Tootfinder

Evaluation Framework for Highlight Explanations of Context Utilisation in Language Models
Jingyi Sun, Pepa Atanasova, Sagnik Ray Choudhury, Sekh Mainul Islam, Isabelle Augenstein
https://arxiv.org/abs/2510.02629

Evaluation Framework for Highlight Explanations of Context Utilisation in Language Models
Context utilisation, the ability of Language Models (LMs) to incorporate relevant information from the provided context when generating responses, remains largely opaque to users, who cannot determine whether models draw from parametric memory or provided context, nor identify which specific context pieces inform the response. Highlight explanations (HEs) offer a natural solution as they can point the exact context pieces and tokens that influenced model outputs. However, no existing work evalu…

Safe and Efficient In-Context Learning via Risk Control
Andrea Wynn, Metod Jazbec, Charith Peris, Rinat Khaziev, Anqi Liu, Daniel Khashabi, Eric Nalisnick
https://arxiv.org/abs/2510.02480

Safe and Efficient In-Context Learning via Risk Control
Large language models (LLMs) demonstrate a remarkable ability to learn new tasks from a few in-context examples. However, this flexibility introduces safety concerns: LLMs can be influenced by incorrect or malicious demonstrations -- for example, if an adversary tampers with or injects harmful examples without a human supervisor noticing. This motivates principled designs in which the system itself includes built-in mechanisms to guard against such attacks. We propose a novel approach to limit …

Variational Uncertainty Decomposition for In-Context Learning
I. Shavindra Jayasekera, Jacob Si, Wenlong Chen, Filippo Valdettaro, A. Aldo Faisal, Yingzhen Li
https://arxiv.org/abs/2509.02327

Variational Uncertainty Decomposition for In-Context Learning
As large language models (LLMs) gain popularity in conducting prediction tasks in-context, understanding the sources of uncertainty in in-context learning becomes essential to ensuring reliability. The recent hypothesis of in-context learning performing predictive Bayesian inference opens the avenue for Bayesian uncertainty estimation, particularly for decomposing uncertainty into epistemic uncertainty due to lack of in-context data and aleatoric uncertainty inherent in the in-context predictio…

MalFlows: Context-aware Fusion of Heterogeneous Flow Semantics for Android Malware Detection
Zhaoyi Meng, Fenglei Xu, Wenxiang Zhao, Wansen Wang, Wenchao Huang, Jie Cui, Hong Zhong, Yan Xiong
https://arxiv.org/abs/2508.03588

MalFlows: Context-aware Fusion of Heterogeneous Flow Semantics for Android Malware Detection
Static analysis, a fundamental technique in Android app examination, enables the extraction of control flows, data flows, and inter-component communications (ICCs), all of which are essential for malware detection. However, existing methods struggle to leverage the semantic complementarity across different types of flows for representing program behaviors, and their context-unaware nature further hinders the accuracy of cross-flow semantic integration. We propose and implement MalFlows, a novel…

Anthropic adds context editing and a memory tool to the Claude API, allowing AI agents to handle long-running tasks without frequently hitting context limits (Anthropic)
https://www.anthropic.com/news/context-management

Managing context on the Claude Developer Platform
We're introducing new capabilities for managing your agents’ context on the Claude Developer Platform: context editing and the memory tool.

Don't Just Chase "Highlighted Tokens" in MLLMs: Revisiting Visual Holistic Context Retention
Xin Zou, Di Lu, Yizhou Wang, Yibo Yan, Yuanhuiyi Lyu, Xu Zheng, Linfeng Zhang, Xuming Hu
https://arxiv.org/abs/2510.02912

Don't Just Chase "Highlighted Tokens" in MLLMs: Revisiting Visual Holistic Context Retention
Despite their powerful capabilities, Multimodal Large Language Models (MLLMs) suffer from considerable computational overhead due to their reliance on massive visual tokens. Recent studies have explored token pruning to alleviate this problem, which typically uses text-vision cross-attention or [\texttt{CLS}] attention to assess and discard redundant visual tokens. In this work, we identify a critical limitation of such attention-first pruning approaches, i.e., they tend to preserve semanticall…

FocusAgent: Simple Yet Effective Ways of Trimming the Large Context of Web Agents
Imene Kerboua, Sahar Omidi Shayegan, Megh Thakkar, Xing Han L\`u, L\'eo Boisvert, Massimo Caccia, J\'er\'emy Espinas, Alexandre Aussem, V\'eronique Eglin, Alexandre Lacoste
https://arxiv.org/abs/2510.03204…

FocusAgent: Simple Yet Effective Ways of Trimming the Large Context of Web Agents
Web agents powered by large language models (LLMs) must process lengthy web page observations to complete user goals; these pages often exceed tens of thousands of tokens. This saturates context limits and increases computational cost processing; moreover, processing full pages exposes agents to security risks such as prompt injection. Existing pruning strategies either discard relevant content or retain irrelevant context, leading to suboptimal action prediction. We introduce FocusAgent, a sim…

MRG-Bench: Evaluating and Exploring the Requirements of Context for Repository-Level Code Generation
Haiyang Li
https://arxiv.org/abs/2508.02998 https://ar…

MRG-Bench: Evaluating and Exploring the Requirements of Context for Repository-Level Code Generation
Large Language Models (LLMs) have demonstrated impressive capabilities in code generation. However, current evaluation datasets suffer from issues such as the lack of runnable test cases, deviation from the distribution of real-world code, and the ability to evaluate only the Python language. These limitations undermine the credibility of the evaluation results. To address these limitations, we introduce \textbf{MRG-Bench} (Multi-language Repository-level Code Generation Benchmark), a novel d…

Disentangled Multi-Context Meta-Learning: Unlocking robust and Generalized Task Learning
Seonsoo Kim, Jun-Gill Kang, Taehong Kim, Seongil Hong
https://arxiv.org/abs/2509.01297 h…

Disentangled Multi-Context Meta-Learning: Unlocking robust and Generalized Task Learning
In meta-learning and its downstream tasks, many methods rely on implicit adaptation to task variations, where multiple factors are mixed together in a single entangled representation. This makes it difficult to interpret which factors drive performance and can hinder generalization. In this work, we introduce a disentangled multi-context meta-learning framework that explicitly assigns each task factor to a distinct context vector. By decoupling these variations, our approach improves robustness…

CADD: Context aware disease deviations via restoration of brain images using normative conditional diffusion models
Ana Lawry Aguila, Ayodeji Ijishakin, Juan Eugenio Iglesias, Tomomi Takenaga, Yukihiro Nomura, Takeharu Yoshikawa, Osamu Abe, Shouhei Hanaoka
https://arxiv.org/abs/2508.03594

CADD: Context aware disease deviations via restoration of brain images using normative conditional diffusion models
Applying machine learning to real-world medical data, e.g. from hospital archives, has the potential to revolutionize disease detection in brain images. However, detecting pathology in such heterogeneous cohorts is a difficult challenge. Normative modeling, a form of unsupervised anomaly detection, offers a promising approach to studying such cohorts where the ``normal'' behavior is modeled and can be used at subject level to detect deviations relating to disease pathology. Diffusion models hav…

Block: Balancing Load in LLM Serving with Context, Knowledge and Predictive Scheduling
Wei Da, Evangelia Kalyvianaki
https://arxiv.org/abs/2508.03611 https://

Block: Balancing Load in LLM Serving with Context, Knowledge and Predictive Scheduling
This paper presents Block, a distributed scheduling framework designed to optimize load balancing and auto-provisioning across instances in large language model serving frameworks by leveraging contextual information from incoming requests. Unlike popular model serving systems that rely on monolithic and heuristic task schedulers, Block operates as a fully distributed, stateless, and predictive scheduling system to achieve low overhead, reliability, and scalability. It leverages the determinist…

It sounds like the judge is allowing in a fair amount here that isn't strictly relevant but hammers home both the absurdity and the political context.
#sandwichgate
https://bsky.app/profile/jamieson.bsky…

Dave Jamieson (@jamieson.bsky.social)
Border Patrol agent Lairmore testifies that he was not injured by the sandwich, but he felt the impact through his ballistic vest. The sandwich came apart and "kind of exploded" on his chest upon impact, he says. "I could smell the onions and mustard."

From Text to Trajectories: GPT-2 as an ODE Solver via In-Context
Ziyang Ma, Baojian Zhou, Deqing Yang, Yanghua Xiao
https://arxiv.org/abs/2508.03031 https://

From Text to Trajectories: GPT-2 as an ODE Solver via In-Context
In-Context Learning (ICL) has emerged as a new paradigm in large language models (LLMs), enabling them to perform novel tasks by conditioning on a few examples embedded in the prompt. Yet, the highly nonlinear behavior of ICL for NLP tasks remains poorly understood. To shed light on its underlying mechanisms, this paper investigates whether LLMs can solve ordinary differential equations (ODEs) under the ICL setting. We formulate standard ODE problems and their solutions as sequential prompts an…

https://www.techmeme.com/251104/p19#a251104p19
The rise of artificial intelligence has produced serial writers to science and medical journals, most likely using chatbots to boost the number of citations they’ve published.

European law enforcement arrests nine suspects involved in an alleged crypto fraud ring that stole €600M+ via fake investment platforms promising high returns
By Sergiu Gatlan / BleepingComputer. View the full context on Techmeme.

Post-leave burnout recovery, week 2:
Lots of mixed feelings in returning to work. I really missed my team and my job. Work and the work schedule are each somewhat helpful for recovery. 1:1s with reports and partnerships are a strong highlight of my week!
Motivation while recovering from the tail end of burnout is hard for some types of work as a manager. Burnout makes motivation hard as it is, and lack of context on what's going on makes that harder, and both of those exacerba…

Do you also enjoy "cracking" random Braille inscriptions based on the context, word lengths, repeating letters and the knowledge how "a" is written?

A substitution lemma for multiple context-free languages
Andrew Duncan, Murray Elder, Lisa Frenkel, Mengfan Lyu
https://arxiv.org/abs/2509.02117 https://ar…

A substitution lemma for multiple context-free languages
We present a new criterion for proving that a language is not multiple context-free, which we call a Substitution Lemma. We apply it to show a sample selection of languages are not multiple context-free, including the word problem of $F_2\times F_2$. Our result is in contrast to Kanazawa et al. [2014, Theory Comput. Syst.] who proved that it was not possible to generalise the standard pumping lemma for context-free languages to multiple context-free languages, and Kanazawa [2019, Inform. and …

I have a habbit of making (too) many (small) packages for functionality that might be reused in different context. {box} might be an alternative by making scripts into modlues that can be loaded: #RStats

‘box’: Write Reusable, Composable and Modular R Code
A modern module system for R. Organise code into hierarchical, composable, reusable modules, and use it effortlessly across projects via a flexible, declarative dependency loading syntax.

In the context of making America grounded again, always check what sources other than Elon Musk say about Neuralink Blindsight https://artificialvision.com/neuralink.htm When something sounds too good to be true...

Evaluating the Robustness of Retrieval-Augmented Generation to Adversarial Evidence in the Health Domain
Shakiba Amirshahi, Amin Bigdeli, Charles L. A. Clarke, Amira Ghenai
https://arxiv.org/abs/2509.03787

Evaluating the Robustness of Retrieval-Augmented Generation to Adversarial Evidence in the Health Domain
Retrieval augmented generation (RAG) systems provide a method for factually grounding the responses of a Large Language Model (LLM) by providing retrieved evidence, or context, as support. Guided by this context, RAG systems can reduce hallucinations and expand the ability of LLMs to accurately answer questions outside the scope of their training data. Unfortunately, this design introduces a critical vulnerability: LLMs may absorb and reproduce misinformation present in retrieved evidence. This…

Not to pat myself on the back, but I'm really happy with how the alt texts of the @… bot that are generated from #openstreetmap tags through an Overpass query.
It makes the alt text show where the image was taken, what's potentially in the frame but also context around it. I enjoy reading them every day!

Proto-planetary disk composition-dependent element volatility in the context of rocky planet formation
Rob J. Spaargaren, Oliver Herbort, Haiyang S. Wang, Stephen J. Mojzsis, Paolo Sossi
https://arxiv.org/abs/2509.03724

Proto-planetary disk composition-dependent element volatility in the context of rocky planet formation
The compositions of the Solar System terrestrial bodies are fractionated from that of the Sun, where elemental depletions in the bulk rocky bodies correlate with element volatility, expressed in its 50% condensation temperature. However, because element volatility depends on disk gas composition, it is not mandated that elemental fractionation trends derived from the solar-terrestrial scenario apply to other planetary systems. Here, we expand upon previous efforts to quantify element volatility…

@… I appreciate the additional context, and in hindsight I was more aggressive than warranted. It’s me who should apologize.
@…

Breaking Down the $\textsf{CosmoGEMS}$: Toward Modeling and Understanding Globular Cluster Stellar Streams in a Fully Cosmological Context
Nondh Panithanpaisal, Robyn E. Sanderson, Carl L. Rodriguez, Tjitske Starkenburg, Sarah Pearson, Ana Bonaca, Philip F. Hopkins, Brian T. Cook, Arpit Arora, Newlin C. Weatherford
https://arxiv.org/abs/25…

Breaking Down the $\textsf{CosmoGEMS}$: Toward Modeling and Understanding Globular Cluster Stellar Streams in a Fully Cosmological Context
Next-generation surveys are expected to uncover thousands of globular cluster (GC) stellar streams, motivating the need for a theoretical framework that produces realistic GC streams in a fully cosmological, Milky Way-like environment. We present $\textsf{CosmoGEMS}$, a star-by-star cosmological GC stream framework that self-consistently links small-scale cluster physics with large-scale Galactic dynamics. The initial phase-space positions of stream stars are informed by post-processed GC popul…

Salesforce öffnet Slack für externe KI
Salesforce öffnet Slack für KI: Über API und Model Context Protocol erhalten Entwickler Zugriff auf Chatdaten, um kontextsensitive Agenten zu erstellen.
https://www.

EntropyLong: Effective Long-Context Training via Predictive Uncertainty
Junlong Jia, Ziyang Chen, Xing Wu, Chaochen Gao, Zijia Lin, Debing Zhang, Songlin Hu, Binghui Guo
https://arxiv.org/abs/2510.02330

EntropyLong: Effective Long-Context Training via Predictive Uncertainty
Training long-context language models to capture long-range dependencies requires specialized data construction. Current approaches, such as generic text concatenation or heuristic-based variants, frequently fail to guarantee genuine long-range dependencies. We propose EntropyLong, a novel data construction method that leverages predictive uncertainty to verify dependency quality. Our approach identifies high-entropy positions in documents, retrieves semantically relevant contexts from large co…

CATMark: A Context-Aware Thresholding Framework for Robust Cross-Task Watermarking in Large Language Models
Yu Zhang, Shuliang Liu, Xu Yang, Xuming Hu
https://arxiv.org/abs/2510.02342

CATMark: A Context-Aware Thresholding Framework for Robust Cross-Task Watermarking in Large Language Models
Watermarking algorithms for Large Language Models (LLMs) effectively identify machine-generated content by embedding and detecting hidden statistical features in text. However, such embedding leads to a decline in text quality, especially in low-entropy scenarios where performance needs improvement. Existing methods that rely on entropy thresholds often require significant computational resources for tuning and demonstrate poor adaptability to unknown or cross-task generation scenarios. We prop…

MonSTeR: a Unified Model for Motion, Scene, Text Retrieval
Luca Collorone, Matteo Gioia, Massimiliano Pappa, Paolo Leoni, Giovanni Ficarra, Or Litany, Indro Spinelli, Fabio Galasso
https://arxiv.org/abs/2510.03200

MonSTeR: a Unified Model for Motion, Scene, Text Retrieval
Intention drives human movement in complex environments, but such movement can only happen if the surrounding context supports it. Despite the intuitive nature of this mechanism, existing research has not yet provided tools to evaluate the alignment between skeletal movement (motion), intention (text), and the surrounding context (scene). In this work, we introduce MonSTeR, the first MOtioN-Scene-TExt Retrieval model. Inspired by the modeling of higher-order relations, MonSTeR constructs a unif…

the context, the context, the context

AttenTrack: Mobile User Attention Awareness Based on Context and External Distractions
Yutong Lin, Suyuan Liu, Kaiwen Guo, Haohua Du, Chao Liu, Xiang-Yang Li
https://arxiv.org/abs/2509.01414

AttenTrack: Mobile User Attention Awareness Based on Context and External Distractions
In the mobile internet era, managing limited attention amid information overload is crucial for enhancing collaboration and information delivery. However, current attention-aware systems often depend on wearables or personalized data, limiting their scalability and cross-context adaptability. Inspired by psychological theories, we attempt to treat mobile notifications as naturally occurring external distractions and infer users' attention states based on their response behaviors and contextual …

A post from the archive 📫:
Building debugging context for Copilot Chat
https://www.poppastring.com/blog/building-debugging-context-for-copilot-chat

How Does the Pretraining Distribution Shape In-Context Learning? Task Selection, Generalization, and Robustness
Wa\"iss Azizian, Ali Hasan
https://arxiv.org/abs/2510.01163 …

How Does the Pretraining Distribution Shape In-Context Learning? Task Selection, Generalization, and Robustness
The emergence of in-context learning (ICL) in large language models (LLMs) remains poorly understood despite its consistent effectiveness, enabling models to adapt to new tasks from only a handful of examples. To clarify and improve these capabilities, we characterize how the statistical properties of the pretraining distribution (e.g., tail behavior, coverage) shape ICL on numerical tasks. We develop a theoretical framework that unifies task selection and generalization, extending and sharpeni…

Comment on Marek Czachor article entitled "On Relativity of Quantumness as Implied by Relativity of Arithmetic and Probability"
Krzysztof Sienicki, Miko{\l}aj Sienicki
https://arxiv.org/abs/2510.02412

Comment on Marek Czachor article entitled "On Relativity of Quantumness as Implied by Relativity of Arithmetic and Probability"
Czachor's model of hierarchical arithmetics begins with a valid formal premise but fixes the key probability mapping g by importing the Born rule and Fubini-Study metric from standard quantum mechanics, where Born probabilities are Kolmogorov within a fixed measurement context. This g is then applied in a non-Newtonian hidden-variable setting, producing a hybrid framework whose agreement with quantum correlations is built in by design, not derived from new physics, and thus does not constitute …

Labeled Plane Trees and Increasing Plane Trees
Lora R. Du, Kathy Q. Ji, Dax T. X. Zhang
https://arxiv.org/abs/2510.03039 https://arxiv.org/pdf/2510.03039…

Labeled Plane Trees and Increasing Plane Trees
This note is dedicated to presenting a polynomial analogue of $(n+1)!C_n=2^n(2n-1)!!$ (with $C_n$ as the $n$-th Catalan number) in the context of labeled plane trees and increasing plane trees, based on the definition of improper edges in labeled plane trees. A new involution on labeled plane trees is constructed to establish this identity, implying that the number of improper edges and the number of proper edges are equidsitributed over the set of labeled plane trees.

An axiomatic approach to analytic $1$-affineness
Matteo Montagnani, Emanuele Pavia
https://arxiv.org/abs/2509.04341 https://arxiv.org/pdf/2509.04341…

An axiomatic approach to analytic $1$-affineness
The notion of $1$-affineness was originally formulated by Gaitsgory in the context of derived algebraic geometry. Motivated by applications to rigid and analytic geometry, we introduce two very general and abstract frameworks where it makes to ask for objects to be $1$-affine with respect to some sheaf of categories. The first framework is suited for studying the problem of $1$-affineness when the sheaf of categories arises from an operation in a six-functor formalism over $\mathscr{C}$; we app…

Theory of Scaling Laws for In-Context Regression: Depth, Width, Context and Time
Blake Bordelon, Mary I. Letey, Cengiz Pehlevan
https://arxiv.org/abs/2510.01098 https://

Theory of Scaling Laws for In-Context Regression: Depth, Width, Context and Time
We study in-context learning (ICL) of linear regression in a deep linear self-attention model, characterizing how performance depends on various computational and statistical resources (width, depth, number of training steps, batch size and data per context). In a joint limit where data dimension, context length, and residual stream width scale proportionally, we analyze the limiting asymptotics for three ICL settings: (1) isotropic covariates and tasks (ISO), (2) fixed and structured covarianc…

Transport coefficients of charged Gauss-Bonnet black holes with arbitrary topology
Mois\'es Bravo-Gaete, Luis Guajardo, Daniel F. Higuita-Borja, Julio A. M\'endez-Zavaleta
https://arxiv.org/abs/2509.03838

Transport coefficients of charged Gauss-Bonnet black holes with arbitrary topology
In this study, we present a novel family of exact black hole solutions constructed in the context of five-dimensional Gauss-Bonnet gravity. These solutions add a non-linear charge to the Bañados-Teitelboim-Zanelli-like configurations known to exist with arbitrary Thurston horizon geometry. We establish constraints on the parameter space defining physically viable black holes, aligning with the standard energy conditions. An explicit proof of the first law of thermodynamics within our scenario …

An Agentic Model Context Protocol Framework for Medical Concept Standardization
Jaerong Ahn, Andrew Wen, Nan Wang, Heling Jia, Zhiyi Yue, Sunyang Fu, Hongfang Liu
https://arxiv.org/abs/2509.03828

An Agentic Model Context Protocol Framework for Medical Concept Standardization
The Observational Medical Outcomes Partnership (OMOP) common data model (CDM) provides a standardized representation of heterogeneous health data to support large-scale, multi-institutional research. One critical step in data standardization using OMOP CDM is the mapping of source medical terms to OMOP standard concepts, a procedure that is resource-intensive and error-prone. While large language models (LLMs) have the potential to facilitate this process, their tendency toward hallucination ma…

Thinking with Nothinking Calibration: A New In-Context Learning Paradigm in Reasoning Large Language Models
Haotian Wu, Bo Xu, Yao Shu, Menglin Yang, Chengwei Qin
https://arxiv.org/abs/2508.03363

Thinking with Nothinking Calibration: A New In-Context Learning Paradigm in Reasoning Large Language Models
Reasoning large language models (RLLMs) have recently demonstrated remarkable capabilities through structured and multi-step reasoning. While prior research has primarily focused on improving their training and inference strategies, their potential for in-context learning (ICL) remains largely underexplored. To fill this gap, we propose Thinking with Nothinking Calibration (JointThinking), a new ICL paradigm that leverages the structured difference between two reasoning modes, i.e., Thinking an…

CARPO: Leveraging Listwise Learning-to-Rank for Context-Aware Query Plan Optimization
Wenrui Zhou, Qiyu Liu, Jingshu Peng, Aoqian Zhang, Lei Chen
https://arxiv.org/abs/2509.03102

CARPO: Leveraging Listwise Learning-to-Rank for Context-Aware Query Plan Optimization
Efficient data processing is increasingly vital, with query optimizers playing a fundamental role in translating SQL queries into optimal execution plans. Traditional cost-based optimizers, however, often generate suboptimal plans due to flawed heuristics and inaccurate cost models, leading to the emergence of Learned Query Optimizers (LQOs). To address challenges in existing LQOs, such as the inconsistency and suboptimality inherent in pairwise ranking methods, we introduce CARPO, a generic fr…

A Comparative Study of Optimal Control and Neural Networks in Asteroid Rendezvous Mission Analysis
Zhong Zhang, Niccol\`o Michelotti, Gon\c{c}alo Oliveira Pinho, Francesco Topputo
https://arxiv.org/abs/2508.02920

A Comparative Study of Optimal Control and Neural Networks in Asteroid Rendezvous Mission Analysis
This paper presents a comparative study of the applicability and accuracy of optimal control methods and neural network-based estimators in the context of porkchop plots for preliminary asteroid rendezvous mission design. The scenario considered involves a deep-space CubeSat equipped with a low-thrust engine, departing from Earth and rendezvousing with a near-Earth asteroid within a three-year launch window. A low-thrust trajectory optimization model is formulated, incorporating variable specif…

On singular limits of relative entropies
Feng Xu
https://arxiv.org/abs/2508.03271 https://arxiv.org/pdf/2508.03271

On singular limits of relative entropies
In this paper we generalize a key result relating singular limits of certain relative entropies with index in the setting of conformal nets, which has played an important role recently in the mathematical theory of relative entropies in the context of Conformal Field Theory.

Holographic F(Q,T) Gravity with Lambert Solution
Houda Filali, Mouhsine Koussou, Mohamed Bennai, Rachid Ahl Laamara
https://arxiv.org/abs/2508.03486 https://

Holographic F(Q,T) Gravity with Lambert Solution
In this work, we study a model of holographic dark energy using FLRW cosmology in the context of modified gravity. An extension of the symmetric teleparallel gravity is obtained by considering the gravitational action L is given by an arbitrary function f of the nonmetricity Q, where the nonmetricity Q is responsible for the gravitational interaction, and the trace of the matter energy momentum tensor T, so that L=f(Q,T). We expand on the features of the derived cosmological model in view of th…

Flamed-TTS: Flow Matching Attention-Free Models for Efficient Generating and Dynamic Pacing Zero-shot Text-to-Speech
Hieu-Nghia Huynh-Nguyen, Huynh Nguyen Dang, Ngoc-Son Nguyen, Van Nguyen
https://arxiv.org/abs/2510.02848

Flamed-TTS: Flow Matching Attention-Free Models for Efficient Generating and Dynamic Pacing Zero-shot Text-to-Speech
Zero-shot Text-to-Speech (TTS) has recently advanced significantly, enabling models to synthesize speech from text using short, limited-context prompts. These prompts serve as voice exemplars, allowing the model to mimic speaker identity, prosody, and other traits without extensive speaker-specific data. Although recent approaches incorporating language models, diffusion, and flow matching have proven their effectiveness in zero-shot TTS, they still encounter challenges such as unreliable synth…

Can Large Language Models Identify Materials from Radar Signals?
Jiangyou Zhu, Hongyu Deng, He Chen
https://arxiv.org/abs/2508.03120 https://arxiv.org/pdf/…

Can Large Language Models Identify Materials from Radar Signals?
Accurately identifying the material composition of objects is a critical capability for AI robots powered by large language models (LLMs) to perform context-aware manipulation. Radar technologies offer a promising sensing modality for material recognition task. When combined with deep learning, radar technologies have demonstrated strong potential in identifying the material of various objects. However, existing radar-based solutions are often constrained to closed-set object categories and typ…

We Have It Covered: A Resampling-based Method for Uplift Model Comparison
Yang Liu, Chaoyu Yuan
https://arxiv.org/abs/2509.04315 https://arxiv.org/pdf/2509…

We Have It Covered: A Resampling-based Method for Uplift Model Comparison
Uplift models play a critical role in modern marketing applications to help understand the incremental benefits of interventions and identify optimal targeting strategies. A variety of techniques exist for building uplift models, and it is essential to understand the model differences in the context of intended applications. The uplift curve is a widely adopted tool for assessing uplift model performance on the selection universe when observations are available for the entire population. Howeve…

Regular Rotating Black Hole: Probing the boundaries of the Radiative Signatures and Jet Power
Subhadip Sau, John W. Moffat
https://arxiv.org/abs/2508.03473 https://

Regular Rotating Black Hole: Probing the boundaries of the Radiative Signatures and Jet Power
We perform a detailed observational analysis of several galactic X-ray binaries, focusing on the interplay between black hole spin, jet power, and radiative efficiency within the context of Blandford-Znajek-powered jets. Using updated measurements from continuum fitting and Fe-line methods, we constrain the spin parameter a and the deviation parameter $β$ for five key black hole systems: H1743-322, XTE J1550-564, GRS 1124-683, GRO J1655-40, and GRS 1915+105. For each system, we compare the all…

AdaptCache: KV Cache Native Storage Hierarchy for Low-Delay and High-Quality Language Model Serving
Shaoting Feng, Hanchen Li, Kuntai Du, Zhuohan Gu, Yuhan Liu, Jiayi Yao, Siddhant Ray, Samuel Shen, Yihua Cheng, Ganesh Ananthanarayanan, Junchen Jiang
https://arxiv.org/abs/2509.00105

AdaptCache: KV Cache Native Storage Hierarchy for Low-Delay and High-Quality Language Model Serving
Large language model (LLM) applications often reuse previously processed context, such as chat history and documents, which introduces significant redundant computation. Existing LLM serving systems address such redundant computation by storing the KV caches of processed context and loading the corresponding KV cache when a new request reuses the context. Further, as these LLM applications scale, the total size of KV caches becomes excessively large and requires both DRAM and SSD for full stora…

Clarifying Semantics of In-Context Examples for Unit Test Generation
Chen Yang, Lin Yang, Ziqi Wang, Dong Wang, Jianyi Zhou, Junjie Chen
https://arxiv.org/abs/2510.01994 https:/…

Clarifying Semantics of In-Context Examples for Unit Test Generation
Recent advances in large language models (LLMs) have enabled promising performance in unit test generation through in-context learning (ICL). However, the quality of in-context examples significantly influences the effectiveness of generated tests-poorly structured or semantically unclear test examples often lead to suboptimal outputs. In this paper, we propose CLAST, a novel technique that systematically refines unit tests to improve their semantic clarity, thereby enhancing their utility as i…

Every damn time I order sandwiches from this cheesesteak place I make the same damn mistake. I have them add "pepper rings" which I think of as the medium-hot peppers most sandwich places use. No, they use Hungarian Hots that are hotter than the average jalapeño.
I can handle heat but this is the wrong context. My own damn fault too...

from my link log —
Context parameters and API design in Kotlin.
https://serranofp.com/blog/context-params.html
saved 2025-10-20 https://

RAGuard: A Novel Approach for in-context Safe Retrieval Augmented Generation for LLMs
Connor Walker, Koorosh Aslansefat, Mohammad Naveed Akram, Yiannis Papadopoulos
https://arxiv.org/abs/2509.03768

RAGuard: A Novel Approach for in-context Safe Retrieval Augmented Generation for LLMs
Accuracy and safety are paramount in Offshore Wind (OSW) maintenance, yet conventional Large Language Models (LLMs) often fail when confronted with highly specialised or unexpected scenarios. We introduce RAGuard, an enhanced Retrieval-Augmented Generation (RAG) framework that explicitly integrates safety-critical documents alongside technical manuals.By issuing parallel queries to two indices and allocating separate retrieval budgets for knowledge and safety, RAGuard guarantees both technical …

Topic Modeling as Long-Form Generation: Can Long-Context LLMs revolutionize NTM via Zero-Shot Prompting?
Xuan Xu, Haolun Li, Zhongliang Yang, Beilin Chu, Jia Song, Moxuan Xu, Linna Zhou
https://arxiv.org/abs/2510.03174

Topic Modeling as Long-Form Generation: Can Long-Context LLMs revolutionize NTM via Zero-Shot Prompting?
Traditional topic models such as neural topic models rely on inference and generation networks to learn latent topic distributions. This paper explores a new paradigm for topic modeling in the era of large language models, framing TM as a long-form generation task whose definition is updated in this paradigm. We propose a simple but practical approach to implement LLM-based topic model tasks out of the box (sample a data subset, generate topics and representative text with our prompt, text assi…

Object-Reconstruction-Aware Whole-body Control of Mobile Manipulators
Fatih Dursun, Bruno Vilhena Adorno, Simon Watson, Wei Pan
https://arxiv.org/abs/2509.04094 https://

Object-Reconstruction-Aware Whole-body Control of Mobile Manipulators
Object reconstruction and inspection tasks play a crucial role in various robotics applications. Identifying paths that reveal the most unknown areas of the object becomes paramount in this context, as it directly affects efficiency, and this problem is known as the view path planning problem. Current methods often use sampling-based path planning techniques, evaluating potential views along the path to enhance reconstruction performance. However, these methods are computationally expensive as …

Probability distributions over CSS codes: two-universality, QKD hashing, collision bounds, security
Pete Rigas
https://arxiv.org/abs/2510.02402 https://arx…

Probability distributions over CSS codes: two-universality, QKD hashing, collision bounds, security
We characterize novel probability distributions for CSS codes. Such classes of error correcting codes, originally introduced by Calderbank, Shor, and Steane, are of great significance in advancing the fidelity of Quantum computation, with implications for future near term applications. Within the context of Quantum key distribution, such codes, as examined by Ostrev in arXiv: 2109.06709 along with two-universal hashing protocols, have greatly simplified Quantum phases of computation for uncondi…

IoT-MCP: Bridging LLMs and IoT Systems Through Model Context Protocol
Ningyuan Yang, Guanliang Lyu, Mingchen Ma, Yiyi Lu, Yiming Li, Zhihui Gao, Hancheng Ye, Jianyi Zhang, Tingjun Chen, Yiran Chen
https://arxiv.org/abs/2510.01260

IoT-MCP: Bridging LLMs and IoT Systems Through Model Context Protocol
The integration of Large Language Models (LLMs) with Internet-of-Things (IoT) systems faces significant challenges in hardware heterogeneity and control complexity. The Model Context Protocol (MCP) emerges as a critical enabler, providing standardized communication between LLMs and physical devices. We propose IoT-MCP, a novel framework that implements MCP through edge-deployed servers to bridge LLMs and IoT ecosystems. To support rigorous evaluation, we introduce IoT-MCP Bench, the first bench…

Variable Rate Image Compression via N-Gram Context based Swin-transformer
Priyanka Mudgal, Feng Liu
https://arxiv.org/abs/2510.00058 https://arxiv.org/pdf/…

Variable Rate Image Compression via N-Gram Context based Swin-transformer
This paper presents an N-gram context-based Swin Transformer for learned image compression. Our method achieves variable-rate compression with a single model. By incorporating N-gram context into the Swin Transformer, we overcome its limitation of neglecting larger regions during high-resolution image reconstruction due to its restricted receptive field. This enhancement expands the regions considered for pixel restoration, thereby improving the quality of high-resolution reconstructions. Our m…

Theta-Categories and Tannakian duality
Joost Nuiten, Bertrand Toen
https://arxiv.org/abs/2508.03145 https://arxiv.org/pdf/2508.03145

Theta-Categories and Tannakian duality
We introduce a notion of $Θ$-categories, which is a refinement of the notion of symmetric monoidal $\infty$-categories. We use this notion to prove a Tannakian duality statement, relating $Θ$-categories with fpqc-stacks by means of a certain stack of fiber functors in the context of $Θ$-categories. This provides, over a base ring of arbitrary characteristic, a strong link between Tannakian $Θ$-categories and the schematic homotopy types.

Strategic Fusion of Vision Language Models: Shapley-Credited Context-Aware Dawid-Skene for Multi-Label Tasks in Autonomous Driving
Yuxiang Feng, Keyang Zhang, Hassane Ouchouid, Ashwil Kaniamparambil, Ioannis Souflas, Panagiotis Angeloudis
https://arxiv.org/abs/2510.01126

Strategic Fusion of Vision Language Models: Shapley-Credited Context-Aware Dawid-Skene for Multi-Label Tasks in Autonomous Driving
Large vision-language models (VLMs) are increasingly used in autonomous-vehicle (AV) stacks, but hallucination limits their reliability in safety-critical pipelines. We present Shapley-credited Context-Aware Dawid-Skene with Agreement, a game-theoretic fusion method for multi-label understanding of ego-view dashcam video. It learns per-model, per-label, context-conditioned reliabilities from labelled history and, at inference, converts each model's report into an agreement-guardrailed log-likel…

Crosslisted article(s) found for cs.FL. https://arxiv.org/list/cs.FL/new
[1/1]:
- Unraveling Syntax: How Language Models Learn Context-Free Grammars
Laura Ying Schulz, Daniel Mitropolsky, Tomaso Poggio

PyLate: Flexible Training and Retrieval for Late Interaction Models
Antoine Chaffin, Rapha\"el Sourty
https://arxiv.org/abs/2508.03555 https://arxiv.o…

PyLate: Flexible Training and Retrieval for Late Interaction Models
Neural ranking has become a cornerstone of modern information retrieval. While single vector search remains the dominant paradigm, it suffers from the shortcoming of compressing all the information into a single vector. This compression leads to notable performance degradation in out-of-domain, long-context, and reasoning-intensive retrieval tasks. Multi-vector approaches pioneered by ColBERT aim to address these limitations by preserving individual token embeddings and computing similarity via…

IndiCASA: A Dataset and Bias Evaluation Framework in LLMs Using Contrastive Embedding Similarity in the Indian Context
Santhosh G S, Akshay Govind S, Gokul S Krishnan, Balaraman Ravindran, Sriraam Natarajan
https://arxiv.org/abs/2510.02742

IndiCASA: A Dataset and Bias Evaluation Framework in LLMs Using Contrastive Embedding Similarity in the Indian Context
Large Language Models (LLMs) have gained significant traction across critical domains owing to their impressive contextual understanding and generative capabilities. However, their increasing deployment in high stakes applications necessitates rigorous evaluation of embedded biases, particularly in culturally diverse contexts like India where existing embedding-based bias assessment methods often fall short in capturing nuanced stereotypes. We propose an evaluation framework based on a encoder …

A JWST/MIRI view of k Andromedae b: Refining its mass, age, and physical parameters
N. Godoy, E. Choquet, E. Serabyn, M. Malin, P. Tremblin, C. Danielski, P. O. Lagage, A. Boccaletti, B. Charnay, M. E. Ressler
https://arxiv.org/abs/2509.03624

A JWST/MIRI view of k Andromedae b: Refining its mass, age, and physical parameters
Context. kAndb is a substellar companion near the planet-brown dwarf boundary, orbiting a B9IV star at 50-100 au. Estimates of its age and mass vary, fueling a decade-long debate. Atmospheric parameters (Teff 1650-2050 K, log(g) 3.5-5.5) remain poorly constrained due to model differences and heterogeneous datasets. Aims. We refine the characterization of kAndb using mid-infrared data to capture its bolometric emission. Combined with NIR measurements, we constrain Teff, log(g), and radius to red…

Investigating The Smells of LLM Generated Code
Debalina Ghosh Paul, Hong Zhu, Ian Bayley
https://arxiv.org/abs/2510.03029 https://arxiv.org/pdf/2510.03029

Investigating The Smells of LLM Generated Code
Context: Large Language Models (LLMs) are increasingly being used to generate program code. Much research has been reported on the functional correctness of generated code, but there is far less on code quality. Objectives: In this study, we propose a scenario-based method of evaluating the quality of LLM-generated code to identify the weakest scenarios in which the quality of LLM generated code should be improved. Methods: The method measures code smells, an important indicator of code qua…

Traversable Kaluza-Klein wormholes?
Christopher Simmonds (Victoria University of Wellington), Matt Visser (Victoria University of Wellington)
https://arxiv.org/abs/2508.02824 ht…

Traversable Kaluza-Klein wormholes?
Various authors have suggested that Kaluza--Klein variants of traversable wormholes might to some extent ameliorate the defocussing properties (the curvature condition violations, and implied energy condition violations) inherent in positing the existence of a traversable wormhole throat. We shall show that in a Kaluza--Klein context the price paid for completely eliminating the defocussing properties of the wormhole throat is extremely high -- the 5th dimension has to become truly enormous (fo…

Linear RNNs for autoregressive generation of long music samples
Konrad Szewczyk, Daniel Gallo Fern\'andez, James Townsend
https://arxiv.org/abs/2510.02401 https://

Linear RNNs for autoregressive generation of long music samples
Directly learning to generate audio waveforms in an autoregressive manner is a challenging task, due to the length of the raw sequences and the existence of important structure on many different timescales. Traditional approaches based on recurrent neural networks, as well as causal convolutions and self-attention, have only had limited success on this task. However, recent work has shown that deep state space models, also referred to as linear RNNs, can be highly efficient in this context. In …

A supermassive black hole under the radar: repeating X-ray variability in a Seyfert galaxy
Matteo Imbrogno, Andrea Sacchi, Giovanni Miniutti, Francesco Tombesi, Gian Luca Israel, Enrico Piconcelli, Roberta Amato
https://arxiv.org/abs/2510.02832

A supermassive black hole under the radar: repeating X-ray variability in a Seyfert galaxy
In the last few years, a few supermassive black holes (SMBHs) have shown short-term (of the order of hours) X-ray variability. Given the limited size of the sample, every new addition to this class of SMBHs can bring invaluable information. Within the context of an automated search for X-ray sources showing flux variability in the \textit{Chandra} archive, we identified peculiar variability patterns in 2MASX J12571076+2724177 (J1257), a SMBH in the Coma cluster, during observations performed in…

Dynamical evolution timescales for the supermassive black hole system in the galaxy NGC 7727 (Arp 222)
P. Berczik, M. Ishchenko, O. Veles, M. Sobolenko, K. Voggel, C. M. Boily, E. Polyachenko, R. State
https://arxiv.org/abs/2510.02974

Dynamical evolution timescales for the supermassive black hole system in the galaxy NGC 7727 (Arp 222)
Context. A dual active galactic nucleus candidate with a separation of only 500 pc was recently found in NGC 7727. According to the hierarchical merging scenario, such objects would be expected to merge on a timescale of a few hundred Myr. However, estimating the accurate merging timescales for the two nuclei is still a complex challenge. Aims. Using our numerical N-body code, we can trace the full evolution of central black holes during all phases: dynamical friction of unbound black holes, bi…

Hedging with memory: shallow and deep learning with signatures
Eduardo Abi Jaber, Louis-Amand G\'erard
https://arxiv.org/abs/2508.02759 https://arxiv.o…

Hedging with memory: shallow and deep learning with signatures
We investigate the use of path signatures in a machine learning context for hedging exotic derivatives under non-Markovian stochastic volatility models. In a deep learning setting, we use signatures as features in feedforward neural networks and show that they outperform LSTMs in most cases, with orders of magnitude less training compute. In a shallow learning setting, we compare two regression approaches: the first directly learns the hedging strategy from the expected signature of the price p…

Agentic-AI Healthcare: Multilingual, Privacy-First Framework with MCP Agents
Mohammed A. Shehab
https://arxiv.org/abs/2510.02325 https://arxiv.org/pdf/2510…

Agentic-AI Healthcare: Multilingual, Privacy-First Framework with MCP Agents
This paper introduces Agentic-AI Healthcare, a privacy-aware, multilingual, and explainable research prototype developed as a single-investigator project. The system leverages the emerging Model Context Protocol (MCP) to orchestrate multiple intelligent agents for patient interaction, including symptom checking, medication suggestions, and appointment scheduling. The platform integrates a dedicated Privacy and Compliance Layer that applies role-based access control (RBAC), AES-GCM field-level e…

In-Context Learning can Perform Continual Learning Like Humans
Liuwang Kang, Fan Wang, Shaoshan Liu, Hung-Chyun Chou, Chuan Lin, Ning Ding
https://arxiv.org/abs/2509.22764 https…

In-Context Learning can Perform Continual Learning Like Humans
Large language models (LLMs) can adapt to new tasks via in-context learning (ICL) without parameter updates, making them powerful learning engines for fast adaptation. While extensive research has examined ICL as a few-shot learner, whether it can achieve long-term retention and cross-task knowledge accumulation when multitasks arrive sequentially remains underexplored. Motivated by human memory studies, we investigate the retention characteristics of ICL in multitask settings and extend it to …

GRASPing Anatomy to Improve Pathology Segmentation
Keyi Li, Alexander Jaus, Jens Kleesiek, Rainer Stiefelhagen
https://arxiv.org/abs/2508.03374 https://arx…

GRASPing Anatomy to Improve Pathology Segmentation
Radiologists rely on anatomical understanding to accurately delineate pathologies, yet most current deep learning approaches use pure pattern recognition and ignore the anatomical context in which pathologies develop. To narrow this gap, we introduce GRASP (Guided Representation Alignment for the Segmentation of Pathologies), a modular plug-and-play framework that enhances pathology segmentation models by leveraging existing anatomy segmentation models through pseudolabel integration and featur…

Neural Video Compression with In-Loop Contextual Filtering and Out-of-Loop Reconstruction Enhancement
Yaojun Wu, Chaoyi Lin, Yiming Wang, Semih Esenlik, Zhaobin Zhang, Kai Zhang, Li Zhang
https://arxiv.org/abs/2509.04051

Neural Video Compression with In-Loop Contextual Filtering and Out-of-Loop Reconstruction Enhancement
This paper explores the application of enhancement filtering techniques in neural video compression. Specifically, we categorize these techniques into in-loop contextual filtering and out-of-loop reconstruction enhancement based on whether the enhanced representation affects the subsequent coding loop. In-loop contextual filtering refines the temporal context by mitigating error propagation during frame-by-frame encoding. However, its influence on both the current and subsequent frames poses ch…

Psychological safety in software workplaces: A systematic literature review
Beatriz Santana, Lidiv\^anio Monte, Bianca Santana de Ara\'ujo Silva, Glauco Carneiro, S\'avio Freire, Jos\'e Amancio Macedo Santos, Manoel Mendon\c{c}a
https://arxiv.org/abs/2508.03369

Psychological safety in software workplaces: A systematic literature review
Context: Psychological safety (PS) is an important factor influencing team well-being and performance, particularly in collaborative and dynamic domains such as software development. Despite its acknowledged significance, research on PS within the field of software engineering remains limited. The socio-technical complexities and fast-paced nature of software development present challenges to cultivating PS. To the best of our knowledge, no systematic secondary study has synthesized existing kn…

ADSeeker: A Knowledge-Infused Framework for Anomaly Detection and Reasoning
Kai Zhang, Zekai Zhang, Xihe Sun, Jingmeng Nie, Qinghui Chen, Han Hao, Jianyuan Guo, Jinglin Zhang
https://arxiv.org/abs/2508.03088

ADSeeker: A Knowledge-Infused Framework for Anomaly Detection and Reasoning
Automatic vision inspection holds significant importance in industry inspection. While multimodal large language models (MLLMs) exhibit strong language understanding capabilities and hold promise for this task, their performance remains significantly inferior to that of human experts. In this context, we identify two key challenges: (i) insufficient integration of anomaly detection (AD) knowledge during pre-training, and (ii) the lack of technically precise and conte-aware language generation f…

Diffusion-RL Based Air Traffic Conflict Detection and Resolution Method
Tonghe Li, Jixin Liu, Weili Zeng, Hao Jiang
https://arxiv.org/abs/2509.03550 https://

Diffusion-RL Based Air Traffic Conflict Detection and Resolution Method
In the context of continuously rising global air traffic, efficient and safe Conflict Detection and Resolution (CD&R) is paramount for air traffic management. Although Deep Reinforcement Learning (DRL) offers a promising pathway for CD&R automation, existing approaches commonly suffer from a "unimodal bias" in their policies. This leads to a critical lack of decision-making flexibility when confronted with complex and dynamic constraints, often resulting in "decision deadlocks." To overcome thi…

Diffusion Generative Models Meet Compressed Sensing, with Applications to Image Data and Financial Time Series
Zhengyi Guo, Jiatu Li, Wenpin Tang, David D. Yao
https://arxiv.org/abs/2509.03898

Diffusion Generative Models Meet Compressed Sensing, with Applications to Image Data and Financial Time Series
This paper develops dimension reduction techniques for accelerating diffusion model inference in the context of synthetic data generation. The idea is to integrate compressed sensing into diffusion models: (i) compress the data into a latent space, (ii) train a diffusion model in the latent space, and (iii) apply a compressed sensing algorithm to the samples generated in the latent space, facilitating the efficiency of both model training and inference. Under suitable sparsity assumptions on da…

Analysis of the Supernova Remnant IC 443 using H.E.S.S. Data
Alison M. W. Mitchell (for the H.E.S.S. Collaboration), Lukas Grosspietsch (for the H.E.S.S. Collaboration), Tina Wach (for the H.E.S.S. Collaboration)
https://arxiv.org/abs/2510.02843

Analysis of the Supernova Remnant IC 443 using H.E.S.S. Data
IC 443 is a well-known supernova remnant that stands out due to its interaction with a dense molecular cloud, creating a complex environment where shocks can efficiently accelerate particles to high energies. This makes it a key target for investigating the mechanisms of cosmic-ray acceleration and gamma-ray production, particularly in the context of supernova remnants as potential sources of PeV cosmic rays. This work presents a first analysis of the region as observed by H.E.S.S.. We detect e…

A-MemGuard: A Proactive Defense Framework for LLM-Based Agent Memory
Qianshan Wei, Tengchao Yang, Yaochen Wang, Xinfeng Li, Lijun Li, Zhenfei Yin, Yi Zhan, Thorsten Holz, Zhiqiang Lin, XiaoFeng Wang
https://arxiv.org/abs/2510.02373

A-MemGuard: A Proactive Defense Framework for LLM-Based Agent Memory
Large Language Model (LLM) agents use memory to learn from past interactions, enabling autonomous planning and decision-making in complex environments. However, this reliance on memory introduces a critical security risk: an adversary can inject seemingly harmless records into an agent's memory to manipulate its future behavior. This vulnerability is characterized by two core aspects: First, the malicious effect of injected records is only activated within a specific context, making them hard t…

VLMQ: Efficient Post-Training Quantization for Large Vision-Language Models via Hessian Augmentation
Yufei Xue, Yushi Huang, Jiawei Shao, Jun Zhang
https://arxiv.org/abs/2508.03351

VLMQ: Efficient Post-Training Quantization for Large Vision-Language Models via Hessian Augmentation
Post-training quantization (PTQ) has emerged as an effective approach for compressing large models and accelerating their inference without retraining. While PTQ has been extensively studied in the context of large language models (LLMs), its applicability to vision-language models (VLMs) remains underexplored. In this paper, we identify a modality discrepancy (\emph{i.e.}, limited text tokens \emph{vs.} excessive and redundant vision tokens) of VLMs. However, existing Hessian-based LLM PTQ met…

Multimodal Function Vectors for Spatial Relations
Shuhao Fu, Esther Goldberg, Ying Nian Wu, Hongjing Lu
https://arxiv.org/abs/2510.02528 https://arxiv.org/…

Multimodal Function Vectors for Spatial Relations
Large Multimodal Models (LMMs) demonstrate impressive in-context learning abilities from limited multimodal demonstrations, yet the internal mechanisms supporting such task learning remain opaque. Building on prior work of large language models, we show that a small subset of attention heads in the vision-language model OpenFlamingo-4B is responsible for transmitting representations of spatial relations. The activations of these attention heads, termed function vectors, can be extracted and man…

ClinicalFMamba: Advancing Clinical Assessment using Mamba-based Multimodal Neuroimaging Fusion
Meng Zhou, Farzad Khalvati
https://arxiv.org/abs/2508.03008 https://

ClinicalFMamba: Advancing Clinical Assessment using Mamba-based Multimodal Neuroimaging Fusion
Multimodal medical image fusion integrates complementary information from different imaging modalities to enhance diagnostic accuracy and treatment planning. While deep learning methods have advanced performance, existing approaches face critical limitations: Convolutional Neural Networks (CNNs) excel at local feature extraction but struggle to model global context effectively, while Transformers achieve superior long-range modeling at the cost of quadratic computational complexity, limiting cl…

An invertible generative model for forward and inverse problems
Tristan van Leeuwen, Christoph Brune, Marcello Carioni
https://arxiv.org/abs/2509.03910 https://

An invertible generative model for forward and inverse problems
We formulate the inverse problem in a Bayesian framework and aim to train a generative model that allows us to simulate (i.e., sample from the likelihood) and do inference (i.e., sample from the posterior). We review the use of triangular normalizing flows for conditional sampling in this context and show how to combine two such triangular maps (an upper and a lower one) in to one invertible mapping that can be used for simulation and inference. We work out several useful properties of this inv…

Efficient Odd-One-Out Anomaly Detection
Silvio Chito, Paolo Rabino, Tatiana Tommasi
https://arxiv.org/abs/2509.04326 https://arxiv.org/pdf/2509.04326

Efficient Odd-One-Out Anomaly Detection
The recently introduced odd-one-out anomaly detection task involves identifying the odd-looking instances within a multi-object scene. This problem presents several challenges for modern deep learning models, demanding spatial reasoning across multiple views and relational reasoning to understand context and generalize across varying object categories and layouts. We argue that these challenges must be addressed with efficiency in mind. To this end, we propose a DINO-based model that reduces th…

BitsAI-Fix: LLM-Driven Approach for Automated Lint Error Resolution in Practice
Yuanpeng Li, Qi Long, Zhiyuan Yao, Jian Xu, Lintao Xie, Xu He, Lu Geng, Xin Han, Yueyan Chen, Wenbo Duan
https://arxiv.org/abs/2508.03487

BitsAI-Fix: LLM-Driven Approach for Automated Lint Error Resolution in Practice
As enterprise codebases continue to grow in scale and complexity, the volume of lint errors far exceeds engineers' manual remediation capacity, leading to continuous accumulation of technical debt and hindered development efficiency. This paper presents BitsAI-Fix, an automated lint error remediation workflow based on Large Language Models (LLMs), designed to address this critical challenge in industrial-scale environments. BitsAI-Fix employs tree-sitter for context expansion and generates sear…

StructCoh: Structured Contrastive Learning for Context-Aware Text Semantic Matching
Chao Xue, Ziyuan Gao
https://arxiv.org/abs/2509.02033 https://arxiv.org…

StructCoh: Structured Contrastive Learning for Context-Aware Text Semantic Matching
Text semantic matching requires nuanced understanding of both structural relationships and fine-grained semantic distinctions. While pre-trained language models excel at capturing token-level interactions, they often overlook hierarchical structural patterns and struggle with subtle semantic discrimination. In this paper, we proposed StructCoh, a graph-enhanced contrastive learning framework that synergistically combines structural reasoning with representation space optimization. Our approach …

The H.E.S.S. Gravitational Wave and Gamma-Ray Burst Follow-Up Programs
Bernardo Cornejo (on behalf of the H.E.S.S. Collaboration), Halim Ashkar (on behalf of the H.E.S.S. Collaboration), Matteo Cerruti (on behalf of the H.E.S.S. Collaboration), Ilja Jaroschewski (on behalf of the H.E.S.S. Collaboration), Pierre Pichard (on behalf of the H.E.S.S. Collaboration), Santiago Pita (on behalf of the H.E.S.S. Collaboration), Fabian Schussler (on behalf of the H.E.S.S. Collaboration)

The H.E.S.S. Gravitational Wave and Gamma-Ray Burst Follow-Up Programs
Multi-wavelength and multi-messenger astrophysics have experienced rapid growth over the past decade, seeking a complete picture of different cosmic phenomena. Transient sources, in particular, benefit from the input of multi-messenger observations, offering complementary perspectives on the same event while maximizing the detection probability of a rapidly fading signal. In this context, Gravitational Wave (GW) detections serve as perfect triggers for potential counterpart detections. Notably,…

ArcMemo: Abstract Reasoning Composition with Lifelong LLM Memory
Matthew Ho, Chen Si, Zhaoxiang Feng, Fangxu Yu, Zhijian Liu, Zhiting Hu, Lianhui Qin
https://arxiv.org/abs/2509.04439

ArcMemo: Abstract Reasoning Composition with Lifelong LLM Memory
While inference-time scaling enables LLMs to carry out increasingly long and capable reasoning traces, the patterns and insights uncovered during these traces are immediately discarded once the context window is reset for a new query. External memory is a natural way to persist these discoveries, and recent work has shown clear benefits for reasoning-intensive tasks. We see an opportunity to make such memories more broadly reusable and scalable by moving beyond instance-based memory entries (e.…

Less is More: Token-Efficient Video-QA via Adaptive Frame-Pruning and Semantic Graph Integration
Shaoguang Wang (The Hong Kong University of Science and Technology), Jianxiang He (The Hong Kong University of Science and Technology), Yijie Xu (The Hong Kong University of Science and Technology), Ziyang Chen (The Hong Kong University of Science and Technology), Weiyu Guo (The Hong Kong University of Science and Technology), Hui Xiong (The Hong Kong University of Science and Technology)

Less is More: Token-Efficient Video-QA via Adaptive Frame-Pruning and Semantic Graph Integration
The practical application of Multimodal Large Language Models (MLLMs) to Video Question Answering (Video-QA) is severely hindered by the high token cost of processing numerous video frames. While increasing the number of sampled frames is a common strategy, we observe a "less is more" phenomenon where excessive frames can paradoxically degrade performance due to context dilution. Concurrently, state-of-the-art keyframe selection methods, while effective, still yield significant temporal redunda…

MultiWikiQA: A Reading Comprehension Benchmark in 300 Languages
Dan Saattrup Smart
https://arxiv.org/abs/2509.04111 https://arxiv.org/pdf/2509.04111

MultiWikiQA: A Reading Comprehension Benchmark in 300+ Languages
We introduce a new reading comprehension dataset, dubbed MultiWikiQA, which covers 306 languages. The context data comes from Wikipedia articles, with questions generated by an LLM and the answers appearing verbatim in the Wikipedia articles. We conduct a crowdsourced human evaluation of the fluency of the generated questions across 30 of the languages, providing evidence that the questions are of good quality. We evaluate 6 different language models, both decoder and encoder models of varying …

Mitigating Modal Imbalance in Multimodal Reasoning
Chen Henry Wu, Neil Kale, Aditi Raghunathan
https://arxiv.org/abs/2510.02608 https://arxiv.org/pdf/2510.…

Mitigating Modal Imbalance in Multimodal Reasoning
Foundation models (FMs) deployed in real-world tasks such as computer-use agents must integrate diverse modalities. How good are FMs at performing joint reasoning, simultaneously reasoning over multiple modalities, especially when the modalities interact and relate to each other to form cross-modal context? To better understand this problem, we study FMs on cross-modal conflicts: scenarios where conflicting evidence is presented across modalities. This allows us to examine whether FMs prioritiz…

Memory Forcing: Spatio-Temporal Memory for Consistent Scene Generation on Minecraft
Junchao Huang, Xinting Hu, Boyao Han, Shaoshuai Shi, Zhuotao Tian, Tianyu He, Li Jiang
https://arxiv.org/abs/2510.03198

Memory Forcing: Spatio-Temporal Memory for Consistent Scene Generation on Minecraft
Autoregressive video diffusion models have proved effective for world modeling and interactive scene generation, with Minecraft gameplay as a representative application. To faithfully simulate play, a model must generate natural content while exploring new scenes and preserve spatial consistency when revisiting explored areas. Under limited computation budgets, it must compress and exploit historical cues within a finite context window, which exposes a trade-off: Temporal-only memory lacks long…

Replaced article(s) found for cs.CL. https://arxiv.org/list/cs.CL/new
[2/3]:
- Context Reasoner: Incentivizing Reasoning Capability for Contextualized Privacy and Safety Compli...
Hu, Li, Jing, Hu, Zeng, Han, Xu, Chu, Hu, Song

Efficient Multi-Slide Visual-Language Feature Fusion for Placental Disease Classification
Hang Guo, Qing Zhang, Zixuan Gao, Siyuan Yang, Shulin Peng, Xiang Tao, Ting Yu, Yan Wang, Qingli Li
https://arxiv.org/abs/2508.03277

Efficient Multi-Slide Visual-Language Feature Fusion for Placental Disease Classification
Accurate prediction of placental diseases via whole slide images (WSIs) is critical for preventing severe maternal and fetal complications. However, WSI analysis presents significant computational challenges due to the massive data volume. Existing WSI classification methods encounter critical limitations: (1) inadequate patch selection strategies that either compromise performance or fail to sufficiently reduce computational demands, and (2) the loss of global histological context resulting fr…

Black-box Context-free Grammar Inference for Readable & Natural Grammars
Mohammad Rifat Arefin, Shanto Rahman, Christoph Csallner
https://arxiv.org/abs/2509.26616 https://…

Black-box Context-free Grammar Inference for Readable & Natural Grammars
Black-box context-free grammar inference is crucial for program analysis, reverse engineering, and security, yet existing tools such as Arvada, TreeVada, and Kedavra struggle with scalability, readability, and accuracy on large, complex languages. We present NatGI, a novel LLM-guided grammar inference framework that extends TreeVada's parse tree recovery with three key innovations: bracket-guided bubble exploration, LLM- driven bubble generation and non-terminal labeling, and hierarchical delta…

Pretrain-Test Task Alignment Governs Generalization in In-Context Learning
Mary I. Letey, Jacob A. Zavatone-Veth, Yue M. Lu, Cengiz Pehlevan
https://arxiv.org/abs/2509.26551 htt…

Pretrain-Test Task Alignment Governs Generalization in In-Context Learning
In-context learning (ICL) is a central capability of Transformer models, but the structures in data that enable its emergence and govern its robustness remain poorly understood. In this work, we study how the structure of pretraining tasks governs generalization in ICL. Using a solvable model for ICL of linear regression by linear attention, we derive an exact expression for ICL generalization error in high dimensions under arbitrary pretraining-testing task covariance mismatch. This leads to a…

Self-Anchor: Large Language Model Reasoning via Step-by-step Attention Alignment
Hongxiang Zhang, Yuan Tian, Tianyi Zhang
https://arxiv.org/abs/2510.03223 https://

Self-Anchor: Large Language Model Reasoning via Step-by-step Attention Alignment
To solve complex reasoning tasks for Large Language Models (LLMs), prompting-based methods offer a lightweight alternative to fine-tuning and reinforcement learning. However, as reasoning chains extend, critical intermediate steps and the original prompt will be buried in the context, receiving insufficient attention and leading to errors. In this paper, we propose Self-Anchor, a novel pipeline that leverages the inherent structure of reasoning to steer LLM attention. Self-Anchor decomposes rea…

Understanding and Leveraging the Expert Specialization of Context Faithfulness in Mixture-of-Experts LLMs
Jun Bai, Minghao Tong, Yang Liu, Zixia Jia, Zilong Zheng
https://arxiv.org/abs/2508.19594

Understanding and Leveraging the Expert Specialization of Context Faithfulness in Mixture-of-Experts LLMs
Context faithfulness is essential for reliable reasoning in context-dependent scenarios. However, large language models often struggle to ground their outputs in the provided context, resulting in irrelevant responses. Inspired by the emergent expert specialization observed in mixture-of-experts architectures, this work investigates whether certain experts exhibit specialization in context utilization, offering a potential pathway toward targeted optimization for improved context faithfulness. …

Personalized Vision via Visual In-Context Learning
Yuxin Jiang, Yuchao Gu, Yiren Song, Ivor Tsang, Mike Zheng Shou
https://arxiv.org/abs/2509.25172 https://

Personalized Vision via Visual In-Context Learning
Modern vision models, trained on large-scale annotated datasets, excel at predefined tasks but struggle with personalized vision -- tasks defined at test time by users with customized objects or novel objectives. Existing personalization approaches rely on costly fine-tuning or synthetic data pipelines, which are inflexible and restricted to fixed task formats. Visual in-context learning (ICL) offers a promising alternative, yet prior methods confine to narrow, in-domain tasks and fail to gener…

Pack and Force Your Memory: Long-form and Consistent Video Generation
Xiaofei Wu, Guozhen Zhang, Zhiyong Xu, Yuan Zhou, Qinglin Lu, Xuming He
https://arxiv.org/abs/2510.01784 ht…

Pack and Force Your Memory: Long-form and Consistent Video Generation
Long-form video generation presents a dual challenge: models must capture long-range dependencies while preventing the error accumulation inherent in autoregressive decoding. To address these challenges, we make two contributions. First, for dynamic context modeling, we propose MemoryPack, a learnable context-retrieval mechanism that leverages both textual and image information as global guidance to jointly model short- and long-term dependencies, achieving minute-level temporal consistency. Th…