Online results show a 1.5\% engagement lift on the HomePage and a 0.35\% increase in bookers on Search, generating millions of dollars in incremental revenue. Our findings suggest that action-space shaping and semantic constraints are primary drivers of performance.

cs.CV · cs.LG

版本更新 · v2arXiv:2610.02247

Toward Controlling Biology with Language: Offline Learning of Prompt-Conditioned Interventions for Cells, Organoids, and Biobots

Nam H. Le, Douglas Blackiston, Michael Levin, Josh Bongard

原文摘要

Natural language has become an interface to code, images and molecules, but not to living matter: a stimulus applied to cells has no fixed linguistic meaning, and every paired example of an instruction and its outcome costs a wet-lab experiment. We ask how far an existing archive of experiments can take this interface, using xenobots, motile multicellular constructs made of frog cells, stimulated by a single electrical pulse of varying duration. We re-track 155 archived recordings with two independent methods, measure what the pulse does from each bot's own behaviour before and after it, and separate behaviours the pulse controls from behaviours it does not. Longer pulses stop bots and short pulses leave them moving, sometimes faster; turning direction and starting from rest do not depend on duration, and restarting at most weakly. We then replace hand-made behaviour labels with a blind language critic: fresh sessions of a multimodal language model read plots of each bot and score how well its behaviour matches each of 176 instructions. The critic agrees with the measured labels whenever its plots contain the information the labels are defined on, and dose curves learned from its scores match those learned from the measurements. A prompt-to-intervention model trained only on critic scores maps unseen instructions, including negations, to the right end of the dose range on unseen bots, as well as the same model trained on measured labels. The archive therefore supports a calibrated, refusing language interface for the three behaviours a pulse can set, and identifies, rather than hides, the behaviours it cannot.

q-bio.QM · cs.AI · cs.CV · cs.NE · cs.RO