Tootfinder

Git Context Controller: Manage the Context of LLM-based Agents like Git
Junde Wu
https://arxiv.org/abs/2508.00031 https://arxiv.org/pdf/2508.00031

Git Context Controller: Manage the Context of LLM-based Agents like Git
Large language model (LLM) based agents have shown impressive capabilities by interleaving internal reasoning with external tool use. However, as these agents are deployed in long-horizon workflows, such as coding for a big, long-term project, context management becomes a critical bottleneck. We introduce Git-Context-Controller (GCC), a structured context management framework inspired by software version control systems. GCC elevates context as versioned memory hierarchy like Git. It structures…

OMD at the Greek Theater in Los Angeles, performing 'Anthropocene' June 21, 2025.
https://open.spotify.com/track/1zTghGUFb7Ajrn5ghO9ZQl?si=XNxEhurBRXaWIu3kJmY42Q&context=spotify:album:2yW1JUK…

Anthropocene
Orchestral Manoeuvres In The Dark · Bauhaus Staircase (The Complete Edition) · Song · 2023

Out-of-Context Abduction: LLMs Make Inferences About Procedural Data Leveraging Declarative Facts in Earlier Training Data
Sohaib Imran, Rob Lamb, Peter M. Atkinson
https://arxiv.org/abs/2508.00741

Out-of-Context Abduction: LLMs Make Inferences About Procedural Data Leveraging Declarative Facts in Earlier Training Data
Large language models (LLMs) are trained on large corpora, yet it is unclear whether they can reason about the information present within their training data. We design experiments to study out-of-context abduction in LLMs, the ability to infer the most plausible explanations for observations using relevant facts present in training data. We train treatment LLMs on names and behavior descriptions of fictitious chatbots, but not on examples of dialogue with the chatbots. We find that OpenAI's GP…

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…

Medical Data Pecking: A Context-Aware Approach for Automated Quality Evaluation of Structured Medical Data
Irena Girshovitz, Atai Ambus, Moni Shahar, Ran Gilad-Bachrach
https://arxiv.org/abs/2507.02628

Medical Data Pecking: A Context-Aware Approach for Automated Quality Evaluation of Structured Medical Data
Background: The use of Electronic Health Records (EHRs) for epidemiological studies and artificial intelligence (AI) training is increasing rapidly. The reliability of the results depends on the accuracy and completeness of EHR data. However, EHR data often contain significant quality issues, including misrepresentations of subpopulations, biases, and systematic errors, as they are primarily collected for clinical and billing purposes. Existing quality assessment methods remain insufficient, la…

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…

Linear Attention with Global Context: A Multipole Attention Mechanism for Vision and Physics
Alex Colagrande, Paul Caillon, Eva Feillet, Alexandre Allauzen
https://arxiv.org/abs/2507.02748

Linear Attention with Global Context: A Multipole Attention Mechanism for Vision and Physics
Transformers have become the de facto standard for a wide range of tasks, from image classification to physics simulations. Despite their impressive performance, the quadratic complexity of standard Transformers in both memory and time with respect to the input length makes them impractical for processing high-resolution inputs. Therefore, several variants have been proposed, the most successful relying on patchification, downsampling, or coarsening techniques, often at the cost of losing the f…

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 …

A developer details how he shipped Context, a native macOS app that was almost 100% built using Claude Code (Indragie Karunaratne/My Portfolio)
https://www.indragie.com/blog/i-shipped-a-macos-app-built-entirely-by-claude-code

I Shipped a macOS App Built Entirely by Claude Code
How I built Context—a native macOS SwiftUI app for debugging MCP servers—almost entirely with Claude Code, and what I learned about building with AI coding agents.

Context-aware gate set tomography: Improving the self-consistent characterization of trapped-ion universal gate sets by leveraging non-Markovianity
Pablo Vi\~nas, Alejandro Bermudez
https://arxiv.org/abs/2507.02542

Context-aware gate set tomography: Improving the self-consistent characterization of trapped-ion universal gate sets by leveraging non-Markovianity
To progress in the characterization of noise for current quantum computers, gate set tomography (GST) has emerged as a self-consistent tomographic protocol that can accurately estimate the complete set of noisy quantum gates, state preparations, and measurements. In its original incarnation, GST improves the estimation precision by applying the gates sequentially, provided that the noise makes them a set of fixed completely-positive and trace preserving (CPTP) maps independent of the history of…

Sonify Anything: Towards Context-Aware Sonic Interactions in AR
Laura Sch\"utz, Sasan Matinfar, Ulrich Eck, Daniel Roth, Nassir Navab
https://arxiv.org/abs/2508.01789 https…

Sonify Anything: Towards Context-Aware Sonic Interactions in AR
In Augmented Reality (AR), virtual objects interact with real objects. However, the lack of physicality of virtual objects leads to the absence of natural sonic interactions. When virtual and real objects collide, either no sound or a generic sound is played. Both lead to an incongruent multisensory experience, reducing interaction and object realism. Unlike in Virtual Reality (VR) and games, where predefined scenes and interactions allow for the playback of pre-recorded sound samples, AR requi…

Placing the Near-Earth Object Impact Probability in Context
C. R. Nugent, K. P. Andersen, James M. Bauer, C. T. Jensen, L. K. Kristiansen, C. P. Hansen, M. M. Nielsen, C. F. Vesterg{\aa}rd
https://arxiv.org/abs/2508.02418

Placing the Near-Earth Object Impact Probability in Context
Near-Earth objects (NEOs) have the potential to cause extensive damage and loss of life on Earth. Advancements in NEO discovery, trajectory prediction, and deflection technology indicate that an impact could be prevented, with sufficient warning time. We derive an impact frequency of NEOs 140m and larger, using the NEOMOD2 NEO population model and JPL Horizons. We then place that frequency in context with other preventable causes of death; allowing for comparison between a planet-wide event and…

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…

Totoro spoilers
Re the quoted post from @…:
The very first time I saw My Neighbor Totoro, it was because a friend cajoled me into attending the student anime club’s showing with absolutely no context whatsoever except “you •have• to see this movie, Paul.” I had no idea what genre it was. I had no idea that it was a movie considered suitable for kids. The last anime I’d seen was IIRC Ghost in the Shell; I was ready for anything.
When Mei went missing, I thought, “omg, is this a tragedy? I think this story is a tragedy!” The whole time they were looking for her, I was absolutely terrified. I thought for sure they’d found her sandal. I still tear up when they find her now, on every rewatching.
I’m so glad for that first viewing. It’s a much better movie that way. When you’re expecting an innocent movie about cute forest plushies, you see that. But when you see that it could be a tragedy — the mother’s shadowy illness, the lost child — it hits hard. https://wandering.shop/@Violinknitter/114792748619083661

A Context-Aware Dual-Metric Framework for Confidence Estimation in Large Language Models
Mingruo Yuan, Shuyi Zhang, Ben Kao
https://arxiv.org/abs/2508.00600 https://

A Context-Aware Dual-Metric Framework for Confidence Estimation in Large Language Models
Accurate confidence estimation is essential for trustworthy large language models (LLMs) systems, as it empowers the user to determine when to trust outputs and enables reliable deployment in safety-critical applications. Current confidence estimation methods for LLMs neglect the relevance between responses and contextual information, a crucial factor in output quality evaluation, particularly in scenarios where background knowledge is provided. To bridge this gap, we propose CRUX (Context-awar…

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…

«I believe that the web, like Nokia, was created to connect people. Some create things, some watch or read those, and connections are formed. LLMs […] destroy the human element of the web, and of the real world too, because those are too interconnected now. […] By using LLMs in a collaborative context, you cheat: you introduce errors and invite everybody to find them»
A really lovely interview with @… ! Come for the map talk, stay for the community insight
https://www.apc.org/en/news/every-door-going-map-user-open-source-map-creator

[Thread] an NYT standards editor responds to criticism of its recent Zohran Mamdani story, says the "ultimate source" was Columbia data that Mamdani confirmed (Patrick Healy/@patrickhealynyt)
https://x.com/patrickhealynyt/status/1941262786006483418

Patrick Healy (@patrickhealynyt) on X
As the @nytimes assistant managing editor for Standards and Trust, I’ve received reader feedback regarding our reporting on Zohran Mamdani’s 2009 application to Columbia University. To provide context on how the reporting came together, I wanted to share some information:

M3LLM: Model Context Protocol-aided Mixture of Vision Experts For Multimodal LLMs in Networks
Yongjie Zeng, Hongyang Du
https://arxiv.org/abs/2508.01805 https://

M3LLM: Model Context Protocol-aided Mixture of Vision Experts For Multimodal LLMs in Networks
Current Multimodal Large Language Models (MLLMs) rely on centralized architectures and often suffer from poor alignment between the input task and their fixed visual encoding modules, which limits performance on diverse and dynamic visual tasks. With the increasing deployment of resource-efficient models on edge devices in wireless networks, a new opportunity emerges to dynamically use distributed vision experts for improved MLLM inference quality. To enable this, we propose M3LLM, where the Mo…

If you're a storyteller of any kind, without any context, which of these terms do you find more appealing?
Optional: in a comment -- share your associations with the term(s), or why you prefer one or the other, or another you prefer!
Thank you!

The new Compendium is here, and it’s more than just data. [Part 1]
Created by and for the #NREN community, this year’s edition brings the 2024 survey results to life through stories, context, and reflections from NRENs and expert contributors.
🔗 Read the report and explore the data:

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

Decision Theory For Large Scale Outlier Detection Using Aleatoric Uncertainty: With a Note on Bayesian FDR
Ryan Warnick
https://arxiv.org/abs/2508.01988 https://

Decision Theory For Large Scale Outlier Detection Using Aleatoric Uncertainty: With a Note on Bayesian FDR
Aleatoric and Epistemic uncertainty have achieved recent attention in the literature as different sources from which uncertainty can emerge in stochastic modeling. Epistemic being intrinsic or model based notions of uncertainty, and aleatoric being the uncertainty inherent in the data. We propose a novel decision theoretic framework for outlier detection in the context of aleatoric uncertainty; in the context of Bayesian modeling. The model incorporates bayesian false discovery rate control for…

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…

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 …

Visual Contextual Attack: Jailbreaking MLLMs with Image-Driven Context Injection
Ziqi Miao, Yi Ding, Lijun Li, Jing Shao
https://arxiv.org/abs/2507.02844 h…

Visual Contextual Attack: Jailbreaking MLLMs with Image-Driven Context Injection
With the emergence of strong visual-language capabilities, multimodal large language models (MLLMs) have demonstrated tremendous potential for real-world applications. However, the security vulnerabilities exhibited by the visual modality pose significant challenges to deploying such models in open-world environments. Recent studies have successfully induced harmful responses from target MLLMs by encoding harmful textual semantics directly into visual inputs. However, in these approaches, the v…

„My new system is, simply, no system at all. I write what I think. I delete what I don’t need. I don’t capture everything. I don’t try to. I read what I feel like. I think in conversation, in movement, in context. I don’t build a second brain. I inhabit the first.”
The PKM-movement never took off for me since it feels like building and bulding and building… Just one more note! It’s like productivity snake-oil.

I Deleted My Second Brain
Why I Erased 10,000 Notes, 7 Years of Ideas, and Every Thought I Tried to Save

Users say AI notetaking tools for meetings can misinterpret context when generating summaries or share content meant for a select audience with all participants (Ann-Marie Alcšntara/Wall Street Journal)
https://www.wsj.com/tech/ai/ai-notetaker-m

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…

Analysing Temporal Reasoning in Description Logics Using Formal Grammars
Camille Bourgaux, Anton Gnatenko, Micha\"el Thomazo
https://arxiv.org/abs/2508.00575 https://

Analysing Temporal Reasoning in Description Logics Using Formal Grammars
We establish a correspondence between (fragments of) $\mathcal{TEL}^\bigcirc$, a temporal extension of the $\mathcal{EL}$ description logic with the LTL operator $\bigcirc^k$, and some specific kinds of formal grammars, in particular, conjunctive grammars (context-free grammars equipped with the operation of intersection). This connection implies that $\mathcal{TEL}^\bigcirc$ does not possess the property of ultimate periodicity of models, and further leads to undecidability of query answering …

SleepWalk: Exploiting Context Switching and Residual Power for Physical Side-Channel Attacks
Sahan Sanjaya, Aruna Jayasena, Prabhat Mishra
https://arxiv.org/abs/2507.22306 https…

SleepWalk: Exploiting Context Switching and Residual Power for Physical Side-Channel Attacks
Context switching is utilized by operating systems to change the execution context between application programs. It involves saving and restoring the states of multiple registers and performing a pipeline flush to remove any pre-fetched instructions, leading to a higher instantaneous power consumption compared to typical program execution. In this paper, we introduce a physical power side-channel leakage source that exploits the power spike observed during a context switch, triggered by the inb…

The president of El Salvador would not recognize Justice if it indicted him in The Hague and sentenced him for his crimes.
But let’s do it anyway just to be sure.
Context: https://apnews.com/article/kilmar-abrego-…

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

Multi-Focus Probes for Context-Preserving Network Exploration and Interaction in Immersive Analytics
Eric Zimmermann, Stefan Bruckner
https://arxiv.org/abs/2507.01140

Multi-Focus Probes for Context-Preserving Network Exploration and Interaction in Immersive Analytics
Immersive visualization of network data enables users to physically navigate and interact with complex structures, but managing transitions between detailed local (egocentric) views and global (exocentric) overviews remains a major challenge. We present a multifocus probe technique for immersive environments that allows users to instantiate multiple egocentric subgraph views while maintaining persistent links to the global network context. Each probe acts as a portable local focus, enabling fin…

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…

Homotopy equivalence of digital pictures in $\mathbb{Z}^2$
Dae-Woong Lee, P. Christopher Staecker
https://arxiv.org/abs/2509.03023 https://arxiv.org/pdf/25…

Homotopy equivalence of digital pictures in $\mathbb{Z}^2$
We investigate the properties of digital homotopy in the context of digital pictures $(X,κ,\bar κ)$, where $X\subsetneq \mathbb{Z}^n$ is a finite set, $κ$ is an adjacency relation on $X$, and $\bar κ$ is an adjacency relation on the complement of $X$. In particular we focus on homotopy equivalence between digital pictures in $\mathbb{Z}^2$. We define a numerical homotopy-type invariant for digital pictures in $\mathbb{Z}^2$ called the outer perimeter, which is a basic tool for distinguishin…

Exposing and Mitigating Calibration Biases and Demographic Unfairness in MLLM Few-Shot In-Context Learning for Medical Image Classification
Xing Shen, Justin Szeto, Mingyang Li, Hengguan Huang, Tal Arbel
https://arxiv.org/abs/2506.23298

Exposing and Mitigating Calibration Biases and Demographic Unfairness in MLLM Few-Shot In-Context Learning for Medical Image Classification
Multimodal large language models (MLLMs) have enormous potential to perform few-shot in-context learning in the context of medical image analysis. However, safe deployment of these models into real-world clinical practice requires an in-depth analysis of the accuracies of their predictions, and their associated calibration errors, particularly across different demographic subgroups. In this work, we present the first investigation into the calibration biases and demographic unfairness of MLLMs'…

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 …

New orientable sequences
Chris J Mitchell, Peter R Wild
https://arxiv.org/abs/2507.02526 https://arxiv.org/pdf/2507.02526

New orientable sequences
Orientable sequences of order n are infinite periodic sequences with symbols drawn from a finite alphabet of size k with the property that any particular subsequence of length n occurs at most once in a period in either direction. They were introduced in the early 1990s in the context of possible applications in position sensing. Bounds on the period of such sequences and a range of methods of construction have been devised, although apart from very small cases a significant gap remains between…

Dynamic Forgetting and Spatio-Temporal Periodic Interest Modeling for Local-Life Service Recommendation
Zhaoyu Hu, Hao Guo, Yuan Tian, Erpeng Xue, Jianyang Wang, Xianyang Qi, Hongxiang Lin, Lei Wang, Sheng Chen
https://arxiv.org/abs/2508.02451

Dynamic Forgetting and Spatio-Temporal Periodic Interest Modeling for Local-Life Service Recommendation
In the context of the booming digital economy, recommendation systems, as a key link connecting users and numerous services, face challenges in modeling user behavior sequences on local-life service platforms, including the sparsity of long sequences and strong spatio-temporal dependence. Such challenges can be addressed by drawing an analogy to the forgetting process in human memory. This is because users' responses to recommended content follow the recency effect and the cyclicality of memory…

J-VAR: Analysis of RR Lyrae light curves in seven optical bands
Sayali Kulkarni, H\'ector V\'azquez Rami\'o, Carlos L\'opez-SanJuan, Alessandro Ederoclite, Stylianos Pyrzas
https://arxiv.org/abs/2509.03120

J-VAR: Analysis of RR Lyrae light curves in seven optical bands
Context. RR Lyrae stars are key distance indicators and tracers of old stellar populations in the Milky Way and its satellites. Accurate period and amplitude measurements are essential for constraining stellar evolution and mapping Galactic structure. The Javalambre VARiability (J-VAR) survey provides valuable time-series data across seven optical bands, including gri and four medium and narrow bands. Aims. Our goal is to construct and analyze light curves for RR Lyrae stars identified in the f…

Dam Management in the Era of Climate Change
Cristina Di Girolami, M'hamed Mrad, Ga\"igi, Vathana Ly Vath, Simone Scotti
https://arxiv.org/abs/2508.02636 https://

Dam Management in the Era of Climate Change
Climate change has a dramatic impact, particularly by concentrating rainfall into a few short periods, interspersed by long dry spells. In this context, the role of dams is crucial. We consider the optimal control of a dam, where the water level must not exceed a designated safety threshold, nor fall below a minimum level to ensure functionality and sustainability for for the outgoing river. To model dry spells and intense rainfall events, commonly referred to as water bombs, we introduce a Haw…

Having legal protection insurance is always a good idea to, but IMHO it’s absolutely essential for professors due to the particular context in which we work. But I find that it’s hardly ever discussed.
An obvious example is the question of the limits of academic freedom, but many potential issues are way more trivial. ⇢
#AcademicChatter

Having legal protection insurance is always a good idea to, but IMHO it’s absolutely essential for professors due to the particular context in which we work. But I find that it’s hardly ever discussed.
An obvious example is the question of the limits of academic freedom, but many potential issues are way more trivial. ⇢
#AcademicChatter

Having legal protection insurance is always a good idea to, but IMHO it’s absolutely essential for professors due to the particular context in which we work. But I find that it’s hardly ever discussed.
An obvious example is the question of the limits of academic freedom, but many potential issues are way more trivial. ⇢
#AcademicChatter

BioBlue: Notable runaway-optimiser-like LLM failure modes on biologically and economically aligned AI safety benchmarks for LLMs with simplified observation format
Roland Pihlakas, Sruthi Kuriakose
https://arxiv.org/abs/2509.02655

BioBlue: Notable runaway-optimiser-like LLM failure modes on biologically and economically aligned AI safety benchmarks for LLMs with simplified observation format
Relatively many past AI safety discussions have centered around the dangers of unbounded utility maximisation by RL agents, illustrated by scenarios like the "paperclip maximiser" or by specification gaming in general. Unbounded maximisation is problematic for many reasons. We wanted to verify whether these RL runaway optimisation problems are still relevant with LLMs as well. Turns out, strangely, this is indeed clearly the case. The problem is not that the LLMs just lose context or become inc…

Fractal Dimension in Nonlinear Wave Dynamics Governed by a Nonlinear Partial Differential Equation
Saugata Dutta, Kajal Kumar Mondal, Prasanta Chatterjee
https://arxiv.org/abs/2508.01816

Fractal Dimension in Nonlinear Wave Dynamics Governed by a Nonlinear Partial Differential Equation
This work presents a detailed analytical and geometrical investigation of the (2+1)-dimensional Boiti-Leon-Pempinelli system, a nonlinear dispersive model arising in the context of fluid and plasma dynamics. By employing a projective Riccati-based ansatz, a new class of exact solutions is systematically derived. These solutions, when visualized, exhibit intricate geometrical features that evolve across multiple spatial scales. To quantify this complexity, a voxel-based box-counting dimension an…

Two-flavor chirally imbalanced quark matter beyond large-$N_c$
Andr\'e G. da Silva, Dyana C. Duarte, Ricardo L. S. Farias, Marcus Benghi Pinto, Rudnei O. Ramos, William R. Tavares
https://arxiv.org/abs/2509.02955

Two-flavor chirally imbalanced quark matter beyond large-$N_c$
We investigate a chirally imbalanced medium in the context of the two-flavor Nambu--Jona-Lasinio model using both the large-$N_c$ (LN) and beyond large-$N_c$ (BLN) approximations. To incorporate BLN effects, we consider the optimized perturbation theory (OPT) to the first nontrivial order, which includes two-loop (exchange) contributions. This procedure allows us to explicitly explore how finite $N_c$ corrections affect the thermodynamics as well as the phase diagram of chirally imbalanced quar…

Dynamically Adaptive Reasoning via LLM-Guided MCTS for Efficient and Context-Aware KGQA
Yingxu Wang, Shiqi Fan, Mengzhu Wang, Siwei Liu
https://arxiv.org/abs/2508.00719 https://…

Dynamically Adaptive Reasoning via LLM-Guided MCTS for Efficient and Context-Aware KGQA
Knowledge Graph Question Answering (KGQA) aims to interpret natural language queries and perform structured reasoning over knowledge graphs by leveraging their relational and semantic structures to retrieve accurate answers. Recent KGQA methods primarily follow either retrieve-then-reason paradigm, relying on GNNs or heuristic rules for static paths extraction, or dynamic path generation strategies that use large language models (LLMs) with prompting to jointly perform retrieval and reasoning. …

Context-Aware Code Wiring Recommendation with LLM-based Agent
Taiming Wang, Yanjie Jiang, Chunhao Dong, Yuxia Zhang, Hui Liu
https://arxiv.org/abs/2507.01315

Context-Aware Code Wiring Recommendation with LLM-based Agent
Copy-paste-modify is a widespread and pragmatic practice in software development, where developers adapt reused code snippets, sourced from platforms such as Stack Overflow, GitHub, or LLM outputs, into their local codebase. A critical yet underexplored aspect of this adaptation is code wiring, which involves substituting unresolved variables in the pasted code with suitable ones from the surrounding context. Existing solutions either rely on heuristic rules or historical templates, often faili…

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!

Overlapping IPv4, IPv6, and TCP data: exploring errors, test case context and multiple overlaps inside network stacks and NIDSes with PYROLYSE
Lucas Aubard, Johan Mazel, Gilles Guette, Pierre Chifflier
https://arxiv.org/abs/2508.00735

Overlapping IPv4, IPv6, and TCP data: exploring errors, test case context and multiple overlaps inside network stacks and NIDSes with PYROLYSE
IP fragmentation and TCP segmentation allow for splitting large data packets into smaller ones, e.g., for transmission across network links of limited capacity. These mechanisms permit complete or partial overlaps with different data on the overlapping portions. IPv4, IPv6, and TCP reassembly policies, i.e., the data chunk preferences that depend on the overlap types, differ across protocol implementations. This leads to vulnerabilities, as NIDSes may interpret the packet differently from the m…

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 …

@… I don’t have anything I can show yet, but I’m happy to introduce the idea and context:
I run a single user* Mastodon server, which strikes me as incredibly wasteful. I never use the web front end, yet I pay the compose cost of running it. I receive and send relatively few activities, yet the software is running constantly.
So that led me to want …

The term "context engineering" is gaining traction over "prompt engineering" as it better describes the skill of providing LLMs with the necessary information (Simon Willison/Simon Willison's Weblog)
https://simonwillison.net/2025/Jun/27/context-…

Exploring the Design Space of Fair Tree Learning Algorithms
Kiara Stempel, Mattia Cerrato, Stefan Kramer
https://arxiv.org/abs/2509.03204 https://arxiv.org…

Exploring the Design Space of Fair Tree Learning Algorithms
Decision trees have been studied extensively in the context of fairness, aiming to maximize prediction performance while ensuring non-discrimination against different groups. Techniques in this space usually focus on imposing constraints at training time, constraining the search space so that solutions which display unacceptable values of relevant metrics are not considered, discarded, or discouraged. If we assume one target variable y and one sensitive attribute s, the design space of tree lea…

Effective Explanations for Belief-Desire-Intention Robots: When and What to Explain
Cong Wang, Roberto Calandra, Verena Kl\"os
https://arxiv.org/abs/2507.02016

Effective Explanations for Belief-Desire-Intention Robots: When and What to Explain
When robots perform complex and context-dependent tasks in our daily lives, deviations from expectations can confuse users. Explanations of the robot's reasoning process can help users to understand the robot intentions. However, when to provide explanations and what they contain are important to avoid user annoyance. We have investigated user preferences for explanation demand and content for a robot that helps with daily cleaning tasks in a kitchen. Our results show that users want explanatio…

LMAE4Eth: Generalizable and Robust Ethereum Fraud Detection by Exploring Transaction Semantics and Masked Graph Embedding
Yifan Jia, Yanbin Wang, Jianguo Sun, Ye Tian, Peng Qian
https://arxiv.org/abs/2509.03939

LMAE4Eth: Generalizable and Robust Ethereum Fraud Detection by Exploring Transaction Semantics and Masked Graph Embedding
Current Ethereum fraud detection methods rely on context-independent, numerical transaction sequences, failing to capture semantic of account transactions. Furthermore, the pervasive homogeneity in Ethereum transaction records renders it challenging to learn discriminative account embeddings. Moreover, current self-supervised graph learning methods primarily learn node representations through graph reconstruction, resulting in suboptimal performance for node-level tasks like fraud account detec…

LCQNN: Linear Combination of Quantum Neural Networks
Hongshun Yao, Xia Liu, Mingrui Jing, Xin Wang
https://arxiv.org/abs/2507.02832 https://

LCQNN: Linear Combination of Quantum Neural Networks
Quantum neural networks combine quantum computing with advanced data-driven methods, offering promising applications in quantum machine learning. However, the optimal paradigm of quantum neural networks in the context trainability and expressivity remains an open question. To overcome this issue, we first introduce a framework called Linear Combination of Quantum Neural Networks (LCQNN), which leverages the linear combination of unitaries concept to coordinate the composition of QNNs, thereby b…

On the Homogeneous Space of a Pair of Associative Composition Algebras
Mahir Bilen, Ana Casimiro, Ferruh \"Ozbudak
https://arxiv.org/abs/2509.04354 https://

On the Homogeneous Space of a Pair of Associative Composition Algebras
The relationship between associative composition algebras of dimensions 2 and 4 within the context of homogeneous spaces, with a particular focus on Hamiltonian quaternions, is explored. In the special case of Hamiltonian quaternions, the equivariant cohomology rings of the homogeneous spaces are computed to gain a deeper understanding of their topological structure. In the general case, equivariant K-theory is utilized to examine the categories of vector bundles on these spaces.Taking this one…

Computing $c$- and $a$-functions from entanglement
Konstantinos Boutivas, Dimitrios Katsinis, Georgios Pastras, Nikolaos Tetradis
https://arxiv.org/abs/2509.03259 https://

Computing $c$- and $a$-functions from entanglement
We confirm the direct connection between entanglement entropy and the notion of irreversibility in the renormalization-group flow in the context of a simple theory for which a calculation from first principles is feasible. The change of the entanglement entropy for a spherical entangling surface as its radius grows from zero to infinity corresponds to the flow from the UV to the IR. Through analytical and numerical means, we compute the entanglement entropy for a free massive scalar theory, mak…

IC-Custom: Diverse Image Customization via In-Context Learning
Yaowei Li, Xiaoyu Li, Zhaoyang Zhang, Yuxuan Bian, Gan Liu, Xinyuan Li, Jiale Xu, Wenbo Hu, Yating Liu, Lingen Li, Jing Cai, Yuexian Zou, Yancheng He, Ying Shan
https://arxiv.org/abs/2507.01926

IC-Custom: Diverse Image Customization via In-Context Learning
Image customization, a crucial technique for industrial media production, aims to generate content that is consistent with reference images. However, current approaches conventionally separate image customization into position-aware and position-free customization paradigms and lack a universal framework for diverse customization, limiting their applications across various scenarios. To overcome these limitations, we propose IC-Custom, a unified framework that seamlessly integrates position-awa…

A Two-Stage Strategy for Mitosis Detection Using Improved YOLO11x Proposals and ConvNeXt Classification
Jie Xiao, Mengye Lyu, Shaojun Liu
https://arxiv.org/abs/2509.02627 https:…

A Two-Stage Strategy for Mitosis Detection Using Improved YOLO11x Proposals and ConvNeXt Classification
MIDOG 2025 Track 1 requires mitosis detection in whole-slide images (WSIs) containing non-tumor, inflamed, and necrotic regions. Due to the complicated and heterogeneous context, as well as possible artifacts, there are often false positives and false negatives, thus degrading the detection F1-score. To address this problem, we propose a two-stage framework. Firstly, an improved YOLO11x, integrated with EMA attention and LSConv, is employed to generate mitosis candidates. We use a low confidenc…

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…

Understanding and Improving Length Generalization in Recurrent Models
Ricardo Buitrago Ruiz, Albert Gu
https://arxiv.org/abs/2507.02782 https://

Understanding and Improving Length Generalization in Recurrent Models
Recently, recurrent models such as state space models and linear attention have become popular due to their linear complexity in the sequence length. Thanks to their recurrent nature, in principle they can process arbitrarily long sequences, but their performance sometimes drops considerably beyond their training context lengths-i.e. they fail to length generalize. In this work, we provide comprehensive empirical and theoretical analysis to support the unexplored states hypothesis, which posits…

Managing Power Gaps as a Topic of Pair Programming Skill: A Grounded Theory
Linus Ververs, Lutz Prechelt
https://arxiv.org/abs/2508.00462 https://arxiv.org…

Managing Power Gaps as a Topic of Pair Programming Skill: A Grounded Theory
Context: Pair Programming as a work mode is used (occasionally or frequently) throughout professional software development. Objective: Understand what power-related phenomena occur in pair programming as it is used in industry; give advice to practitioners on how to do better pair programming. Method: Analyze 22 industrial pair programming sessions using Grounded Theory Methodology. Formulate a Grounded Theory on power-related behaviors. Run a survey with 292 participants about that theory. Use…

Quantum Key-Recovery Attacks on FBC Algorithm
Yan-Ying Zhu, Bin-Bin Cai, Fei Gao, Song Lin
https://arxiv.org/abs/2508.00448 https://arxiv.org/pdf/2508.0044…

Quantum Key-Recovery Attacks on FBC Algorithm
With the advancement of quantum computing, symmetric cryptography faces new challenges from quantum attacks. These attacks are typically classified into two models: Q1 (classical queries) and Q2 (quantum superposition queries). In this context, we present a comprehensive security analysis of the FBC algorithm considering quantum adversaries with different query capabilities. In the Q2 model, we first design 4-round polynomial-time quantum distinguishers for FBC-F and FBC-KF structures, and then…

EIM-TRNG: Obfuscating Deep Neural Network Weights with Encoding-in-Memory True Random Number Generator via RowHammer
Ranyang Zhou, Abeer Matar A. Almalky, Gamana Aragonda, Sabbir Ahmed, Filip Roth Tr{\o}nnes-Christensen, Adnan Siraj Rakin, Shaahin Angizi
https://arxiv.org/abs/2507.02206 …

EIM-TRNG: Obfuscating Deep Neural Network Weights with Encoding-in-Memory True Random Number Generator via RowHammer
True Random Number Generators (TRNGs) play a fundamental role in hardware security, cryptographic systems, and data protection. In the context of Deep NeuralNetworks (DNNs), safeguarding model parameters, particularly weights, is critical to ensure the integrity, privacy, and intel-lectual property of AI systems. While software-based pseudo-random number generators are widely used, they lack the unpredictability and resilience offered by hardware-based TRNGs. In this work, we propose a novel an…

Path Planning using a One-shot-sampling Skeleton Map
Gabriel O. Flores-Aquino, Octavio Gutierrez-Frias, Juan Irving Vasquez
https://arxiv.org/abs/2507.02328

Path Planning using a One-shot-sampling Skeleton Map
Path planning algorithms aim to compute a collision-free path, and many works focus on finding the optimal distance path. However, for some applications, a more suitable approach is to balance response time, safety of the paths, and path length. In this context, a skeleton map is a useful tool in graph-based schemes, as it provides an intrinsic representation of free configuration space. However, skeletonization algorithms are very resource-intensive, being primarily oriented towards image proc…

Finding My Way: Influence of Different Audio Augmented Reality Navigation Cues on User Experience and Subjective Usefulness
Sina Hinzmann, Francesco Vona, Juliane Henning, Mohamed Amer, Omar Abdellatif, Tanja Kojic, Jan-Niklas Voigt-Antons
https://arxiv.org/abs/2509.03199

Finding My Way: Influence of Different Audio Augmented Reality Navigation Cues on User Experience and Subjective Usefulness
As augmented reality (AR) becomes increasingly prevalent in mobile and context-aware applications, the role of auditory cues in guiding users through physical environments is becoming critical. This study investigates the effectiveness and user experience of various categories of audio cues, including fully non-verbal sounds and speech-derived Spearcons, during outdoor navigation tasks using the Meta Quest 3 headset. Twenty participants navigated five outdoor routes using audio-only cue types: …

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 …

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.…

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…

Prompt-Guided Patch UNet-VAE with Adversarial Supervision for Adrenal Gland Segmentation in Computed Tomography Medical Images
Hania Ghouse, Muzammil Behzad
https://arxiv.org/abs/2509.03188

Prompt-Guided Patch UNet-VAE with Adversarial Supervision for Adrenal Gland Segmentation in Computed Tomography Medical Images
Segmentation of small and irregularly shaped abdominal organs, such as the adrenal glands in CT imaging, remains a persistent challenge due to severe class imbalance, poor spatial context, and limited annotated data. In this work, we propose a unified framework that combines variational reconstruction, supervised segmentation, and adversarial patch-based feedback to address these limitations in a principled and scalable manner. Our architecture is built upon a VAE-UNet backbone that jointly rec…

Cosmic-Quantum Connections: Assessing the Viability of Weak Gravity and Weak Cosmic Censorship Conjectures in Kerr-Newman-Kiselev-Letelier Black Hole
Saeed Noori Gashti, Behnam Pourhassan
https://arxiv.org/abs/2508.01487

Cosmic-Quantum Connections: Assessing the Viability of Weak Gravity and Weak Cosmic Censorship Conjectures in Kerr-Newman-Kiselev-Letelier Black Hole
This paper addresses a potential validation of the weak gravity conjecture (WGC) with the weak cosmic censorship conjecture (WCCC), as a significant challenge in quantum gravity. We explore the viability of the WGC and WCCC in the context of the Kerr-Newman-Kiselev-Letelier (KNKL) black hole. Although these conjectures appear unrelated, but surprising connection between these conjectures, It establishes a bridge between the quantum and the cosmic. By imposing specific constraints on the black h…

Bug Fixing with Broader Context: Enhancing LLM-Based Program Repair via Layered Knowledge Injection
Ramtin Ehsani, Esteban Parra, Sonia Haiduc, Preetha Chatterjee
https://arxiv.org/abs/2506.24015

Bug Fixing with Broader Context: Enhancing LLM-Based Program Repair via Layered Knowledge Injection
Prompting LLMs with bug-related context (e.g., error messages, stack traces) improves automated program repair, but many bugs still remain unresolved. In real-world projects, developers often rely on broader repository and project-level context beyond the local code to resolve such bugs. In this paper, we investigate how automatically extracting and providing such knowledge can improve LLM-based program repair. We propose a layered knowledge injection framework that incrementally augments LLMs …

The Role of the Satellite in Quantum Information Networks
Luca Paccard, Valentin Leloup, Luca Lazzarini, Agathe Blaise, Mailys Guerault, Mickael Faugeron, Fabrice Arnal, Mathieu Bertrand, Raphael Aymeric, Michel Sotom, St\'ephanie Molin, Patrick G\'elard, Pierre Besancenot, Cyrille Laborde, Laurent de Forges de Parny, Mathias van den Bossche
https://

The Role of the Satellite in Quantum Information Networks
Quantum Information Networks (QIN) attract increasing interest, as they will enable interconnection of multiple quantum devices in a distributed organization thus enhancing intrinsic computing, sensing, and security capabilities. The core mechanism of a QIN is quantum state swapping, based on teleportation, which consumes quantum entanglement, and which can be seen in this context as a new kind of network resource. The satellite is expected to play a central role for supporting global connectiv…

SmartPoser: Arm Pose Estimation with a Smartphone and Smartwatch Using UWB and IMU Data
Nathan DeVrio, Vimal Mollyn, Chris Harrison
https://arxiv.org/abs/2509.03451 https://

SmartPoser: Arm Pose Estimation with a Smartphone and Smartwatch Using UWB and IMU Data
The ability to track a user's arm pose could be valuable in a wide range of applications, including fitness, rehabilitation, augmented reality input, life logging, and context-aware assistants. Unfortunately, this capability is not readily available to consumers. Systems either require cameras, which carry privacy issues, or utilize multiple worn IMUs or markers. In this work, we describe how an off-the-shelf smartphone and smartwatch can work together to accurately estimate arm pose. Moving be…

LeakSealer: A Semisupervised Defense for LLMs Against Prompt Injection and Leakage Attacks
Francesco Panebianco, Stefano Bonfanti, Francesco Trov\`o, Michele Carminati
https://arxiv.org/abs/2508.00602 …

LeakSealer: A Semisupervised Defense for LLMs Against Prompt Injection and Leakage Attacks
The generalization capabilities of Large Language Models (LLMs) have led to their widespread deployment across various applications. However, this increased adoption has introduced several security threats, notably in the forms of jailbreaking and data leakage attacks. Additionally, Retrieval Augmented Generation (RAG), while enhancing context-awareness in LLM responses, has inadvertently introduced vulnerabilities that can result in the leakage of sensitive information. Our contributions are t…

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 …

From Long Videos to Engaging Clips: A Human-Inspired Video Editing Framework with Multimodal Narrative Understanding
Xiangfeng Wang, Xiao Li, Yadong Wei, Xueyu Song, Yang Song, Xiaoqiang Xia, Fangrui Zeng, Zaiyi Chen, Liu Liu, Gu Xu, Tong Xu
https://arxiv.org/abs/2507.02790

From Long Videos to Engaging Clips: A Human-Inspired Video Editing Framework with Multimodal Narrative Understanding
The rapid growth of online video content, especially on short video platforms, has created a growing demand for efficient video editing techniques that can condense long-form videos into concise and engaging clips. Existing automatic editing methods predominantly rely on textual cues from ASR transcripts and end-to-end segment selection, often neglecting the rich visual context and leading to incoherent outputs. In this paper, we propose a human-inspired automatic video editing framework (HIVE)…

Dynamic Symbolic Execution for Semantic Difference Analysis of Component and Connector Architectures
Johanna Grahl, Bernhard Rumpe, Max Stachon, Sebastian St\"uber
https://arxiv.org/abs/2508.00749

Dynamic Symbolic Execution for Semantic Difference Analysis of Component and Connector Architectures
In the context of model-driven development, ensuring the correctness and consistency of evolving models is paramount. This paper investigates the application of Dynamic Symbolic Execution (DSE) for semantic difference analysis of component-and-connector architectures, specifically utilizing MontiArc models. We have enhanced the existing MontiArc-to-Java generator to gather both symbolic and concrete execution data at runtime, encompassing transition conditions, visited states, and internal vari…

Context-Driven Knowledge Graph Completion with Semantic-Aware Relational Message Passing
Siyuan Li, Ruitong Liu, Yan Wen, Te Sun
https://arxiv.org/abs/2506.23141

Context-Driven Knowledge Graph Completion with Semantic-Aware Relational Message Passing
Semantic context surrounding a triplet $(h, r, t)$ is crucial for Knowledge Graph Completion (KGC), providing vital cues for prediction. However, traditional node-based message passing mechanisms, when applied to knowledge graphs, often introduce noise and suffer from information dilution or over-smoothing by indiscriminately aggregating information from all neighboring edges. To address this challenge, we propose a semantic-aware relational message passing. A core innovation of this framework …

Pull Requests From The Classroom: Co-Developing Curriculum And Code
Dennis Zyska, Ilia Kuznetsov, Florian M\"uller, Iryna Gurevych
https://arxiv.org/abs/2508.00646 https://…

Pull Requests From The Classroom: Co-Developing Curriculum And Code
Educational technologies often misalign with instructors' pedagogical goals, forcing adaptations that compromise teaching efficacy. In this paper, we present a case study on the co-development of curriculum and technology in the context of a university course on scientific writing. Specifically, we examine how a custom-built peer feedback system was iteratively developed alongside the course to support annotation, feedback exchange, and revision. Results show that while co-development fostered …

KGBERT4Eth: A Feature-Complete Transformer Powered by Knowledge Graph for Multi-Task Ethereum Fraud Detection
Yifan Jia, Ye Tian, Liguo Zhang, Yanbin Wang, Jianguo Sun, Liangliang Song
https://arxiv.org/abs/2509.03860

KGBERT4Eth: A Feature-Complete Transformer Powered by Knowledge Graph for Multi-Task Ethereum Fraud Detection
Ethereum's rapid ecosystem expansion and transaction anonymity have triggered a surge in malicious activity. Detection mechanisms currently bifurcate into three technical strands: expert-defined features, graph embeddings, and sequential transaction patterns, collectively spanning the complete feature sets of Ethereum's native data layer. Yet the absence of cross-paradigm integration mechanisms forces practitioners to choose between sacrificing sequential context awareness, structured fund-flow…

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

Beyond Fixed: Variable-Length Denoising for Diffusion Large Language Models
Jinsong Li, Xiaoyi Dong, Yuhang Zang, Yuhang Cao, Jiaqi Wang, Dahua Lin
https://arxiv.org/abs/2508.00819

Beyond Fixed: Variable-Length Denoising for Diffusion Large Language Models
Diffusion Large Language Models (DLLMs) are emerging as a powerful alternative to the dominant Autoregressive Large Language Models, offering efficient parallel generation and capable global context modeling. However, the practical application of DLLMs is hindered by a critical architectural constraint: the need for a statically predefined generation length. This static length allocation leads to a problematic trade-off: insufficient lengths cripple performance on complex tasks, while excessive…

Thwart Me If You Can: An Empirical Analysis of Android Platform Armoring Against Stalkerware
Malvika Jadhav, Wenxuan Bao, Vincent Bindschaedler
https://arxiv.org/abs/2508.02454 …

Thwart Me If You Can: An Empirical Analysis of Android Platform Armoring Against Stalkerware
Stalkerware is a serious threat to individuals' privacy that is receiving increased attention from the security and privacy research communities. Existing works have largely focused on studying leading stalkerware apps, dual-purpose apps, monetization of stalkerware, or the experience of survivors. However, there remains a need to understand potential defenses beyond the detection-and-removal approach, which may not necessarily be effective in the context of stalkerware. In this paper, we per…

Highly Interactive Testing for Uninterrupted Development Flow
Andrew Tropin
https://arxiv.org/abs/2508.02176 https://arxiv.org/pdf/2508.02176

Highly Interactive Testing for Uninterrupted Development Flow
Highly interactive development environments (HIDEs) enable uninterrupted development flow through continuous program evolution and rapid hypothesis checking. However, traditional testing approaches -- typically executed separately via CLI -- isolate tests from HIDE tooling (interactive debuggers, value and stack inspectors, etc.) and introduce disruptive delays due to coarse execution granularity and lack of runtime context. This disconnect breaks development flow by exceeding critical attentio…

CooT: Learning to Coordinate In-Context with Coordination Transformers
Huai-Chih Wang, Hsiang-Chun Chuang, Hsi-Chun Cheng, Dai-Jie Wu, Shao-Hua Sun
https://arxiv.org/abs/2506.23549

CooT: Learning to Coordinate In-Context with Coordination Transformers
Effective coordination among artificial agents in dynamic and uncertain environments remains a significant challenge in multi-agent systems. Existing approaches, such as self-play and population-based methods, either generalize poorly to unseen partners or require extensive training. To overcome these limitations, we propose Coordination Transformers (CooT), a novel in-context coordination framework that uses recent interaction histories to adapt to unseen partners rapidly. Unlike previous appr…

NyayaRAG: Realistic Legal Judgment Prediction with RAG under the Indian Common Law System
Shubham Kumar Nigam, Balaramamahanthi Deepak Patnaik, Shivam Mishra, Ajay Varghese Thomas, Noel Shallum, Kripabandhu Ghosh, Arnab Bhattacharya
https://arxiv.org/abs/2508.00709

NyayaRAG: Realistic Legal Judgment Prediction with RAG under the Indian Common Law System
Legal Judgment Prediction (LJP) has emerged as a key area in AI for law, aiming to automate judicial outcome forecasting and enhance interpretability in legal reasoning. While previous approaches in the Indian context have relied on internal case content such as facts, issues, and reasoning, they often overlook a core element of common law systems, which is reliance on statutory provisions and judicial precedents. In this work, we propose NyayaRAG, a Retrieval-Augmented Generation (RAG) framewo…

CyberRAG: An agentic RAG cyber attack classification and reporting tool
Francesco Blefari, Cristian Cosentino, Francesco Aurelio Pironti, Angelo Furfaro, Fabrizio Marozzo
https://arxiv.org/abs/2507.02424

CyberRAG: An agentic RAG cyber attack classification and reporting tool
Intrusion Detection and Prevention Systems (IDS/IPS) in large enterprises can generate hundreds of thousands of alerts per hour, overwhelming security analysts with logs that demand deep, rapidly evolving domain expertise. Conventional machine-learning detectors trim the alert volume but still yield high false-positive rates, while standard single-pass Retrieval-Augmented Generation (RAG) pipelines often retrieve irrelevant context and fail to justify their predictions. To overcome these shortc…

Testing the Untestable? An Empirical Study on the Testing Process of LLM-Powered Software Systems
Cleyton Magalhaes, Italo Santos, Brody Stuart-Verner, Ronnie de Souza Santos
https://arxiv.org/abs/2508.00198

Testing the Untestable? An Empirical Study on the Testing Process of LLM-Powered Software Systems
Background: Software systems powered by large language models are becoming a routine part of everyday technologies, supporting applications across a wide range of domains. In software engineering, many studies have focused on how LLMs support tasks such as code generation, debugging, and documentation. However, there has been limited focus on how full systems that integrate LLMs are tested during development. Aims: This study explores how LLM-powered systems are tested in the context of real-wo…

MOTIF: Modular Thinking via Reinforcement Fine-tuning in LLMs
Purbesh Mitra, Sennur Ulukus
https://arxiv.org/abs/2507.02851 https://a…

MOTIF: Modular Thinking via Reinforcement Fine-tuning in LLMs
Recent advancements in the reasoning capabilities of large language models (LLMs) show that employing group relative policy optimization (GRPO) algorithm for reinforcement learning (RL) training allows the models to use more thinking/reasoning tokens for generating better responses. However, LLMs can generate only a finite amount of tokens while maintaining attention to the previously generated tokens. This limit, also known as the context size of an LLM, is a bottleneck in LLM reasoning with a…

In-context learning for the classification of manipulation techniques in phishing emails
Antony Dalmiere (LAAS-TRUST, LAAS), Guillaume Auriol (LAAS-TRUST, INSA Toulouse), Vincent Nicomette (LAAS-TSF, LAAS), Pascal Marchand (LERASS)
https://arxiv.org/abs/2506.22515

In-context learning for the classification of manipulation techniques in phishing emails
Traditional phishing detection often overlooks psychological manipulation. This study investigates using Large Language Model (LLM) In-Context Learning (ICL) for fine-grained classification of phishing emails based on a taxonomy of 40 manipulation techniques. Using few-shot examples with GPT-4o-mini on real-world French phishing emails (SignalSpam), we evaluated performance against a human-annotated test set (100 emails). The approach effectively identifies prevalent techniques (e.g., Baiting, …

MMBERT: Scaled Mixture-of-Experts Multimodal BERT for Robust Chinese Hate Speech Detection under Cloaking Perturbations
Qiyao Xue, Yuchen Dou, Ryan Shi, Xiang Lorraine Li, Wei Gao
https://arxiv.org/abs/2508.00760

MMBERT: Scaled Mixture-of-Experts Multimodal BERT for Robust Chinese Hate Speech Detection under Cloaking Perturbations
Hate speech detection on Chinese social networks presents distinct challenges, particularly due to the widespread use of cloaking techniques designed to evade conventional text-based detection systems. Although large language models (LLMs) have recently improved hate speech detection capabilities, the majority of existing work has concentrated on English datasets, with limited attention given to multimodal strategies in the Chinese context. In this study, we propose MMBERT, a novel BERT-based m…

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. …

Leveraging In-Context Learning for Political Bias Testing of LLMs
Patrick Haller, Jannis Vamvas, Rico Sennrich, Lena A. J\"ager
https://arxiv.org/abs/2506.22232

Leveraging In-Context Learning for Political Bias Testing of LLMs
A growing body of work has been querying LLMs with political questions to evaluate their potential biases. However, this probing method has limited stability, making comparisons between models unreliable. In this paper, we argue that LLMs need more context. We propose a new probing task, Questionnaire Modeling (QM), that uses human survey data as in-context examples. We show that QM improves the stability of question-based bias evaluation, and demonstrate that it may be used to compare instruct…

NeedleChain: Measuring Intact Long-Context Reasoning Capability of Large Language Models
Hyeonseok Moon, Heuiseok Lim
https://arxiv.org/abs/2507.22411 https://

NeedleChain: Measuring Intact Long-Context Reasoning Capability of Large Language Models
The Needle-in-a-Haystack (NIAH) benchmark is widely used to evaluate Large Language Models' (LLMs) ability to understand long contexts (LC). It evaluates the capability to identify query-relevant context within extensive query-irrelevant passages. Although this method serves as a widely accepted standard for evaluating long-context understanding, our findings suggest it may overestimate the true LC capability of LLMs. We demonstrate that even state-of-the-art models such as GPT-4o struggle to i…