Tootfinder

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…

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.

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.

the context, the context, the context

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…

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…

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…

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…

A Measurement Study of Model Context Protocol
Hechuan Guo, Yongle Hao, Yue Zhang, Minghui Xu, Peizhuo Lyu, Jiezhi Chen, Xiuzhen Cheng
https://arxiv.org/abs/2509.25292 https://…

A Measurement Study of Model Context Protocol
The Model Context Protocol (MCP) has been proposed as a unifying standard for connecting large language models (LLMs) with external tools and resources, promising the same role for AI integration that HTTP and USB played for the Web and peripherals. Yet, despite rapid adoption and hype, its trajectory remains uncertain. Are MCP marketplaces truly growing, or merely inflated by placeholders and abandoned prototypes? Are servers secure and privacy-preserving, or do they expose users to systemic r…

Memory-Augmented Log Analysis with Phi-4-mini: Enhancing Threat Detection in Structured Security Logs
Anbi Guo, Mahfuza Farooque
https://arxiv.org/abs/2510.00529 https://…

Memory-Augmented Log Analysis with Phi-4-mini: Enhancing Threat Detection in Structured Security Logs
Structured security logs are critical for detecting advanced persistent threats (APTs). Large language models (LLMs) struggle in this domain due to limited context and domain mismatch. We propose \textbf{DM-RAG}, a dual-memory retrieval-augmented generation framework for structured log analysis. It integrates a short-term memory buffer for recent summaries and a long-term FAISS-indexed memory for historical patterns. An instruction-tuned Phi-4-mini processes the combined context and outputs str…

Leveraging Scene Context with Dual Networks for Sequential User Behavior Modeling
Xu Chen, Yunmeng Shu, Yuangang Pan, Jinsong Lan, Xiaoyong Zhu, Shuai Xiao, Haojin Zhu, Ivor W. Tsang, Bo Zheng
https://arxiv.org/abs/2509.26172

Leveraging Scene Context with Dual Networks for Sequential User Behavior Modeling
Modeling sequential user behaviors for future behavior prediction is crucial in improving user's information retrieval experience. Recent studies highlight the importance of incorporating contextual information to enhance prediction performance. One crucial but usually neglected contextual information is the scene feature which we define as sub-interfaces within an app, created by developers to provide specific functionalities, such as ``text2product search" and ``live" modules in e-commence ap…

SparseServe: Unlocking Parallelism for Dynamic Sparse Attention in Long-Context LLM Serving
Qihui Zhou, Peiqi Yin, Pengfei Zuo, James Cheng
https://arxiv.org/abs/2509.24626 http…

SparseServe: Unlocking Parallelism for Dynamic Sparse Attention in Long-Context LLM Serving
Serving long-context LLMs is costly because attention computation grows linearly with context length. Dynamic sparse attention algorithms (DSAs) mitigate this by attending only to the key-value (KV) cache of critical tokens. However, with DSAs, the main performance bottleneck shifts from HBM bandwidth to HBM capacity: KV caches for unselected tokens must remain in HBM for low-latency decoding, constraining parallel batch size and stalling further throughput gains. Offloading these underutilized…

#FotoVorschlag 'Fortbewegungsmittel' 'Means of transportation'
1/.
I do not have that many photos of my beloved means of transportation.
The photo itself has also some context which I've written in my blog:

Does anybody know of aggregations of dcat data (or any published dcat on data collections/services) in the context of cultural heritage institutions?
#openglam #dcat #culturalheritage

If you think there might be an AI bubble, but don't worry, even if it pops, how bad can it be?
Here's one number: seven tech companies (NVIDIA, Microsoft, Google, Apple, Meta, Tesla, Amazon) are worth ~20 trillion, ⇓ of total US stock market
For context: total subprime mortgages in 2008 were 1.3 trillion, total US mortgage debt 10 trillion
You have a 401k or index fonds? Your money is likely tied with the AI market. If you want that, that's fine. If you're …

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…

My sabbatical officially starts today!
https://mastodon.social/@nocontexttrek/115073662238539617

Star Trek Minus Context (@nocontexttrek@mastodon.social)
Attached: 1 image #StarTrek

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. …

Also a big thanks to the unknown-to-me person who already made the first PRs to add more configs over the weekend: https://codeberg.org/gedankenstuecke/freshrss-fulltext-settings
I've also expanded the README a bit to hopefully provide more context 🙂

The Complexity Trap: Simple Observation Masking Is as Efficient as LLM Summarization for Agent Context Management
Tobias Lindenbauer, Igor Slinko, Ludwig Felder, Egor Bogomolov, Yaroslav Zharov
https://arxiv.org/abs/2508.21433

The Complexity Trap: Simple Observation Masking Is as Efficient as LLM Summarization for Agent Context Management
Large Language Model (LLM)-based agents solve complex tasks through iterative reasoning, exploration, and tool-use, a process that can result in long, expensive context histories. While state-of-the-art Software Engineering ( SE) agents like OpenHands or Cursor use LLM-based summarization to tackle this issue, it is unclear whether the increased complexity offers tangible performance benefits compared to simply omitting older observations. We present a systematic comparison of these strategies …

Given the remarkable #GreensSurge in members and polling, and the ongoing chaotic infighting at 'Your Party', it's maybe worth reposting my questioning of both leaderships on the eco-planetary crisis and overshoot.
As UK politics turns both right and left, how do we get degrowth onto the agenda? – degrowthUK

As UK politics turns both right and left, how do we get degrowth onto the agenda?
By Mark H Burton In the series, Prospects for Degrowth The last few weeks have seen two opposing developments in British politics, both in the context of the Starmer Labour government and its failu…

Trying to figure out why Gemini AI in Gmail is so very bad. It seems to not have access to all of my email and whatever RAG-like thing they are doing is very stupid.
Here's Gemini saying my oldest email matching some keywords in 2021 is from March 2021. If I explicitly tell it to look in February 2021, it finds it. There's only 11 emails that match in all of 2021, it's not like I'm blowing up the context window.

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

Debating degrowth: A response to Jason Hickel.
https://www.resilience.org/stories/2025-10-02/debating-degrowth-a-response-to-jason-hickel/
We (coordinators of the DegrowthUK website) will possibly write something about this. Broadly…

Debating degrowth: A response to Jason Hickel
We really believe that degrowth is a more complex, substantial and significant concept and movement than its treatment in the Hickel interview. We think that the degrowth movement has a very important role to play in the challenging political context.

Why bother? He'll just want to rip them up so that he can add fake gold in their place.
https://www.theguardian.com/commentisfree/2025/sep/01/soldiers-landscaping-washington-dc-crime

Permutation closure for multiple context-free languages
Andrew Duncan, Murray Elder, Lisa Frenkel, Mengfan Lyu
https://arxiv.org/abs/2509.22239 https://arx…

Permutation closure for multiple context-free languages
We prove that the \emph{permutation closure} of a multiple context-free language is multiple context-free, which extends work of Okhotin and Sorokin [LATA 2020] who showed closure under \emph{cyclic shift}, and complements work of Brandstädt [1981, RAIRO Inform. Théor.] (resp. Brough \emph{et al.} [2016, Discrete Math. Theor. Comput. Sci.]) who showed the same result for regular, context-sensitive, recursively enumerable (resp. EDT0L and ET0L) languages. In contrast to Okhotin and Sorokin who…

Honestly, #emoji and icons in #Unicode are a true horror.
Yeah, sure. It's great that you don't have to use <img/> anymore and you can just paste a random Unicode character. You can get graphics into fields where only text was originally intended (like bug summaries). Even better, you can now easily get cool colorful icons on terminal with almost no effort.
However, it is an #accessibility nightmare. People are now encoding *information* in random graphical symbols. Symbols that require huge fonts to render, or huge character tables to describe.
Yeah, a bare <img/> carrying information sucks. However, you can add a *meaningful* alt-text to the image, and accessibility tools can use that text to provide meaningful context. Like "bug fix".
However, emojis and icons are symbolic. The best you can get is some description like "hammer and wrench", so people can kinda figure out that it's probably a "bug fix". Or maybe it was a "maintenance task"? Or you'll get a "unknown character 0x1F6E0". And I'm sure people will surely enjoy cross-referencing a "legend" of such "unknown characters".

Addiction (Speculatve)
Kind of a fucked-up metaphor, but I was thinking yesterday that parenting is a lot like addiction. If you separate me from my child, I'll take completely irrational and desperate actions to get them back, driven by a deep instinct that goes well beyond "love." I'll also make self-disadvantageous long-term decisions like forgoing sleep, working an extra job, or quitting a job to do some combination of providing for and/or being present with my child.
Even in parenting situations where love is absent, and beyond, I think, the possessiveness that sometimes festers in those situations, there's often (although not always) a craving for simple presence of the child.
In a healthy relationship, there's a whole lot more than this, but it's interesting to me that the same obsessive craving and absolute priority that we think of as diseased and/or monstrous in someone addicted to a hard drug can be healthy in the right context (that is, when it doesn't contribute to abusive or twisted parental relationships but instead exists alongside a healthy amount of love and respect).
Makes me wonder if there are ways to have a truly healthy drug addiction, although I recognize the answer might well be "no" and that even if it's "technically/theoretically yes" it might still be harmful to hype up or even merely discuss that possibility since it might help addicted people in harmful addictions more easily justify inaction. At minimum I think any "yes" answer here involves assuming utopian-level differences from our current society.
#Parenting #Addiction

📊 Full observability: Log every state change, measure latency, generate audit trails
🎨 #PHP8 features: String-backed enums for states/events, readonly value objects for context data
🔄 Domain events pattern: Emit events to outbox, handle side effects idempotently with retries
📋 Easy testing: Unit test state transitions without hitting database or network
📈 Auto-documentation: …

SlimPack: Fine-Grained Asymmetric Packing for Balanced and Efficient Variable-Length LLM Training
Yuliang Liu, Guohao Wu, Shenglong Zhang, Wei Zhang, Qianchao Zhu, Zhouyang Li, Chenyu Wang
https://arxiv.org/abs/2509.26246

SlimPack: Fine-Grained Asymmetric Packing for Balanced and Efficient Variable-Length LLM Training
The efficient distributed training of Large Language Models (LLMs) is severely hampered by the extreme variance in context lengths. This data heterogeneity, amplified by conventional packing strategies and asymmetric forward-backward costs, leads to critical inefficiencies such as cascading workload imbalances and severe hardware underutilization. Existing solutions attempt to mitigate these challenges, but often at the expense of memory or communication efficiency. To address these challenge…

DynaMIC: Dynamic Multimodal In-Context Learning Enabled Embodied Robot Counterfactual Resistance Ability
Tianqiang Yan, Ziqiao Lin, Sicheng Wang, Tianwei Zhang, Zhenglong Sun
https://arxiv.org/abs/2509.24413

DynaMIC: Dynamic Multimodal In-Context Learning Enabled Embodied Robot Counterfactual Resistance Ability
The emergence of large pre-trained models based on natural language has breathed new life into robotics development. Extensive research has integrated large models with robots, utilizing the powerful semantic understanding and generation capabilities of large models to facilitate robot control through natural language instructions gradually. However, we found that robots that strictly adhere to human instructions, especially those containing misleading information, may encounter errors during t…

POVQA: Preference-Optimized Video Question Answering with Rationales for Data Efficiency
Ashim Dahal, Ankit Ghimire, Saydul Akbar Murad, Nick Rahimi
https://arxiv.org/abs/2510.01009

POVQA: Preference-Optimized Video Question Answering with Rationales for Data Efficiency
Video Question Answering (VQA) with Large Vision Language Models (LVLMs) has gained significant traction in research ever since the Flamingo was introduced by Deepmind. Recent advancements in large context/long video question answering have allowed VQA tasks to have context window of 1500+ frames. However, this only leads to 50 seconds of video footage without losing any significant information. We introduce POVQA, a data-efficient pipeline that compresses each second of video into a single tem…

Rethinking Thinking Tokens: LLMs as Improvement Operators
Lovish Madaan, Aniket Didolkar, Suchin Gururangan, John Quan, Ruan Silva, Ruslan Salakhutdinov, Manzil Zaheer, Sanjeev Arora, Anirudh Goyal
https://arxiv.org/abs/2510.01123

Rethinking Thinking Tokens: LLMs as Improvement Operators
Reasoning training incentivizes LLMs to produce long chains of thought (long CoT), which among other things, allows them to explore solution strategies with self-checking. This results in higher accuracy, but inflates context length, token/compute cost, and answer latency. We ask: Can current models leverage their metacognition to provide other combinations on this Pareto frontier, e.g., better accuracy with lower context length and/or latency? Abstractly, we view the model as an improvement op…

Dynamics of Majorana zero modes across hybrid Kitaev chain
Rajiv Kumar, Rohit Kumar Shukla, Levan Chotorlishvili, Sunil Kumar Mishra
https://arxiv.org/abs/2509.26134 https://

Dynamics of Majorana zero modes across hybrid Kitaev chain
The Kitaev chain has been extensively explored in the context of uniform couplings, with studies focusing either on purely nearest-neighbor interactions or on systems dominated by long-range superconducting pairing. Building on these investigations, we introduce a hybrid Kitaev chain in which the lattice is partitioned into two segments: the left segment comprises nearest-neighbor couplings, while the right segment incorporates long-range pairing. To probe the role of the interface, we study tw…

Test time training enhances in-context learning of nonlinear functions
Kento Kuwataka, Taiji Suzuki
https://arxiv.org/abs/2509.25741 https://arxiv.org/pdf/…

Test time training enhances in-context learning of nonlinear functions
Test-time training (TTT) enhances model performance by explicitly updating designated parameters prior to each prediction to adapt to the test data. While TTT has demonstrated considerable empirical success, its theoretical underpinnings remain limited, particularly for nonlinear models. In this paper, we investigate the combination of TTT with in-context learning (ICL), where the model is given a few examples from the target distribution at inference time. We analyze this framework in the sett…

Analysis of Semantic Communication for Logic-based Hypothesis Deduction
Ahmet Faruk Saz, Siheng Xiong, Faramarz Fekri
https://arxiv.org/abs/2508.21755 https://

Analysis of Semantic Communication for Logic-based Hypothesis Deduction
This work presents an analysis of semantic communication in the context of First-Order Logic (FOL)-based deduction. Specifically, the receiver holds a set of hypotheses about the State of the World (SotW), while the transmitter has incomplete evidence about the true SotW but lacks access to the ground truth. The transmitter aims to communicate limited information to help the receiver identify the hypothesis most consistent with true SotW. We formulate the objective as approximating the posterio…

Designing Wine Tasting Experiences for All: The role of Human Diversity and Personal food memory
Xinyang Shan, Yuanyuan Xu, Yuqing Wang, Tian Xia, Yinshan Lin
https://arxiv.org/abs/2510.00607

Designing Wine Tasting Experiences for All: The role of Human Diversity and Personal food memory
This study investigates the design of inclusive wine-tasting experiences by examining the roles of human diversity and personal food memory. Through field studies conducted in various wine regions, we explored how Chinese visitors engage with wine-tasting activities during winery tours, highlighting the cross-cultural challenges they face. Our findings underscore the importance of experiencers' abilities, necessities, and aspirations (ANAs), the authenticity of wine tasting within the context o…

Throttling for metric dimension and its variants
Boris Brimkov, Peter Diao, Jesse Geneson, Carolyn Reinhart, Shen-Fu Tsai, William Wang, Kyle Worley
https://arxiv.org/abs/2510.00530

Throttling for metric dimension and its variants
Metric dimension is a graph parameter that has been applied to robot navigation and finding low-dimensional vector embeddings. Throttling entails minimizing the sum of two available resources when solving certain graph problems. In this paper, we introduce throttling for metric dimension, edge metric dimension, and mixed metric dimension. In the context of vector embeddings, metric dimension throttling finds a low-dimensional, low-magnitude embedding with integer coordinates. We show that compu…

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

Test particle sampling and particle acceleration in a 2D coronal plasmoid-mediated reconnecting current sheet
Eilif S. {\O}yre, Boris V. Gudiksen, Lyndsay Fletcher
https://arxiv.org/abs/2509.25447

Test particle sampling and particle acceleration in a 2D coronal plasmoid-mediated reconnecting current sheet
Context: Solar flares accelerate electrons, creating non-thermal energy distributions. However, the acceleration sites and dominant acceleration mechanisms remain largely unknown. Aims: We study the characteristics of electron acceleration and subsequent non-thermal energy distribution in a 2D coronal plasmoid-mediated reconnecting current sheet. Methods: We used test particles and the guiding centre approximation to transport electrons in a static coronal 2D fan-spine topology magnetohydro…

Does anybody know of aggregations of dcat data (or any published dcat on data collections/services) in the context of cultural heritage institutions?
#openglam #dcat #culturalheritage

Data Quality Taxonomy for Data Monetization
Eduardo Vyhmeister, Bastien Pietropoli, Andrea Visentin
https://arxiv.org/abs/2510.00089 https://arxiv.org/pdf/…

Data Quality Taxonomy for Data Monetization
This chapter presents a comprehensive taxonomy for assessing data quality in the context of data monetisation, developed through a systematic literature review. Organising over one hundred metrics and Key Performance Indicators (KPIs) into four subclusters (Fundamental, Contextual, Resolution, and Specialised) within the Balanced Scorecard (BSC) framework, the taxonomy integrates both universal and domain-specific quality dimensions. By positioning data quality as a strategic connector across t…

Asynchronous Nonlinear Sheaf Diffusion for Multi-Agent Coordination
Yichen Zhao, Tyler Hanks, Hans Riess, Samuel Cohen, Matthew Hale, James Fairbanks
https://arxiv.org/abs/2510.00270

Asynchronous Nonlinear Sheaf Diffusion for Multi-Agent Coordination
Cellular sheaves and sheaf Laplacians provide a far-reaching generalization of graphs and graph Laplacians, resulting in a wide array of applications ranging from machine learning to multi-agent control. In the context of multi-agent systems, so called coordination sheaves provide a unifying formalism that models heterogeneous agents and coordination goals over undirected communication topologies, and applying sheaf diffusion drives agents to achieve their coordination goals. Existing literatur…

Whenever I see a clip of someone throwing around the Video Game Line after a tragedy I gotta check the upload date to see if it was just an out of context clip from 2004 or something new.

Energy conditions and gravitational baryogenesis in $f(R, {\cal R})$ gravity
K. Atazadeh, S. Golsanamlou
https://arxiv.org/abs/2510.01148 https://arxiv.org…

Energy conditions and gravitational baryogenesis in $f(R, {\cal R})$ gravity
In this work, first we examine the energy conditions in the context of the generalized metric-Palatini hybrid gravity, known as $f(R, {\cal R})$ gravity. We show that for the proposed model in this study, {\it i.e.} $ f(R, {\cal R})= R +α{\cal R}^{n} $, one of the four fundamental energy conditions, specifically the strong energy condition, does not hold for some values of $n$. Therefore, it seems that hybrid gravity can provide a model for the accelerated expansion of the universe. In continu…

In Apple Music context does adding a song to a playlist imply adding the album to your library?
As an aside, as a subscribed Apple Music customer is there some support that I can trach out to with similar questions and issues?

Enhancing Connectivity for Emergency Vehicles Through UAV Trajectory and Resource Allocation Optimization
S. Fatemeh Bozorgi, S. Mohammad Razavizadeh, Mohsen Rezaee
https://arxiv.org/abs/2509.26067

Enhancing Connectivity for Emergency Vehicles Through UAV Trajectory and Resource Allocation Optimization
Effective communication for emergency vehicles - such as ambulances and fire trucks - is essential to support their operations in various traffic and environmental conditions. In this context, this paper investigates a vehicular communication system assisted by an Unmanned Aerial Vehicle (UAV), which adjusts its trajectory and resource allocation according to communication needs. The system classifies vehicles into two groups to address their varying service requirements: emergency vehicles, wh…

Context-Driven Performance Modeling for Causal Inference Operators on Neural Processing Units
Neelesh Gupta, Rakshith Jayanth, Dhruv Parikh, Viktor Prasanna
https://arxiv.org/abs/2509.25155

Context-Driven Performance Modeling for Causal Inference Operators on Neural Processing Units
The proliferation of large language models (LLMs) has driven demand for long context inference on resource constrained edge devices. However, deploying these models on Neural Processing Units (NPUs) presents significant challenges due to the architectural mismatch: quadratic complexity of standard attention mechanisms conflicts with memory and compute patterns of edge accelerators. This paper presents a comprehensive performance analysis of various causal inference operators on a modern NPU. We…

DeepCodeSeek: Real-Time API Retrieval for Context-Aware Code Generation
Esakkivel Esakkiraja, Denis Akhiyarov, Aditya Shanmugham, Chitra Ganapathy
https://arxiv.org/abs/2509.25716

DeepCodeSeek: Real-Time API Retrieval for Context-Aware Code Generation
Current search techniques are limited to standard RAG query-document applications. In this paper, we propose a novel technique to expand the code and index for predicting the required APIs, directly enabling high-quality, end-to-end code generation for auto-completion and agentic AI applications. We address the problem of API leaks in current code-to-code benchmark datasets by introducing a new dataset built from real-world ServiceNow Script Includes that capture the challenge of unclear API us…

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 …

Fine-Grained Detection of Context-Grounded Hallucinations Using LLMs
Yehonatan Pesiakhovsky, Zorik Gekhman, Yosi Mass, Liat Ein-Dor, Roi Reichart
https://arxiv.org/abs/2509.22582

Fine-Grained Detection of Context-Grounded Hallucinations Using LLMs
Context-grounded hallucinations are cases where model outputs contain information not verifiable against the source text. We study the applicability of LLMs for localizing such hallucinations, as a more practical alternative to existing complex evaluation pipelines. In the absence of established benchmarks for meta-evaluation of hallucinations localization, we construct one tailored to LLMs, involving a challenging human annotation of over 1,000 examples. We complement the benchmark with an LLM…

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…

Crosslisted article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[5/10]:
- In-Context Curiosity: Distilling Exploration for Decision-Pretrained Transformers on Bandit Tasks
Huitao Yang, Guanting Chen

Learnable Conformal Prediction with Context-Aware Nonconformity Functions for Robotic Planning and Perception
Divake Kumar, Sina Tayebati, Francesco Migliarba, Ranganath Krishnan, Amit Ranjan Trivedi
https://arxiv.org/abs/2509.21955

Learnable Conformal Prediction with Context-Aware Nonconformity Functions for Robotic Planning and Perception
Deep learning models in robotics often output point estimates with poorly calibrated confidences, offering no native mechanism to quantify predictive reliability under novel, noisy, or out-of-distribution inputs. Conformal prediction (CP) addresses this gap by providing distribution-free coverage guarantees, yet its reliance on fixed nonconformity scores ignores context and can yield intervals that are overly conservative or unsafe. We address this with Learnable Conformal Prediction (LCP), whi…

Computable measures of non-Markovianity for Gaussian free fermion systems
Giuliano Chiriac\`o
https://arxiv.org/abs/2509.25953 https://arxiv.org/pdf/2509.2…

Computable measures of non-Markovianity for Gaussian free fermion systems
We investigate measures of non-Markovianity in open quantum systems governed by Gaussian free fermionic dynamics. Standard indicators of non-Markovian behavior, such as the BLP and LFS measures, are revisited in this context. We show that for Gaussian states, trace-based distances -- specifically the Hilbert-Schmidt norm -- and second-order Rényi mutual information can be efficiently expressed in terms of two-point correlation functions, enabling practical computation even in systems where the…

Explainable and Resilient ML-Based Physical-Layer Attack Detectors
Aleksandra Knapi\'nska, Marija Furdek
https://arxiv.org/abs/2509.26530 https://arxiv…

Explainable and Resilient ML-Based Physical-Layer Attack Detectors
Detection of emerging attacks on network infrastructure is a critical aspect of security management. To meet the growing scale and complexity of modern threats, machine learning (ML) techniques offer valuable tools for automating the detection of malicious activities. However, as these techniques become more complex, their internal operations grow increasingly opaque. In this context, we address the need for explainable physical-layer attack detection methods. First, we analyze the inner workin…

Next Point-of-interest (POI) Recommendation Model Based on Multi-modal Spatio-temporal Context Feature Embedding
Lingyu Zhang, Guobin Wu, Yan Wang, Pengfei Xu, Jian Liang, Xuan Song, Yunhai Wang
https://arxiv.org/abs/2509.22661

Next Point-of-interest (POI) Recommendation Model Based on Multi-modal Spatio-temporal Context Feature Embedding
The next Point-of-interest (POI) recommendation is mainly based on sequential traffic information to predict the user's next boarding point location. This is a highly regarded and widely applied research task in the field of intelligent transportation, and there have been many research results to date. Traditional POI prediction models primarily rely on short-term traffic sequence information, often neglecting both long-term and short-term preference data, as well as crucial spatiotemporal cont…

Carbon and nitrogen as indicators of stellar evolution and age. A homogeneous sample of 44 open clusters from the Gaia-ESO Survey
G. Tautvai\v{s}ien\.e, A. Drazdauskas, \v{S}. Mikolaitis, R. Minkevi\v{c}i\=ut\.e, E. Stonkut\.e, S. Randich, A. Bragaglia, L. Magrini, R. Smiljanic, M. Ambrosch, V. Bagdonas, G. Casali, Y. Chorniy, C. Viscasillas V\'azquez
https…

Carbon and nitrogen as indicators of stellar evolution and age. A homogeneous sample of 44 open clusters from the Gaia-ESO Survey
Context. Low- and intermediate-mass giants undergo a complex chemical evolution that has yet to be observationally probed. The influence of core helium flash on the chemical composition of stellar atmospheres has been an open question since its theoretical prediction 60 years ago. Aims. Based on high-resolution spectral observations of 44 open star clusters in the Gaia-ESO survey, our aim is to perform the first large-scale homogeneous investigation into the carbon and nitrogen photospheric con…

U-DFA: A Unified DINOv2-Unet with Dual Fusion Attention for Multi-Dataset Medical Segmentation
Zulkaif Sajjad, Furqan Shaukat, Junaid Mir
https://arxiv.org/abs/2510.00585 https:…

U-DFA: A Unified DINOv2-Unet with Dual Fusion Attention for Multi-Dataset Medical Segmentation
Accurate medical image segmentation plays a crucial role in overall diagnosis and is one of the most essential tasks in the diagnostic pipeline. CNN-based models, despite their extensive use, suffer from a local receptive field and fail to capture the global context. A common approach that combines CNNs with transformers attempts to bridge this gap but fails to effectively fuse the local and global features. With the recent emergence of VLMs and foundation models, they have been adapted for dow…

Orchid: Orchestrating Context Across Creative Workflows with Generative AI
Srishti Palani, Gonzalo Ramos
https://arxiv.org/abs/2508.19517 https://arxiv.org…

Orchid: Orchestrating Context Across Creative Workflows with Generative AI
Context is critical for meaningful interactions between people and Generative AI (GenAI). Yet mainstream tools offer limited means to orchestrate it, particularly across workflows that span multiple interactions, sessions, and models, as often occurs in creative projects. Re specifying prior details, juggling diverse artifacts, and dealing with context drift overwhelm users, obscure intent, and curtail creativity. To address these challenges, we present Orchid, a system that gives its users aff…

Mapping Toxic Comments Across Demographics: A Dataset from German Public Broadcasting
Jan Fillies, Michael Peter Hoffmann, Rebecca Reichel, Roman Salzwedel, Sven Bodemer, Adrian Paschke
https://arxiv.org/abs/2508.21084

Mapping Toxic Comments Across Demographics: A Dataset from German Public Broadcasting
A lack of demographic context in existing toxic speech datasets limits our understanding of how different age groups communicate online. In collaboration with funk, a German public service content network, this research introduces the first large-scale German dataset annotated for toxicity and enriched with platform-provided age estimates. The dataset includes 3,024 human-annotated and 30,024 LLM-annotated anonymized comments from Instagram, TikTok, and YouTube. To ensure relevance, comments we…

Novel very-high-frequency quasi-periodic oscillations of compact, non-singular objects
Jens Boos, Felix Wunsch
https://arxiv.org/abs/2510.00986 https://arx…

Novel very-high-frequency quasi-periodic oscillations of compact, non-singular objects
We report on a novel set of very-high-frequency quasi-periodic oscillations (VHFQPO's) in the context of compact, non-singular horizonless objects. Focussing on the static, spherically symmetric case we utilize metrics of non-singular black holes that are accompanied by a regulator length scale $L > 0$. The choice $L \gtrsim GM$ generically removes the horizon from these metrics leading to compact, horizonless but non-singular objects. This generically guarantees the existence of a stable orbit…

from my link log —
libfringe: an old Rust library for stackful coroutines.
https://github.com/edef1c/libfringe
saved 2025-10-28 https://dotat.at/:/S…

'Too much alignment; not enough culture': Re-balancing cultural alignment practices in LLMs
Eric J. W. Orlowski, Hakim Norhashim, Tristan Koh Ly Wey
https://arxiv.org/abs/2509.26167

'Too much alignment; not enough culture': Re-balancing cultural alignment practices in LLMs
While cultural alignment has increasingly become a focal point within AI research, current approaches relying predominantly on quantitative benchmarks and simplistic proxies fail to capture the deeply nuanced and context-dependent nature of human cultures. Existing alignment practices typically reduce culture to static demographic categories or superficial cultural facts, thereby sidestepping critical questions about what it truly means to be culturally aligned. This paper argues for a fundamen…

Context-Specific Instruction: A Longitudinal Study on Debugging Skill Acquisition and Retention for Novice Programmers
Ziyi Zhang, Devjeet Roy, Venera Arnaoudova
https://arxiv.org/abs/2509.22420

Context-Specific Instruction: A Longitudinal Study on Debugging Skill Acquisition and Retention for Novice Programmers
Bug localization is a critical skill, yet novices often lack systematic approaches. Prior work tested abstract guidelines and general concrete steps; the impact of context-specific instruction is unclear. We ran an eight-week longitudinal study with four conditions: no instruction (G1), abstract guidelines (G2), concrete steps (G3), and our context-specific instruction that pairs concrete bug-localization steps with problem-specific details (G4). Forty-four undergraduates participated; 41 compl…

TASP: Topology-aware Sequence Parallelism
Yida Wang (Capital Normal University, Infinigence-AI), Ke Hong (Tsinghua University, Infinigence-AI), Xiuhong Li (Infinigence-AI), Yuanchao Xu (Capital Normal University), Wenxun Wang (Tsinghua University), Guohao Dai (Infinigence-AI, Shanghai Jiao Tong University), Yu Wang (Tsinghua University)
https://

TASP: Topology-aware Sequence Parallelism
Long-context large language models (LLMs) face constraints due to the quadratic complexity of the self-attention mechanism. The mainstream sequence parallelism (SP) method, Ring Attention, attempts to solve this by distributing the query into multiple query chunks across accelerators and enable each Q tensor to access all KV tensors from other accelerators via the Ring AllGather communication primitive. However, it exhibits low communication efficiency, restricting its practical applicability. …

Strata: Hierarchical Context Caching for Long Context Language Model Serving
Zhiqiang Xie, Ziyi Xu, Mark Zhao, Yuwei An, Vikram Sharma Mailthody, Scott Mahlke, Michael Garland, Christos Kozyrakis
https://arxiv.org/abs/2508.18572

Strata: Hierarchical Context Caching for Long Context Language Model Serving
Large Language Models (LLMs) with expanding context windows face significant performance hurdles. While caching key-value (KV) states is critical for avoiding redundant computation, the storage footprint of long-context caches quickly exceeds GPU memory capacity, forcing production systems to adopt hierarchical caching across memory hierarchies. However, transferring large cached contexts back to the GPU introduces severe performance bottlenecks: fragmented I/O from paged layouts prevents full …

A Contextual Seven-Valued Logic (\emph{Saptabhang\=inaya}) for Quantum Systems
Partha Ghose
https://arxiv.org/abs/2510.01120 https://arxiv.org/pdf/2510.011…

A Contextual Seven-Valued Logic (\emph{Saptabhangīnaya}) for Quantum Systems
The quantum measurement problem is often presented as a conflict between unitary evolution and non-unitary collapse. Drawing on Wittgenstein's later philosophy of language and Bohr's principle of complementarity, we argue that this conflict is a grammatical illusion arising from cross-context conflations. To address this, we introduce a contextual seven-valued logic modeled on the Jaina doctrine of \emph{saptabhangīnaya} (sevenfold predication). In one formulation, each proposition is assigned…

Uncertainty-Aware Concept Bottleneck Models with Enhanced Interpretability
Haifei Zhang, Patrick Barry, Eduardo Brandao
https://arxiv.org/abs/2510.00773 https://

Uncertainty-Aware Concept Bottleneck Models with Enhanced Interpretability
In the context of image classification, Concept Bottleneck Models (CBMs) first embed images into a set of human-understandable concepts, followed by an intrinsically interpretable classifier that predicts labels based on these intermediate representations. While CBMs offer a semantically meaningful and interpretable classification pipeline, they often sacrifice predictive performance compared to end-to-end convolutional neural networks. Moreover, the propagation of uncertainty from concept pred…

Finding Phones Fast: Low-Latency and Scalable Monitoring of Cellular Communications in Sensitive Areas
Martin Kotuliak, Simon Erni, Jakub Pol\'ak, Marc Roeschlin, Richard Baker, Ivan Martinovic, Srdjan \v{C}apkun
https://arxiv.org/abs/2509.25430

Finding Phones Fast: Low-Latency and Scalable Monitoring of Cellular Communications in Sensitive Areas
The widespread availability of cellular devices introduces new threat vectors that allow users or attackers to bypass security policies and physical barriers and bring unauthorized devices into sensitive areas. These threats can arise from user non-compliance or deliberate actions aimed at data exfiltration/infiltration via hidden devices, drones, etc. We identify a critical gap in this context: the absence of low-latency systems for high-quality and instantaneous monitoring of cellular transmi…

MCM-DPO: Multifaceted Cross-Modal Direct Preference Optimization for Alt-text Generation
Jinlan Fu, Shenzhen Huangfu, Hao Fei, Yichong Huang, Xiaoyu Shen, Xipeng Qiu, See-Kiong Ng
https://arxiv.org/abs/2510.00647

MCM-DPO: Multifaceted Cross-Modal Direct Preference Optimization for Alt-text Generation
The alt-text generation task produces concise, context-relevant descriptions of images, enabling blind and low-vision users to access online images. Despite the capabilities of large vision-language models, alt-text generation performance remains limited due to noisy user annotations, inconsistent standards, and MLLMs' insensitivity to contextual information. Previous efforts to fine-tune MLLMs using supervised fine-tuning (SFT) have struggled, as SFT relies on accurate target annotations, whic…

Counterfactual Scenarios for Automated Planning
Nicola Gigante, Francesco Leofante, Andrea Micheli
https://arxiv.org/abs/2508.21521 https://arxiv.org/pdf/2…

Counterfactual Scenarios for Automated Planning
Counterfactual Explanations (CEs) are a powerful technique used to explain Machine Learning models by showing how the input to a model should be minimally changed for the model to produce a different output. Similar proposals have been made in the context of Automated Planning, where CEs have been characterised in terms of minimal modifications to an existing plan that would result in the satisfaction of a different goal. While such explanations may help diagnose faults and reason about the cha…

Anatomy-DT: A Cross-Diffusion Digital Twin for Anatomical Evolution
Moinak Bhattacharya, Gagandeep Singh, Prateek Prasanna
https://arxiv.org/abs/2509.25280 https://

Anatomy-DT: A Cross-Diffusion Digital Twin for Anatomical Evolution
Accurately modeling the spatiotemporal evolution of tumor morphology from baseline imaging is a pre-requisite for developing digital twin frameworks that can simulate disease progression and treatment response. Most existing approaches primarily characterize tumor growth while neglecting the concomitant alterations in adjacent anatomical structures. In reality, tumor evolution is highly non-linear and heterogeneous, shaped not only by therapeutic interventions but also by its spatial context an…

Asymptotic Schwarzschild solutions in $f(R)$ gravity and their observable effects on the photon sphere of black holes
Miguel Aparicio Resco
https://arxiv.org/abs/2510.00702 http…

Asymptotic Schwarzschild solutions in $f(R)$ gravity and their observable effects on the photon sphere of black holes
We investigate asymptotic Schwarzschild exterior solutions in the context of modified gravity theories, specifically within the framework of $f(R)$ gravity, where the asymptotic behavior recovers the standard Schwarzschild solution of General Relativity. Unlike previous studies that rely mainly on analytical approximations, our approach combines asymptotic analysis with numerical integration of the underlying differential equations. Using these solutions, we analyze strong lensing effects to ob…

The Demon is in Ambiguity: Revisiting Situation Recognition with Single Positive Multi-Label Learning
Yiming Lin, Yuchen Niu, Shang Wang, Kaizhu Huang, Qiufeng Wang, Xiao-Bo Jin
https://arxiv.org/abs/2508.21816

The Demon is in Ambiguity: Revisiting Situation Recognition with Single Positive Multi-Label Learning
Context recognition (SR) is a fundamental task in computer vision that aims to extract structured semantic summaries from images by identifying key events and their associated entities. Specifically, given an input image, the model must first classify the main visual events (verb classification), then identify the participating entities and their semantic roles (semantic role labeling), and finally localize these entities in the image (semantic role localization). Existing methods treat verb cl…

COM-BOM: Bayesian Exemplar Search for Efficiently Exploring the Accuracy-Calibration Pareto Frontier
Gaoxiang Luo, Aryan Deshwal
https://arxiv.org/abs/2510.01178 https://…

COM-BOM: Bayesian Exemplar Search for Efficiently Exploring the Accuracy-Calibration Pareto Frontier
Selecting an optimal set of exemplars is critical for good performance of in-context learning. However, prior exemplar search methods narrowly optimize for predictive accuracy, critically neglecting model calibration--a key determinant of trustworthiness and safe deployment. In this paper, we formulate exemplar selection as a multi-objective optimization problem, explicitly targeting both the maximization of predictive accuracy and the minimization of expected calibration error. We solve this p…

Interpreting Language Models Through Concept Descriptions: A Survey
Nils Feldhus, Laura Kopf
https://arxiv.org/abs/2510.01048 https://arxiv.org/pdf/2510.01…

Interpreting Language Models Through Concept Descriptions: A Survey
Understanding the decision-making processes of neural networks is a central goal of mechanistic interpretability. In the context of Large Language Models (LLMs), this involves uncovering the underlying mechanisms and identifying the roles of individual model components such as neurons and attention heads, as well as model abstractions such as the learned sparse features extracted by Sparse Autoencoders (SAEs). A rapidly growing line of work tackles this challenge by using powerful generator mod…

Unit Test Update through LLM-Driven Context Collection and Error-Type-Aware Refinement
Yuanhe Zhang, Zhiquan Yang, Shengyi Pan, Zhongxin Liu
https://arxiv.org/abs/2509.24419 htt…

Unit Test Update through LLM-Driven Context Collection and Error-Type-Aware Refinement
Unit testing is critical for ensuring software quality and software system stability. The current practice of manually maintaining unit tests suffers from low efficiency and the risk of delayed or overlooked fixes. Therefore, an automated approach is required to instantly update unit tests, with the capability to both repair and enhance unit tests. However, existing automated test maintenance methods primarily focus on repairing broken tests, neglecting the scenario of enhancing existing tests …

MC-GNNAS-Dock: Multi-criteria GNN-based Algorithm Selection for Molecular Docking
Siyuan Cao, Hongxuan Wu, Jiabao Brad Wang, Yiliang Yuan, Mustafa Misir
https://arxiv.org/abs/2509.26377

MC-GNNAS-Dock: Multi-criteria GNN-based Algorithm Selection for Molecular Docking
Molecular docking is a core tool in drug discovery for predicting ligand-target interactions. Despite the availability of diverse search-based and machine learning approaches, no single docking algorithm consistently dominates, as performance varies by context. To overcome this challenge, algorithm selection frameworks such as GNNAS-Dock, built on graph neural networks, have been proposed. This study introduces an enhanced system, MC-GNNAS-Dock, with three key advances. First, a multi-criteria …

ProfVLM: A Lightweight Video-Language Model for Multi-View Proficiency Estimation
Edoardo Bianchi, Jacopo Staiano, Antonio Liotta
https://arxiv.org/abs/2509.26278 https://

ProfVLM: A Lightweight Video-Language Model for Multi-View Proficiency Estimation
Existing approaches to skill proficiency estimation often rely on black-box video classifiers, ignoring multi-view context and lacking explainability. We present ProfVLM, a compact vision-language model that reformulates this task as generative reasoning: it jointly predicts skill level and generates expert-like feedback from egocentric and exocentric videos. Central to our method is an AttentiveGatedProjector that dynamically fuses multi-view features, projected from a frozen TimeSformer backb…

Data-Centric Elastic Pipeline Parallelism for Efficient Long-Context LLM Training
Shiju Wang, Yujie Wang, Ao Sun, Fangcheng Fu, Zijian Zhu, Bin Cui, Xu Han, Kaisheng Ma
https://arxiv.org/abs/2509.21275

Data-Centric Elastic Pipeline Parallelism for Efficient Long-Context LLM Training
Long context training is crucial for LLM's context extension. Existing schemes, such as sequence parallelism, incur substantial communication overhead. Pipeline parallelism (PP) reduces this cost, but its effectiveness hinges on partitioning granularity. Batch-level PP dividing input samples exhibits high memory consumption in long-context scenario, whereas token-level PP splitting sequences into slices alleviates memory overhead but may incur hardware under-utilization. This trade-off motivate…

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[7/7]:
- Pretrain-Test Task Alignment Governs Generalization in In-Context Learning
Mary I. Letey, Jacob A. Zavatone-Veth, Yue M. Lu, Cengiz Pehlevan

Hybrid Dialogue State Tracking for Persian Chatbots: A Language Model-Based Approach
Samin Mahdipour Aghabagher, Saeedeh Momtazi
https://arxiv.org/abs/2510.01052 https://…

Hybrid Dialogue State Tracking for Persian Chatbots: A Language Model-Based Approach
Dialogue State Tracking (DST) is an essential element of conversational AI with the objective of deeply understanding the conversation context and leading it toward answering user requests. Due to high demands for open-domain and multi-turn chatbots, the traditional rule-based DST is not efficient enough, since it cannot provide the required adaptability and coherence for human-like experiences in complex conversations. This study proposes a hybrid DST model that utilizes rule-based methods alo…

AI Compute Architecture and Evolution Trends
Bor-Sung Liang
https://arxiv.org/abs/2508.21394 https://arxiv.org/pdf/2508.21394

AI Compute Architecture and Evolution Trends
The focus of AI development has shifted from academic research to practical applications. However, AI development faces numerous challenges at various levels. This article will attempt to analyze the opportunities and challenges of AI from several different perspectives using a structured approach. This article proposes a seven-layer model for AI compute architecture, including Physical Layer, Link Layer, Neural Network Layer, Context Layer, Agent Layer, Orchestrator Layer, and Application Laye…

Why Stop at Words? Unveiling the Bigger Picture through Line-Level OCR
Shashank Vempati, Nishit Anand, Gaurav Talebailkar, Arpan Garai, Chetan Arora
https://arxiv.org/abs/2508.21693

Why Stop at Words? Unveiling the Bigger Picture through Line-Level OCR
Conventional optical character recognition (OCR) techniques segmented each character and then recognized. This made them prone to error in character segmentation, and devoid of context to exploit language models. Advances in sequence to sequence translation in last decade led to modern techniques first detecting words and then inputting one word at a time to a model to directly output full words as sequence of characters. This allowed better utilization of language models and bypass error-prone…

ErrorPrism: Reconstructing Error Propagation Paths in Cloud Service Systems
Junsong Pu, Yichen Li, Zhuangbin Chen, Jinyang Liu, Zhihan Jiang, Jianjun Chen, Rui Shi, Zibin Zheng, Tieying Zhang
https://arxiv.org/abs/2509.26463

ErrorPrism: Reconstructing Error Propagation Paths in Cloud Service Systems
Reliability management in cloud service systems is challenging due to the cascading effect of failures. Error wrapping, a practice prevalent in modern microservice development, enriches errors with context at each layer of the function call stack, constructing an error chain that describes a failure from its technical origin to its business impact. However, this also presents a significant traceability problem when recovering the complete error propagation path from the final log message back t…

GRAD: Generative Retrieval-Aligned Demonstration Sampler for Efficient Few-Shot Reasoning
Oussama Gabouj, Kamel Charaf, Ivan Zakazov, Nicolas Baldwin, Robert West
https://arxiv.org/abs/2510.01165

GRAD: Generative Retrieval-Aligned Demonstration Sampler for Efficient Few-Shot Reasoning
Large Language Models (LLMs) achieve strong performance across diverse tasks, but their effectiveness often depends on the quality of the provided context. Retrieval-Augmented Generation (RAG) enriches prompts with external information, but its reliance on static databases constrains adaptability and can result in irrelevant demonstrations. In this work, we propose a Generative Retrieval-Aligned Demonstrator (GRAD), a dynamic demonstration-based approach where an LLM model is trained to generat…

MultiFluxAI Enhancing Platform Engineering with Advanced Agent-Orchestrated Retrieval Systems
Sri Ram Macharla, Sridhar Murthy J, Anjaneyulu Pasala
https://arxiv.org/abs/2508.21307

MultiFluxAI Enhancing Platform Engineering with Advanced Agent-Orchestrated Retrieval Systems
MultiFluxAI is an innovative AI platform developed to address the challenges of managing and integrating vast, disparate data sources in product engineering across application domains. It addresses both current and new service related queries that enhance user engagement in the digital ecosystem. This platform leverages advanced AI techniques, such as Generative AI, vectorization, and agentic orchestration to provide dynamic and context-aware responses to complex user queries.

TTT3R: 3D Reconstruction as Test-Time Training
Xingyu Chen, Yue Chen, Yuliang Xiu, Andreas Geiger, Anpei Chen
https://arxiv.org/abs/2509.26645 https://arxi…

TTT3R: 3D Reconstruction as Test-Time Training
Modern Recurrent Neural Networks have become a competitive architecture for 3D reconstruction due to their linear-time complexity. However, their performance degrades significantly when applied beyond the training context length, revealing limited length generalization. In this work, we revisit the 3D reconstruction foundation models from a Test-Time Training perspective, framing their designs as an online learning problem. Building on this perspective, we leverage the alignment confidence betw…

Bridging Developer Instructions and Code Completion Through Instruction-Aware Fill-in-the-Middle Paradigm
Zhensu Sun, Chengran Yang, Chao Peng, Pengfei Gao, Xiaoning Du, Li Li, David Lo
https://arxiv.org/abs/2509.24637

Bridging Developer Instructions and Code Completion Through Instruction-Aware Fill-in-the-Middle Paradigm
Large Language Models (LLMs) have significantly advanced code completion, yet they often fail when the developer's intent is underspecified in the code context. To address this, developers usually add natural language instructions (e.g., comments) into the code context to clarify their intent. However, existing code LLMs applied for code completion systems merely undergo a fill-in-the-middle (FIM) pre-training, which struggles to leverage this information effectively due to the lack of instruct…

CoT Vectors: Transferring and Probing the Reasoning Mechanisms of LLMs
Li Li, Ziyi Wang, Yongliang Wu, Jianfei Cai, Xu Yang
https://arxiv.org/abs/2510.00579 https://

CoT Vectors: Transferring and Probing the Reasoning Mechanisms of LLMs
Chain-of-Thought (CoT) prompting has emerged as a powerful approach to enhancing the reasoning capabilities of Large Language Models (LLMs). However, existing implementations, such as in-context learning and fine-tuning, remain costly and inefficient. To improve CoT reasoning at a lower cost, and inspired by the task vector paradigm, we introduce CoT Vectors, compact representations that encode task-general, multi-step reasoning knowledge. Through experiments with Extracted CoT Vectors, we obse…

Rethinking Transformer Connectivity: TLinFormer, A Path to Exact, Full Context-Aware Linear Attention
Zhongpan Tang
https://arxiv.org/abs/2508.20407 https://

Rethinking Transformer Connectivity: TLinFormer, A Path to Exact, Full Context-Aware Linear Attention
The Transformer architecture has become a cornerstone of modern artificial intelligence, but its core self-attention mechanism suffers from a complexity bottleneck that scales quadratically with sequence length, severely limiting its application in long-sequence tasks. To address this challenge, existing linear attention methods typically sacrifice model performance by relying on data-agnostic kernel approximations or restrictive context selection. This paper returns to the first principles of …

Gather-Scatter Mamba: Accelerating Propagation with Efficient State Space Model
Hyun-kyu Ko, Youbin Kim, Jihyeon Park, Dongheok Park, Gyeongjin Kang, Wonjun Cho, Hyung Yi, Eunbyung Park
https://arxiv.org/abs/2510.00862

Gather-Scatter Mamba: Accelerating Propagation with Efficient State Space Model
State Space Models (SSMs)-most notably RNNs-have historically played a central role in sequential modeling. Although attention mechanisms such as Transformers have since dominated due to their ability to model global context, their quadratic complexity and limited scalability make them less suited for long sequences. Video super-resolution (VSR) methods have traditionally relied on recurrent architectures to propagate features across frames. However, such approaches suffer from well-known issue…

Model Context Protocols in Adaptive Transport Systems: A Survey
Gaurab Chhetri, Shriyank Somvanshi, Md Monzurul Islam, Shamyo Brotee, Mahmuda Sultana Mimi, Dipti Koirala, Biplov Pandey, Subasish Das
https://arxiv.org/abs/2508.19239

Model Context Protocols in Adaptive Transport Systems: A Survey
The rapid expansion of interconnected devices, autonomous systems, and AI applications has created severe fragmentation in adaptive transport systems, where diverse protocols and context sources remain isolated. This survey provides the first systematic investigation of the Model Context Protocol (MCP) as a unifying paradigm, highlighting its ability to bridge protocol-level adaptation with context-aware decision making. Analyzing established literature, we show that existing efforts have impli…

Med-RewardBench: Benchmarking Reward Models and Judges for Medical Multimodal Large Language Models
Meidan Ding, Jipeng Zhang, Wenxuan Wang, Cheng-Yi Li, Wei-Chieh Fang, Hsin-Yu Wu, Haiqin Zhong, Wenting Chen, Linlin Shen
https://arxiv.org/abs/2508.21430

Med-RewardBench: Benchmarking Reward Models and Judges for Medical Multimodal Large Language Models
Multimodal large language models (MLLMs) hold significant potential in medical applications, including disease diagnosis and clinical decision-making. However, these tasks require highly accurate, context-sensitive, and professionally aligned responses, making reliable reward models and judges critical. Despite their importance, medical reward models (MRMs) and judges remain underexplored, with no dedicated benchmarks addressing clinical requirements. Existing benchmarks focus on general MLLM c…

KeySG: Hierarchical Keyframe-Based 3D Scene Graphs
Abdelrhman Werby, Dennis Rotondi, Fabio Scaparro, Kai O. Arras
https://arxiv.org/abs/2510.01049 https://…

KeySG: Hierarchical Keyframe-Based 3D Scene Graphs
In recent years, 3D scene graphs have emerged as a powerful world representation, offering both geometric accuracy and semantic richness. Combining 3D scene graphs with large language models enables robots to reason, plan, and navigate in complex human-centered environments. However, current approaches for constructing 3D scene graphs are semantically limited to a predefined set of relationships, and their serialization in large environments can easily exceed an LLM's context window. We introdu…

Granite Embedding R2 Models
Parul Awasthy, Aashka Trivedi, Yulong Li, Meet Doshi, Riyaz Bhat, Vignesh P, Vishwajeet Kumar, Yushu Yang, Bhavani Iyer, Abraham Daniels, Rudra Murthy, Ken Barker, Martin Franz, Madison Lee, Todd Ward, Salim Roukos, David Cox, Luis Lastras, Jaydeep Sen, Radu Florian
https://arxiv.org/abs/2508.21085

Granite Embedding R2 Models
We introduce the Granite Embedding R2 models, a comprehensive family of high-performance English encoder-based embedding models engineered for enterprise-scale dense retrieval applications. Building upon our first-generation release, these models deliver substantial improvements, including 16x expanded context length (8,192 tokens), state-of-the-art performance across diverse retrieval domains - text, code, long-document search, multi-turn conversational, and tabular data - and measurable speed…

Middo: Model-Informed Dynamic Data Optimization for Enhanced LLM Fine-Tuning via Closed-Loop Learning
Zinan Tang, Xin Gao, Qizhi Pei, Zhuoshi Pan, Mengzhang Cai, Jiang Wu, Conghui He, Lijun Wu
https://arxiv.org/abs/2508.21589

Middo: Model-Informed Dynamic Data Optimization for Enhanced LLM Fine-Tuning via Closed-Loop Learning
Supervised Fine-Tuning (SFT) Large Language Models (LLM) fundamentally rely on high-quality training data. While data selection and data synthesis are two common strategies to improve data quality, existing approaches often face limitations in static dataset curation that fail to adapt to evolving model capabilities. In this paper, we introduce Middo, a self-evolving Model-informed dynamic data optimization framework that uses model-aware data selection and context-preserving data refinement. U…

ILRe: Intermediate Layer Retrieval for Context Compression in Causal Language Models
Manlai Liang, Mandi Liu, Jiangzhou Ji, Huaijun Li, Haobo Yang, Yaohan He, Jinlong Li
https://arxiv.org/abs/2508.17892

ILRe: Intermediate Layer Retrieval for Context Compression in Causal Language Models
Large Language Models (LLMs) have demonstrated success across many benchmarks. However, they still exhibit limitations in long-context scenarios, primarily due to their short effective context length, quadratic computational complexity, and high memory overhead when processing lengthy inputs. To mitigate these issues, we introduce a novel context compression pipeline, called Intermediate Layer Retrieval (ILRe), which determines one intermediate decoder layer offline, encodes context by streamin…