# Research Papers — 2025-11 · bundle page 51 of 197 > Full-text markdown of papers #2501–2550 of 9830 published in 2025-11, newest first. AI-readable mirrors of the full 9830-paper 2025-11 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2025-11/page/51 [← Previous bundle](https://aipod.dev/papers/archive/2025-11/page/50.md) [Next bundle →](https://aipod.dev/papers/archive/2025-11/page/52.md) ### DL-CapsNet: A Deep and Light Capsule Network Authors: Pouya Shiri, Amirali Baniasadi Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.00061v1 Capsule Network (CapsNet) is among the promising classifiers and a possible successor of the classifiers built based on Convolutional Neural Network (CNN). CapsNet is more accurate than CNNs in detecting images with overlapping categories and those with applied affine transformations. In this work, we propose a deep variant of CapsNet consisting of several capsule layers. In addition, we design the Capsule Summarization layer to reduce the complexity by reducing the number of parameters. DL-CapsNet, while being highly accurate, employs a small number of parameters and delivers faster training and inference. DL-CapsNet can process complex datasets with a high number of categories. --- ### PEFT-DML: Parameter-Efficient Fine-Tuning Deep Metric Learning for Robust Multi-Modal 3D Object Detection in Autonomous Driving Authors: Abdolazim Rezaei, Mehdi Sookhak Published: 2025-11-23 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.00060v1 This study introduces PEFT-DML, a parameter-efficient deep metric learning framework for robust multi-modal 3D object detection in autonomous driving. Unlike conventional models that assume fixed sensor availability, PEFT-DML maps diverse modalities (LiDAR, radar, camera, IMU, GNSS) into a shared latent space, enabling reliable detection even under sensor dropout or unseen modality class combinations. By integrating Low-Rank Adaptation (LoRA) and adapter layers, PEFT-DML achieves significant training efficiency while enhancing robustness to fast motion, weather variability, and domain shifts. Experiments on benchmarks nuScenes demonstrate superior accuracy. --- ### SafeCiM: Investigating Resilience of Hybrid Floating-Point Compute-in-Memory Deep Learning Accelerators Authors: Swastik Bhattacharya, Sanjay Das, Anand Menon, Shamik Kundu, Arnab Raha, Kanad Basu Published: 2025-11-23 Categories: cs.AR, cs.LG arXiv: https://arxiv.org/abs/2512.00059v1 Deep Neural Networks (DNNs) continue to grow in complexity with Large Language Models (LLMs) incorporating vast numbers of parameters. Handling these parameters efficiently in traditional accelerators is limited by data-transmission bottlenecks, motivating Compute-in-Memory (CiM) architectures that integrate computation within or near memory to reduce data movement. Recent work has explored CiM designs using Floating-Point (FP) and Integer (INT) operations. FP computations typically deliver higher output quality due to their wider dynamic range and precision, benefiting precision-sensitive Generative AI applications. These include models such as LLMs, thus driving advancements in FP-CiM accelerators. However, the vulnerability of FP-CiM to hardware faults remains underexplored, posing a major reliability concern in mission-critical settings. To address this gap, we systematically analyze hardware fault effects in FP-CiM by introducing bit-flip faults at key computational stages, including digital multipliers, CiM memory cells, and digital adder trees. Experiments with Convolutional Neural Networks (CNNs) such as AlexNet and state-of-the-art LLMs including LLaMA-3.2-1B and Qwen-0.3B-Base reveal how faults at each stage affect inference accuracy. Notably, a single adder fault can reduce LLM accuracy to 0%. Based on these insights, we propose a fault-resilient design, SafeCiM, that mitigates fault impact far better than a naive FP-CiM with a pre-alignment stage. For example, with 4096 MAC units, SafeCiM reduces accuracy degradation by up to 49x for a single adder fault compared to the baseline FP-CiM architecture. --- ### Semantics as a Shield: Label Disguise Defense (LDD) against Prompt Injection in LLM Sentiment Classification Authors: Yanxi Li, Ruocheng Shan Published: 2025-11-23 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2511.21752v3 Prompt injection can cause large language models (LLMs) to deviate from their intended tasks by following instructions embedded in untrusted input. We investigate whether changing a classifier's output labels improves accuracy under such attacks, using sentiment classification as a case study. Label Disguise Defense (LDD) replaces positive/negative with alternatives such as good/bad or cat/dog, introduced through demonstrations, and maps predictions back to sentiment classes. We analyze retained aggregate results for 200 IMDb reviews, nine recorded model configurations, and eight alternative label pairs under a fixed attack that always requests the incorrect class using the original labels. Averaged over four pairs per group, sentiment-associated labels outperform the other tested labels in all nine configurations, but exceed a baseline using the original labels and the same demonstration and attack conditions in only three. At least one alternative pair improves on this baseline in seven configurations, with the best pairs identified retrospectively. These results show that a consistent advantage between label groups does not imply consistent improvement over standard labels. Missing matched evaluations on unattacked reviews prevent separating changes in ordinary classification performance from changes in the effect of the attack. --- ### SO-Bench: A Structural Output Evaluation of Multimodal LLMs Authors: Di Feng, Kaixin Ma, Feng Nan, Haofeng Chen, Bohan Zhai, David Griffiths, Mingfei Gao, Zhe Gan, Eshan Verma, Yinfei Yang, Zhifeng Chen, Afshin Dehghan Published: 2025-11-23 Categories: cs.CV, cs.AI, cs.CL, cs.RO arXiv: https://arxiv.org/abs/2511.21750v3 Multimodal large language models (MLLMs) are increasingly deployed in real-world, agentic settings where outputs must not only be correct, but also conform to predefined data schemas. Despite recent progress in structured generation in textual domain, there is still no benchmark that systematically evaluates schema-grounded information extraction and reasoning over visual inputs. In this work, we conduct a comprehensive study of visual structural output capabilities for MLLMs with our carefully designed SO-Bench benchmark. Covering four visual domains, including UI screens, natural images, documents, and charts, SO-Bench is built from over 6.5K diverse JSON schemas and 1.8K curated image-schema pairs with human-verified quality. Benchmarking experiments on open-sourced and frontier proprietary models reveal persistent gaps in predicting accurate, schema compliant outputs, highlighting the need for better multimodal structured reasoning. Beyond benchmarking, we further conduct training experiments to largely improve the model's structured output capability. We make the benchmark and evaluation publicly available at https://github.com/apple/ml-sobench --- ### Proactive Defense: Compound AI for Detecting Persuasion Attacks and Measuring Inoculation Effectiveness Authors: Svitlana Volkova, Will Dupree, Hsien-Te Kao, Peter Bautista, Gabe Ganberg, Jeff Beaubien, Laura Cassani Published: 2025-11-23 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2511.21749v1 This paper introduces BRIES, a novel compound AI architecture designed to detect and measure the effectiveness of persuasion attacks across information environments. We present a system with specialized agents: a Twister that generates adversarial content employing targeted persuasion tactics, a Detector that identifies attack types with configurable parameters, a Defender that creates resilient content through content inoculation, and an Assessor that employs causal inference to evaluate inoculation effectiveness. Experimenting with the SemEval 2023 Task 3 taxonomy across the synthetic persuasion dataset, we demonstrate significant variations in detection performance across language agents. Our comparative analysis reveals significant performance disparities with GPT-4 achieving superior detection accuracy on complex persuasion techniques, while open-source models like Llama3 and Mistral demonstrated notable weaknesses in identifying subtle rhetorical, suggesting that different architectures encode and process persuasive language patterns in fundamentally different ways. We show that prompt engineering dramatically affects detection efficacy, with temperature settings and confidence scoring producing model-specific variations; Gemma and GPT-4 perform optimally at lower temperatures while Llama3 and Mistral show improved capabilities at higher temperatures. Our causal analysis provides novel insights into socio-emotional-cognitive signatures of persuasion attacks, revealing that different attack types target specific cognitive dimensions. This research advances generative AI safety and cognitive security by quantifying LLM-specific vulnerabilities to persuasion attacks and delivers a framework for enhancing human cognitive resilience through structured interventions before exposure to harmful content. --- ### Building Domain-Specific Small Language Models via Guided Data Generation Authors: Aman Kumar, Ekant Muljibhai Amin, Xian Yeow Lee, Lasitha Vidyaratne, Ahmed K. Farahat, Dipanjan D. Ghosh, Yuta Koreeda, Chetan Gupta Published: 2025-11-23 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2511.21748v1 Large Language Models (LLMs) have shown remarkable success in supporting a wide range of knowledge-intensive tasks. In specialized domains, there is growing interest in leveraging LLMs to assist subject matter experts with domain-specific challenges. However, deploying LLMs as SaaS solutions raises data privacy concerns, while many open-source models demand significant computational resources for effective domain adaptation and deployment. A promising alternative is to develop smaller, domain-specialized LLMs, though this approach is often constrained by the lack of high-quality domain-specific training data. In this work, we address these limitations by presenting a cost-efficient and scalable training pipeline that combines guided synthetic data generation from a small seed corpus with bottom-up domain data curation. Our pipeline integrates Domain-Adaptive Pretraining (DAPT), Domain-specific Supervised Fine-tuning (DSFT), and Direct Preference Optimization (DPO) to train effective small-scale models for specialized use cases. We demonstrate this approach through DiagnosticSLM, a 3B-parameter domain-specific model tailored for fault diagnosis, root cause analysis, and repair recommendation in industrial settings. To evaluate model performance, we introduce four domain-specific benchmarks: multiple-choice questions (DiagnosticMCQ), question answering (DiagnosticQA), sentence completion (DiagnosticComp), and summarization (DiagnosticSum). DiagnosticSLM achieves up to 25% accuracy improvement over open-source models of comparable or larger size (2B-9B) on the MCQ task, while also outperforming or matching them in other tasks, demonstrating effective domain-specific reasoning and generalization capabilities. --- ### QuantumChem-200K: A Large-Scale Open Organic Molecular Dataset for Quantum-Chemistry Property Screening and Language Model Benchmarking Authors: Yinqi Zeng, Renjie Li Published: 2025-11-23 Categories: physics.chem-ph, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.21747v1 The discovery of next-generation photoinitiators for two-photon polymerization (TPP) is hindered by the absence of large, open datasets containing the quantum-chemical and photophysical properties required to model photodissociation and excited-state behavior. Existing molecular datasets typically provide only basic physicochemical descriptors and therefore cannot support data-driven screening or AI-assisted design of photoinitiators. To address this gap, we introduce QuantumChem-200K, a large-scale dataset of over 200,000 organic molecules annotated with eleven quantum-chemical properties, including two-photon absorption (TPA) cross sections, TPA spectral ranges, singlet-triplet intersystem crossing (ISC) energies, toxicity and synthetic accessibility scores, hydrophilicity, solubility, boiling point, molecular weight, and aromaticity. These values are computed using a hybrid workflow that integrates density function theory (DFT), semi-empirical excited-state methods, atomistic quantum solvers, and neural-network predictors. Using QuantumChem-200K, we fine tune the open-source Qwen2.5-32B large language model to create a chemistry AI assistant capable of forward property prediction from SMILES. Benchmarking on 3000 unseen molecules from VQM24 and ZINC20 demonstrates that domain-specific fine-tuning significantly improves accuracy over GPT-4o, Llama-3.1-70B, and the base Qwen2.5-32B model, particularly for TPA and ISC predictions central to photoinitiator design. QuantumChem-200K and the corresponding AI assistant together provide the first scalable platform for high-throughput, LLM-driven photoinitiator screening and accelerated discovery of photosensitive materials. --- ### Reasoning With a Star: A Heliophysics Dataset and Benchmark for Agentic Scientific Reasoning Authors: Kevin Lee, Russell Spiewak, James Walsh Published: 2025-11-23 Categories: cs.AI, astro-ph.SR, cs.LG, physics.space-ph arXiv: https://arxiv.org/abs/2511.20694v2 Scientific reasoning through Large Language Models in heliophysics involves more than just recalling facts: it requires incorporating physical assumptions, maintaining consistent units, and providing clear scientific formats through coordinated approaches. To address these challenges, we present Reasoning With a Star, a newly contributed heliophysics dataset applicable to reasoning; we also provide an initial benchmarking approach. Our data are constructed from National Aeronautics and Space Administration & University Corporation for Atmospheric Research Living With a Star summer school problem sets and compiled into a readily consumable question-and-answer structure with question contexts, reasoning steps, expected answer type, ground-truth targets, format hints, and metadata. A programmatic grader checks the predictions using unit-aware numerical tolerance, symbolic equivalence, and schema validation. We benchmark a single-shot baseline and four multi-agent patterns, finding that decomposing workflows through systems engineering principles outperforms direct prompting on problems requiring deductive reasoning rather than pure inductive recall. --- ### $A^2Flow:$ Automating Agentic Workflow Generation via Self-Adaptive Abstraction Operators Authors: Mingming Zhao, Xiaokang Wei, Yuanqi Shao, Kaiwen Zhou, Lin Yang, Siwei Rao, Junhui Zhan, Zhitang Chen Published: 2025-11-23 Categories: cs.AI, cs.MA arXiv: https://arxiv.org/abs/2511.20693v1 Large language models (LLMs) have shown strong potential in automating the design of agentic workflows. However, existing methods still rely heavily on manually predefined operators, limiting generalization and scalability. To address this issue, we propose $A^2Flow$, a fully automated framework for agentic workflow generation based on self-adaptive abstraction operators. $A^2Flow$ employs a three-stage operator extraction process: 1) Case-based Initial Operator Generation: leveraging expert demonstrations and LLM reasoning to generate case-specific operators; 2) Operator Clustering and Preliminary Abstraction: grouping similar operators across tasks to form preliminary abstractions; and 3) Deep Extraction for Abstract Execution Operators: applying long chain-of-thought prompting and multi-path reasoning to derive compact and generalizable execution operators. These operators serve as reusable building blocks for workflow construction without manual predefinition. Furthermore, we enhance node-level workflow search with an operator memory mechanism, which retains historical outputs to enrich context and improve decision-making. Experiments on general and embodied benchmarks show that $A^2Flow$ achieves a 2.4\% and 19.3\% average performance improvement and reduces resource usage by 37\% over state-of-the-art baselines. Homepage:https://github.com/pandawei-ele/A2FLOW --- ### Profile Generators: A Link between the Narrative and the Binary Matrix Representation Authors: Raoul H. Kutil, Georg Zimmermann, Barbara Strasser-Kirchweger, Christian Borgelt Published: 2025-11-23 Categories: cs.LG, cs.LO arXiv: https://arxiv.org/abs/2511.19506v1 Mental health disorders, particularly cognitive disorders defined by deficits in cognitive abilities, are described in detail in the DSM-5, which includes definitions and examples of signs and symptoms. A simplified, machine-actionable representation was developed to assess the similarity and separability of these disorders, but it is not suited for the most complex cases. Generating or applying a full binary matrix for similarity calculations is infeasible due to the vast number of symptom combinations. This research develops an alternative representation that links the narrative form of the DSM-5 with the binary matrix representation and enables automated generation of valid symptom combinations. Using a strict pre-defined format of lists, sets, and numbers with slight variations, complex diagnostic pathways involving numerous symptom combinations can be represented. This format, called the symptom profile generator (or simply generator), provides a readable, adaptable, and comprehensive alternative to a binary matrix while enabling easy generation of symptom combinations (profiles). Cognitive disorders, which typically involve multiple diagnostic criteria with several symptoms, can thus be expressed as lists of generators. Representing several psychotic disorders in generator form and generating all symptom combinations showed that matrix representations of complex disorders become too large to manage. The MPCS (maximum pairwise cosine similarity) algorithm cannot handle matrices of this size, prompting the development of a profile reduction method using targeted generator manipulation to find specific MPCS values between disorders. The generators allow easier creation of binary representations for large matrices and make it possible to calculate specific MPCS cases between complex disorders through conditional generators. --- ### Position: The Complexity of Perfect AI Alignment -- Formalizing the RLHF Trilemma Authors: Subramanyam Sahoo, Aman Chadha, Vinija Jain, Divya Chaudhary Published: 2025-11-23 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2511.19504v1 Reinforcement Learning from Human Feedback (RLHF) is widely used for aligning large language models, yet practitioners face a persistent puzzle: improving safety often reduces fairness, scaling to diverse populations becomes computationally intractable, and making systems robust often amplifies majority biases. We formalize this tension as the Alignment Trilemma: no RLHF system can simultaneously achieve (i) epsilon-representativeness across diverse human values, (ii) polynomial tractability in sample and compute complexity, and (iii) delta-robustness against adversarial perturbations and distribution shift. Through a complexity-theoretic analysis integrating statistical learning theory and robust optimization, we prove that achieving both representativeness (epsilon <= 0.01) and robustness (delta <= 0.001) for global-scale populations requires Omega(2^{d_context}) operations, which is super-polynomial in the context dimensionality. We show that current RLHF implementations resolve this trilemma by sacrificing representativeness: they collect only 10^3--10^4 samples from homogeneous annotator pools while 10^7--10^8 samples are needed for true global representation. Our framework provides a unified explanation for documented RLHF pathologies including preference collapse, sycophancy, and systematic bias amplification. We conclude with concrete directions for navigating these fundamental trade-offs through strategic relaxations of alignment requirements. --- ### CycleChemist: A Dual-Pronged Machine Learning Framework for Organic Photovoltaic Discovery Authors: Hou Hei Lam, Jiangjie Qiu, Xiuyuan Hu, Wentao Li, Fankun Zeng, Siwei Fu, Hao Zhang, Xiaonan Wang Published: 2025-11-23 Categories: cond-mat.mtrl-sci, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.19500v2 Organic photovoltaic (OPV) materials offer a promising path toward sustainable energy generation, but their development is limited by the difficulty of identifying high performance donor and acceptor pairs with strong power conversion efficiencies (PCEs). Existing design strategies typically focus on either the donor or the acceptor alone, rather than using a unified approach capable of modeling both components. In this work, we introduce a dual machine learning framework for OPV discovery that combines predictive modeling with generative molecular design. We present the Organic Photovoltaic Donor Acceptor Dataset (OPV2D), the largest curated dataset of its kind, containing 2000 experimentally characterized donor acceptor pairs. Using this dataset, we develop the Organic Photovoltaic Classifier (OPVC) to predict whether a material exhibits OPV behavior, and a hierarchical graph neural network that incorporates multi task learning and donor acceptor interaction modeling. This framework includes the Molecular Orbital Energy Estimator (MOE2) for predicting HOMO and LUMO energy levels, and the Photovoltaic Performance Predictor (P3) for estimating PCE. In addition, we introduce the Material Generative Pretrained Transformer (MatGPT) to produce synthetically accessible organic semiconductors, guided by a reinforcement learning strategy with three objective policy optimization. By linking molecular representation learning with performance prediction, our framework advances data driven discovery of high performance OPV materials. --- ### Beyond Binary Classification: A Semi-supervised Approach to Generalized AI-generated Image Detection Authors: Hong-Hanh Nguyen-Le, Van-Tuan Tran, Dinh-Thuc Nguyen, Nhien-An Le-Khac Published: 2025-11-23 Categories: cs.LG, cs.AI, cs.CR, cs.CV arXiv: https://arxiv.org/abs/2511.19499v1 The rapid advancement of generators (e.g., StyleGAN, Midjourney, DALL-E) has produced highly realistic synthetic images, posing significant challenges to digital media authenticity. These generators are typically based on a few core architectural families, primarily Generative Adversarial Networks (GANs) and Diffusion Models (DMs). A critical vulnerability in current forensics is the failure of detectors to achieve cross-generator generalization, especially when crossing architectural boundaries (e.g., from GANs to DMs). We hypothesize that this gap stems from fundamental differences in the artifacts produced by these \textbf{distinct architectures}. In this work, we provide a theoretical analysis explaining how the distinct optimization objectives of the GAN and DM architectures lead to different manifold coverage behaviors. We demonstrate that GANs permit partial coverage, often leading to boundary artifacts, while DMs enforce complete coverage, resulting in over-smoothing patterns. Motivated by this analysis, we propose the \textbf{Tri}archy \textbf{Detect}or (TriDetect), a semi-supervised approach that enhances binary classification by discovering latent architectural patterns within the "fake" class. TriDetect employs balanced cluster assignment via the Sinkhorn-Knopp algorithm and a cross-view consistency mechanism, encouraging the model to learn fundamental architectural distincts. We evaluate our approach on two standard benchmarks and three in-the-wild datasets against 13 baselines to demonstrate its generalization capability to unseen generators. --- ### Hierarchical Dual-Strategy Unlearning for Biomedical and Healthcare Intelligence Using Imperfect and Privacy-Sensitive Medical Data Authors: Yi Zhang, Chao Zhang, Zijian Li, Tianxiang Xu, Kunyu Zhang, Zhan Gao, Meinuo Li, Xiaohan Zhang, Qichao Qi, Bing Chen Published: 2025-11-23 Categories: cs.LG, cs.AI, cs.CR arXiv: https://arxiv.org/abs/2511.19498v2 Large language models (LLMs) exhibit exceptional performance but pose substantial privacy risks due to training data memorization, particularly within healthcare contexts involving imperfect or privacy-sensitive patient information. We present a hierarchical dual-strategy framework for selective knowledge unlearning that precisely removes specialized knowledge while preserving fundamental medical competencies. Our approach synergistically integrates geometric-constrained gradient updates to selectively modulate target parameters with concept-aware token-level interventions that distinguish between preservation-critical and unlearning-targeted tokens via a unified four-level medical concept hierarchy. Comprehensive evaluations on the MedMCQA (surgical) and MHQA (anxiety, depression, trauma) datasets demonstrate superior performance, achieving an 82.7% forgetting rate and 88.5% knowledge preservation. Notably, our framework maintains robust privacy guarantees while requiring modification of only 0.1% of parameters, addressing critical needs for regulatory compliance, auditability, and ethical standards in clinical research. --- ### PeriodNet: Boosting the Potential of Attention Mechanism for Time Series Forecasting Authors: Bowen Zhao, Huanlai Xing, Zhiwen Xiao, Jincheng Peng, Li Feng, Xinhan Wang, Rong Qu, Hui Li Published: 2025-11-23 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.19497v1 The attention mechanism has demonstrated remarkable potential in sequence modeling, exemplified by its successful application in natural language processing with models such as Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-trained Transformer (GPT). Despite these advancements, its utilization in time series forecasting (TSF) has yet to meet expectations. Exploring a better network structure for attention in TSF holds immense significance across various domains. In this paper, we present PeriodNet with a brand new structure to forecast univariate and multivariate time series. PeriodNet incorporates period attention and sparse period attention mechanism for analyzing adjacent periods. It enhances the mining of local characteristics, periodic patterns, and global dependencies. For efficient cross-variable modeling, we introduce an iterative grouping mechanism which can directly reduce the cross-variable redundancy. To fully leverage the extracted features on the encoder side, we redesign the entire architecture of the vanilla Transformer and propose a period diffuser for precise multi-period prediction. Through comprehensive experiments conducted on eight datasets, we demonstrate that PeriodNet outperforms six state-of-the-art models in both univariate and multivariate TSF scenarios in terms of mean square error and mean absolute error. In particular, PeriodNet achieves a relative improvement of 22% when forecasting time series with a length of 720, in comparison to other models based on the conventional encoder-decoder Transformer architecture. --- ### Xmodel-2.5: 1.3B Data-Efficient Reasoning SLM Authors: Yang Liu, Xiaolong Zhong, Ling Jiang Published: 2025-11-23 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.19496v1 Large language models deliver strong reasoning and tool-use skills, yet their computational demands make them impractical for edge or cost-sensitive deployments. We present \textbf{Xmodel-2.5}, a 1.3-billion-parameter small language model designed as a \emph{drop-in agent core}. Training with maximal-update parameterization ($μ$P) allows hyper-parameters tuned on a 20M-parameter proxy to transfer directly to the full model, even under the parameter-tied \emph{tie-word-embedding} architecture. A 1.4T-token Warmup--Stable--Decay curriculum is used, and we further show that \textbf{switching from AdamW to Muon during the decay phase} improves the 13-task reasoning average by 4.58\,\% while keeping every other hyper-parameter fixed, verifying that early AdamW stability can be paired with late Muon sharpening for better downstream performance. FP8-mixed-precision training balances accuracy and throughput. All checkpoints, recipes, and evaluation code are released under the Apache-2.0 license.\footnote{https://huggingface.co/XiaoduoAILab/Xmodel-2.5 and https://huggingface.co/XiaoduoAILab/Xmodel-2.5-history (training checkpoints).} Training code and evaluation harness: https://github.com/XiaoduoAILab/Xmodel-2.5. --- ### A Systematic Study of Compression Ordering for Large Language Models Authors: Shivansh Chhawri, Rahul Mahadik, Suparna Rooj Published: 2025-11-23 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.19495v1 Large Language Models (LLMs) require substantial computational resources, making model compression essential for efficient deployment in constrained environments. Among the dominant compression techniques: knowledge distillation, structured pruning, and low-bit quantization, their individual effects are well studied, but their interactions and optimal sequencing remain unclear. This work systematically examines how these techniques perform both independently and in combination when applied to the Qwen2.5 3B model. We evaluate multiple compression pipelines, including single, and proposed three-technique sequences, using perplexity, G-Eval, clarity, prompt alignment, and compression ratio as metrics. Our experiments show that quantization provides the greatest standalone compression, while pruning introduces moderate quality degradation. Critically, the ordering of techniques significantly affects the final model quality: the sequence Pruning, Knowledge Distillation, Quantization (P-KD-Q) yields the best balance, achieving a 3.68x compression ratio while preserving strong instruction-following and language understanding capabilities. Conversely, pipelines applying quantization early suffer severe performance degradation due to irreversible information loss that impairs subsequent training. Overall, this study offers practical insight into designing effective, ordering-aware compression pipelines for deploying LLMs in resource-limited settings. --- ### RFX: High-Performance Random Forests with GPU Acceleration and QLORA Compression Authors: Chris Kuchar Published: 2025-11-23 Categories: cs.LG, stat.ME, stat.ML arXiv: https://arxiv.org/abs/2511.19493v1 RFX (Random Forests X), where X stands for compression or quantization, presents a production-ready implementation of Breiman and Cutler's Random Forest classification methodology in Python. RFX v1.0 provides complete classification: out-of-bag error estimation, overall and local importance measures, proximity matrices with QLORA compression, case-wise analysis, and interactive visualization (rfviz)--all with CPU and GPU acceleration. Regression, unsupervised learning, CLIQUE importance, and RF-GAP proximity are planned for v2.0. This work introduces four solutions addressing the proximity matrix memory bottleneck limiting Random Forest analysis to ~60,000 samples: (1) QLORA (Quantized Low-Rank Adaptation) compression for GPU proximity matrices, reducing memory from 80GB to 6.4MB for 100k samples (12,500x compression with INT8 quantization) while maintaining 99% geometric structure preservation, (2) CPU TriBlock proximity--combining upper-triangle storage with block-sparse thresholding--achieving 2.7x memory reduction with lossless quality, (3) SM-aware GPU batch sizing achieving 95% GPU utilization, and (4) GPU-accelerated 3D MDS visualization computing embeddings directly from low-rank factors using power iteration. Validation across four implementation modes (GPU/CPU x case-wise/non-case-wise) demonstrates correct implementation. GPU achieves 1.4x speedup over CPU for overall importance with 500+ trees. Proximity computation scales from 1,000 to 200,000+ samples (requiring GPU QLORA), with CPU TriBlock filling the gap for medium-scale datasets (10K-50K samples). RFX v1.0 eliminates the proximity memory bottleneck, enabling proximity-based Random Forest analysis on datasets orders of magnitude larger than previously feasible. Open-source production-ready classification following Breiman and Cutler's original methodology. --- ### Forecasting AI Time Horizon Under Compute Slowdowns Authors: Parker Whitfill, Ben Snodin, Joel Becker Published: 2025-11-23 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2511.19492v1 METR's time horizon metric has grown exponentially since 2019, along with compute. However, it is unclear whether compute scaling will persist at current rates through 2030, raising the question of how possible compute slowdowns might impact AI agent capability forecasts. Given a model of time horizon as a function of training compute and algorithms, along with a model of how compute investment spills into algorithmic progress (which, notably, precludes the possibility of a software-only singularity), and the empirical fact that both time horizon and compute have grown at constant rates over 2019--2025, we derive that time horizon growth must be proportional to compute growth. We provide additional, albeit limited, experimental evidence consistent with this theory. We use our model to project time horizon growth under OpenAI's compute projection, finding substantial projected delays in some cases. For example, 1-month time horizons at $80\%$ reliability occur $7$ years later than simple trend extrapolation suggests. --- ### OpenCML: End-to-End Framework of Open-world Machine Learning to Learn Unknown Classes Incrementally Authors: Jitendra Parmar, Praveen Singh Thakur Published: 2025-11-23 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.19491v1 Open-world machine learning is an emerging technique in artificial intelligence, where conventional machine learning models often follow closed-world assumptions, which can hinder their ability to retain previously learned knowledge for future tasks. However, automated intelligence systems must learn about novel classes and previously known tasks. The proposed model offers novel learning classes in an open and continuous learning environment. It consists of two different but connected tasks. First, it discovers unknown classes in the data and creates novel classes; next, it learns how to perform class incrementally for each new class. Together, they enable continual learning, allowing the system to expand its understanding of the data and improve over time. The proposed model also outperformed existing approaches in open-world learning. Furthermore, it demonstrated strong performance in continuous learning, achieving a highest average accuracy of 82.54% over four iterations and a minimum accuracy of 65.87%. --- ### Generative Model-Aided Continual Learning for CSI Feedback in FDD mMIMO-OFDM Systems Authors: Guijun Liu, Yuwen Cao, Tomoaki Ohtsuki, Jiguang He, Shahid Mumtaz Published: 2025-11-23 Categories: cs.LG, cs.AI, cs.IT arXiv: https://arxiv.org/abs/2511.19490v1 Deep autoencoder (DAE) frameworks have demonstrated their effectiveness in reducing channel state information (CSI) feedback overhead in massive multiple-input multiple-output (mMIMO) orthogonal frequency division multiplexing (OFDM) systems. However, existing CSI feedback models struggle to adapt to dynamic environments caused by user mobility, requiring retraining when encountering new CSI distributions. Moreover, returning to previously encountered environments often leads to performance degradation due to catastrophic forgetting. Continual learning involves enabling models to incorporate new information while maintaining performance on previously learned tasks. To address these challenges, we propose a generative adversarial network (GAN)-based learning approach for CSI feedback. By using a GAN generator as a memory unit, our method preserves knowledge from past environments and ensures consistently high performance across diverse scenarios without forgetting. Simulation results show that the proposed approach enhances the generalization capability of the DAE framework while maintaining low memory overhead. Furthermore, it can be seamlessly integrated with other advanced CSI feedback models, highlighting its robustness and adaptability. --- ### Evolution without an Oracle: Driving Effective Evolution with LLM Judges Authors: Zhe Zhao, Yuheng Yang, Haibin Wen, Xiaojie Qiu, Zaixi Zhang, Qingfu Zhang Published: 2025-11-23 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2511.19489v1 The integration of Large Language Models (LLMs) with Evolutionary Computation (EC) has unlocked new frontiers in scientific discovery but remains shackled by a fundamental constraint: the reliance on an Oracle--an objective, machine-computable fitness function. This paper breaks this barrier by asking: Can evolution thrive in a purely subjective landscape governed solely by LLM judges? We introduce MADE (Multi-Agent Decomposed Evolution), a framework that tames the inherent noise of subjective evaluation through "Problem Specification." By decomposing vague instructions into specific, verifiable sub-requirements, MADE transforms high-variance LLM feedback into stable, precise selection pressure. The results are transformative: across complex benchmarks like DevAI and InfoBench, MADE outperforms strong baselines by over 50% in software requirement satisfaction (39.9% to 61.9%) and achieves a 95% perfect pass rate on complex instruction following. This work validates a fundamental paradigm shift: moving from optimizing "computable metrics" to "describable qualities," thereby unlocking evolutionary optimization for the vast open-ended domains where no ground truth exists. --- ### Building Resilient Information Ecosystems: Large LLM-Generated Dataset of Persuasion Attacks Authors: Hsien-Te Kao, Aleksey Panasyuk, Peter Bautista, William Dupree, Gabriel Ganberg, Jeffrey M. Beaubien, Laura Cassani, Svitlana Volkova Published: 2025-11-23 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2511.19488v1 Organization's communication is essential for public trust, but the rise of generative AI models has introduced significant challenges by generating persuasive content that can form competing narratives with official messages from government and commercial organizations at speed and scale. This has left agencies in a reactive position, often unaware of how these models construct their persuasive strategies, making it more difficult to sustain communication effectiveness. In this paper, we introduce a large LLM-generated persuasion attack dataset, which includes 134,136 attacks generated by GPT-4, Gemma 2, and Llama 3.1 on agency news. These attacks span 23 persuasive techniques from SemEval 2023 Task 3, directed toward 972 press releases from ten agencies. The generated attacks come in two mediums, press release statements and social media posts, covering both long-form and short-form communication strategies. We analyzed the moral resonance of these persuasion attacks to understand their attack vectors. GPT-4's attacks mainly focus on Care, with Authority and Loyalty also playing a role. Gemma 2 emphasizes Care and Authority, while Llama 3.1 centers on Loyalty and Care. Analyzing LLM-generated persuasive attacks across models will enable proactive defense, allow to create the reputation armor for organizations, and propel the development of both effective and resilient communications in the information ecosystem. --- ### The Generalized Proximity Forest Authors: Ben Shaw, Adam Rustad, Sofia Pelagalli Maia, Jake S. Rhodes, Kevin R. Moon Published: 2025-11-23 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2511.19487v1 Recent work has demonstrated the utility of Random Forest (RF) proximities for various supervised machine learning tasks, including outlier detection, missing data imputation, and visualization. However, the utility of the RF proximities depends upon the success of the RF model, which itself is not the ideal model in all contexts. RF proximities have recently been extended to time series by means of the distance-based Proximity Forest (PF) model, among others, affording time series analysis with the benefits of RF proximities. In this work, we introduce the generalized PF model, thereby extending RF proximities to all contexts in which supervised distance-based machine learning can occur. Additionally, we introduce a variant of the PF model for regression tasks. We also introduce the notion of using the generalized PF model as a meta-learning framework, extending supervised imputation capability to any pre-trained classifier. We experimentally demonstrate the unique advantages of the generalized PF model compared with both the RF model and the $k$-nearest neighbors model. --- ### Fine-Tune, Then Rectify Authors: Zikun Ye, Jinglong Zhao, Lei Wang Published: 2025-11-23 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.19486v3 Driven by recent advances in artificial intelligence, a growing literature has demonstrated the potential of using large language models (LLMs) as scalable surrogates to generate human-like responses. Two common approaches to improve the performance of LLMs include: fine-tuning, which aligns the LLM more closely with human responses, and rectification, which corrects biases in LLM outputs. In this paper, we develop a two-stage framework that combines fine-tuning and rectification, and optimally allocates limited labeled samples across the two stages. A key insight is that the conventional fine-tuning objective of minimizing mean squared prediction error is generally not aligned with the downstream rectification stage. For mean estimation, we propose to minimize the variance of the prediction errors; for general M-estimation, we propose to minimize a scalarized variance metric as the fine-tuning objective. Building on this insight, we leverage the scaling law of fine-tuning to optimally allocate the limited labeled human data between the fine-tuning and rectification stages. Our empirical analysis validates the fine-tuning scaling law and confirms that our proposed optimal allocation rule reliably identifies the optimal sample allocation. We demonstrate substantial efficiency gains in estimation and inference performance relative to fine-tuning or rectification alone, or to employing the conventional mean squared error objective within the fine-tuning then rectification framework. Such efficiency gains translate to significant cost savings for making reliable decisions. --- ### OmniTFT: Omni Target Forecasting for Vital Signs and Laboratory Result Trajectories in Multi Center ICU Data Authors: Wanzhe Xu, Yutong Dai, Yitao Yang, Martin Loza, Weihang Zhang, Yang Cui, Xin Zeng, Sung Joon Park, Kenta Nakai Published: 2025-11-23 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.19485v1 Accurate multivariate time-series prediction of vital signs and laboratory results is crucial for early intervention and precision medicine in intensive care units (ICUs). However, vital signs are often noisy and exhibit rapid fluctuations, while laboratory tests suffer from missing values, measurement lags, and device-specific bias, making integrative forecasting highly challenging. To address these issues, we propose OmniTFT, a deep learning framework that jointly learns and forecasts high-frequency vital signs and sparsely sampled laboratory results based on the Temporal Fusion Transformer (TFT). Specifically, OmniTFT implements four novel strategies to enhance performance: sliding window equalized sampling to balance physiological states, frequency-aware embedding shrinkage to stabilize rare-class representations, hierarchical variable selection to guide model attention toward informative feature clusters, and influence-aligned attention calibration to enhance robustness during abrupt physiological changes. By reducing the reliance on target-specific architectures and extensive feature engineering, OmniTFT enables unified modeling of multiple heterogeneous clinical targets while preserving cross-institutional generalizability. Across forecasting tasks, OmniTFT achieves substantial performance improvement for both vital signs and laboratory results on the MIMIC-III, MIMIC-IV, and eICU datasets. Its attention patterns are interpretable and consistent with known pathophysiology, underscoring its potential utility for quantitative decision support in clinical care. --- ### stable-pretraining-v1: Foundation Model Research Made Simple Authors: Randall Balestriero, Hugues Van Assel, Sami BuGhanem, Lucas Maes Published: 2025-11-23 Categories: cs.SE, cs.LG arXiv: https://arxiv.org/abs/2511.19484v1 Foundation models and self-supervised learning (SSL) have become central to modern AI, yet research in this area remains hindered by complex codebases, redundant re-implementations, and the heavy engineering burden of scaling experiments. We present stable-pretraining, a modular, extensible, and performance-optimized library built on top of PyTorch, Lightning, Hugging Face, and TorchMetrics. Unlike prior toolkits focused narrowly on reproducing state-of-the-art results, stable-pretraining is designed for flexibility and iteration speed: it unifies essential SSL utilities--including probes, collapse detection metrics, augmentation pipelines, and extensible evaluation routines--within a coherent and reliable framework. A central design principle is logging everything, enabling fine-grained visibility into training dynamics that makes debugging, monitoring, and reproducibility seamless. We validate the library by demonstrating its ability to generate new research insights with minimal overhead, including depthwise representation probing and the analysis of CLIP degradation under synthetic data finetuning. By lowering barriers to entry while remaining scalable to large experiments, stable-pretraining aims to accelerate discovery and expand the possibilities of foundation model research. --- ### Z-Space: A Multi-Agent Tool Orchestration Framework for Enterprise-Grade LLM Automation Authors: Qingsong He, Jing Nan, Jiayu Jiao, Liangjie Tang, Xiaodong Xu, Mengmeng Sun, Qingyao Wang, Minghui Yan Published: 2025-11-23 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2511.19483v1 Large Language Models can break through knowledge and timeliness limitations by invoking external tools within the Model Context Protocol framework to achieve automated execution of complex tasks. However, with the rapid growth of enterprise-scale MCP services, efficiently and accurately matching target functionalities among thousands of heterogeneous tools has become a core challenge restricting system practicality. Existing approaches generally rely on full-prompt injection or static semantic retrieval, facing issues including semantic disconnection between user queries and tool descriptions, context inflation in LLM input, and high inference latency. To address these challenges, this paper proposes Z-Space, a data-generation-oriented multi-agent collaborative tool invocation framework Z-Space. The Z-Space framework establishes a multi-agent collaborative architecture and tool filtering algorithm: (1) A structured semantic understanding of user queries is achieved through an intent parsing model; (2) A tool filtering module (FSWW) based on fused subspace weighted algorithm realizes fine-grained semantic alignment between intents and tools without parameter tuning; (3) An inference execution agent is constructed to support dynamic planning and fault-tolerant execution for multi-step tasks. This framework has been deployed in the Eleme platform's technical division, serving large-scale test data generation scenarios across multiple business units including Taotian, Gaode, and Hema. Production data demonstrates that the system reduces average token consumption in tool inference by 96.26\% while achieving a 92\% tool invocation accuracy rate, significantly enhancing the efficiency and reliability of intelligent test data generation systems. --- ### Human Experts' Evaluation of Generative AI for Contextualizing STEAM Education in the Global South Authors: Matthew Nyaaba, Macharious Nabang, Patrick Kyeremeh, Ibrahim Nantomah, Collins Owusu-Fordjour, Martin Ako, Bismark Nyaaba Akanzire, Kassim Korah Nantomah, Cecilia Issaka, Xiaoming Zhai Published: 2025-11-23 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2511.19482v3 STEAM education in many parts of the Global South remains abstract and weakly connected to learners sociocultural realities. This study examines how human experts evaluate the capacity of Generative AI (GenAI) to contextualize STEAM instruction in these settings. Using a convergent mixed-methods design grounded in human-centered and culturally responsive pedagogy, four STEAM education experts reviewed standardized Ghana NaCCA lesson plans and GenAI-generated lessons created with a customized Culturally Responsive Lesson Planner (CRLP). Quantitative data were collected with a validated 25-item Culturally Responsive Pedagogy Rubric assessing bias awareness, cultural representation, contextual relevance, linguistic responsiveness, and teacher agency. Qualitative reflections provided additional insight into the pedagogical and cultural dynamics of each lesson. Findings show that GenAI, especially through the CRLP, improved connections between abstract standards and learners lived experiences. Teacher Agency was the strongest domain, while Cultural Representation was the weakest. CRLP-generated lessons were rated as more culturally grounded and pedagogically engaging. However, GenAI struggled to represent Ghana's cultural diversity, often producing surface-level references, especially in Mathematics and Computing. Experts stressed the need for teacher mediation, community input, and culturally informed refinement of AI outputs. Future work should involve classroom trials, broader expert participation, and fine-tuning with Indigenous corpora. --- ### Robust Physical Adversarial Patches Using Dynamically Optimized Clusters Authors: Harrison Bagley, Will Meakin, Simon Lucey, Yee Wei Law, Tat-Jun Chin Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18656v1 Physical adversarial attacks on deep learning systems is concerning due to the ease of deploying such attacks, usually by placing an adversarial patch in a scene to manipulate the outcomes of a deep learning model. Training such patches typically requires regularization that improves physical realizability (e.g., printability, smoothness) and/or robustness to real-world variability (e.g. deformations, viewing angle, noise). One type of variability that has received little attention is scale variability. When a patch is rescaled, either digitally through downsampling/upsampling or physically through changing imaging distances, interpolation-induced color mixing occurs. This smooths out pixel values, resulting in a loss of high-frequency patterns and degrading the adversarial signal. To address this, we present a novel superpixel-based regularization method that guides patch optimization to scale-resilient structures. Our ap proach employs the Simple Linear Iterative Clustering (SLIC) algorithm to dynamically cluster pixels in an adversarial patch during optimization. The Implicit Function Theorem is used to backpropagate gradients through SLIC to update the superpixel boundaries and color. This produces patches that maintain their structure over scale and are less susceptible to interpolation losses. Our method achieves greater performance in the digital domain, and when realized physically, these performance gains are preserved, leading to improved physical performance. Real-world performance was objectively assessed using a novel physical evaluation protocol that utilizes screens and cardboard cut-outs to systematically vary real-world conditions. --- ### From Healthy Scans to Annotated Tumors: A Tumor Fabrication Framework for 3D Brain MRI Synthesis Authors: Nayu Dong, Townim Chowdhury, Hieu Phan, Mark Jenkinson, Johan Verjans, Zhibin Liao Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18654v1 The scarcity of annotated Magnetic Resonance Imaging (MRI) tumor data presents a major obstacle to accurate and automated tumor segmentation. While existing data synthesis methods offer promising solutions, they often suffer from key limitations: manual modeling is labor intensive and requires expert knowledge. Deep generative models may be used to augment data and annotation, but they typically demand large amounts of training pairs in the first place, which is impractical in data limited clinical settings. In this work, we propose Tumor Fabrication (TF), a novel two-stage framework for unpaired 3D brain tumor synthesis. The framework comprises a coarse tumor synthesis process followed by a refinement process powered by a generative model. TF is fully automated and leverages only healthy image scans along with a limited amount of real annotated data to synthesize large volumes of paired synthetic data for enriching downstream supervised segmentation training. We demonstrate that our synthetic image-label pairs used as data enrichment can significantly improve performance on downstream tumor segmentation tasks in low-data regimes, offering a scalable and reliable solution for medical image enrichment and addressing critical challenges in data scarcity for clinical AI applications. --- ### FHE-Agent: Automating CKKS Configuration for Practical Encrypted Inference via an LLM-Guided Agentic Framework Authors: Nuo Xu, Zhaoting Gong, Ran Ran, Jinwei Tang, Wujie Wen, Caiwen Ding Published: 2025-11-23 Categories: cs.CR, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.18653v1 Fully Homomorphic Encryption (FHE), particularly the CKKS scheme, is a promising enabler for privacy-preserving MLaaS, but its practical deployment faces a prohibitive barrier: it heavily relies on domain expertise. Configuring CKKS involves a tightly coupled space of ring dimensions, modulus chains, and packing layouts. Without deep cryptographic knowledge to navigate these interactions, practitioners are restricted to compilers that rely on fixed heuristics. These "one-shot" tools often emit rigid configurations that are either severely over-provisioned in latency or fail to find a feasible solution entirely for deeper networks. We present FHE-Agent, an agentic framework that automates this expert reasoning process. By coupling a Large Language Model (LLM) controller with a deterministic tool suite, FHE-Agent decomposes the search into global parameter selection and layer-wise bottleneck repair. The agents operate within a multi-fidelity workflow, pruning invalid regimes using cheap static analysis and reserving expensive encrypted evaluations for the most promising candidates. We instantiate FHE-Agent on the Orion compiler and evaluate it on standard benchmarks (MLP, LeNet, LoLa) and deeper architectures (AlexNet). FHE-Agent consistently achieves better precision and lower latency than naïve search strategies. Crucially, it automatically discovers feasible, 128-bit secure configurations for complex models where baseline heuristics and one-shot prompts fail to produce a valid setup. --- ### Lean 5.0: A Predictive, Human-AI, and Ethically Grounded Paradigm for Construction Management Authors: Atena Khoshkonesh, Mohsen Mohammadagha, Navid Ebrahimi, Narges Sadeghigolshan Published: 2025-11-23 Categories: cs.CE, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.18651v1 This paper introduces Lean 5.0, a human-centric evolution of Lean-Digital integration that connects predictive analytics, AI collaboration, and continuous learning within Industry 5.0 and Construction 5.0 contexts. A systematic literature review (2019-2024) and a 12-week empirical validation study demonstrate measurable performance gains, including a 13% increase in Plan Percent Complete (PPC), 22% reduction in rework, and 42% improvement in forecast accuracy. The study adopts a mixed-method Design Science Research (DSR) approach aligned with PRISMA 2020 guidelines. The paper also examines integration with digital twin and blockchain technologies to improve traceability, auditability, and lifecycle transparency. Despite limitations related to sample size, single-case design, and study duration, the findings show that Lean 5.0 provides a transformative paradigm connecting human cognition with predictive control in construction management. --- ### Evaluating Large Language Models on the 2026 Korean CSAT Mathematics Exam: Measuring Mathematical Ability in a Zero-Data-Leakage Setting Authors: Goun Pyeon, Inbum Heo, Jeesu Jung, Taewook Hwang, Hyuk Namgoong, Hyein Seo, Yerim Han, Eunbin Kim, Hyeonseok Kang, Sangkeun Jung Published: 2025-11-23 Categories: cs.CL arXiv: https://arxiv.org/abs/2511.18649v2 This study systematically evaluated the mathematical reasoning capabilities of Large Language Models (LLMs) using the 2026 Korean College Scholastic Ability Test (CSAT) Mathematics section, ensuring a completely contamination-free evaluation environment. To address data leakage issues in existing benchmarks, we digitized all 46 questions (22 common and 24 elective) within two hours of the exam's public release, eliminating any possibility of inclusion in model training data. We conducted comprehensive evaluations of 24 state-of-the-art LLMs across varying input modalities (Text-only, Image-only, Text+Figure) and prompt languages (Korean, English). The GPT-5 family models achieved perfect scores (100 points) under a limited set of language-modality configurations, while Grok 4, Qwen 3 235B, and Gemini 2.5 pro also scored above 97 points. Notably, gpt-oss-20B achieved 95.7 points despite its relatively small size, demonstrating high cost-effectiveness. Problem-specific analysis revealed Calculus as the weakest domain with significant performance degradation on 4-point high-difficulty problems. Text input consistently outperformed image input, while prompt language effects varied by model scale. In reasoning enhancement experiments with GPT-5 series, increased reasoning intensity improved performance (82.6->100 points) but quadrupled token usage and drastically reduced efficiency, suggesting that models with minimal reasoning may be more practical. This research contributes: (1) implementation of a completely unexposed evaluation environment, (2) a standardized digitization pipeline that converts human-targeted exam materials into LLM-ready evaluation data, and (3) a practical evaluation perspective integrating performance, cost, and time considerations. Detailed results and model comparisons are available at the 2026 Korean CSAT LLM Evaluation Leaderboard; https://isoft.cnu.ac.kr/csat2026/ --- ### Kitty: Accurate and Efficient 2-bit KV Cache Quantization with Dynamic Channel-wise Precision Boost Authors: Haojun Xia, Xiaoxia Wu, Jisen Li, Robert Wu, Junxiong Wang, Jue Wang, Chenxi Li, Aman Singhal, Alay Dilipbhai Shah, Alpay Ariyak, Donglin Zhuang, Zhongzhu Zhou et al. Published: 2025-11-23 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.18643v1 The KV cache is a dominant memory bottleneck for LLM inference. While 4-bit KV quantization preserves accuracy, 2-bit often degrades it, especially on long-context reasoning. We close this gap via an algorithm-system co-design for mixed-precision KV caching: Kitty. On the algorithm side, extensive experiments show that Dynamic Channel-wise Precision Boost -- which ranks Key-cache channels by sensitivity and keeps only a small fraction at higher precision -- maintains near-zero loss in accuracy drop while approaching 2-bit memory. The main challenge is handling dynamic 4-bit channel boosts while keeping the page layout coalesced and the dequantization uniform, with no scattered reads or hard-coded masks. Kitty addresses these issues by decompose each mixed-precision Key page into two tensors with unified 2-bit precision. Based on this, Kitty provides a page-centric KV layout, Triton-compatible page dequantization kernels, and a lightweight runtime pipeline that preserves coalescing and avoids divergence. Across seven tasks and two model families (Qwen3, LLaMA3), Kitty cuts KV memory by nearly 8x with negligible accuracy loss, enabling up to 8x larger batches and 2.1x-4.1x higher throughput under the same memory budget. We release the full implementation of Kitty at https://github.com/Summer-Summer/Kitty. --- ### Health system learning achieves generalist neuroimaging models Authors: Akhil Kondepudi, Akshay Rao, Chenhui Zhao, Yiwei Lyu, Samir Harake, Soumyanil Banerjee, Rushikesh Joshi, Anna-Katharina Meissner, Renly Hou, Cheng Jiang, Asadur Chowdury, Ashok Srinivasan et al. Published: 2025-11-23 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.18640v1 Frontier artificial intelligence (AI) models, such as OpenAI's GPT-5 and Meta's DINOv3, have advanced rapidly through training on internet-scale public data, yet such systems lack access to private clinical data. Neuroimaging, in particular, is underrepresented in the public domain due to identifiable facial features within MRI and CT scans, fundamentally restricting model performance in clinical medicine. Here, we show that frontier models underperform on neuroimaging tasks and that learning directly from uncurated data generated during routine clinical care at health systems, a paradigm we call health system learning, yields high-performance, generalist neuroimaging models. We introduce NeuroVFM, a visual foundation model trained on 5.24 million clinical MRI and CT volumes using a scalable volumetric joint-embedding predictive architecture. NeuroVFM learns comprehensive representations of brain anatomy and pathology, achieving state-of-the-art performance across multiple clinical tasks, including radiologic diagnosis and report generation. The model exhibits emergent neuroanatomic understanding and interpretable visual grounding of diagnostic findings. When paired with open-source language models through lightweight visual instruction tuning, NeuroVFM generates radiology reports that surpass frontier models in accuracy, clinical triage, and expert preference. Through clinically grounded visual understanding, NeuroVFM reduces hallucinated findings and critical errors, offering safer clinical decision support. These results establish health system learning as a paradigm for building generalist medical AI and provide a scalable framework for clinical foundation models. --- ### No Free Lunch in Language Model Bias Mitigation? Targeted Bias Reduction Can Exacerbate Unmitigated LLM Biases Authors: Shireen Chand, Faith Baca, Emilio Ferrara Published: 2025-11-23 Categories: cs.CL, cs.AI, cs.CY arXiv: https://arxiv.org/abs/2511.18635v1 Large Language Models (LLMs) inherit societal biases from their training data, potentially leading to harmful or unfair outputs. While various techniques aim to mitigate these biases, their effects are often evaluated only along the dimension of the bias being targeted. This work investigates the cross-category consequences of targeted bias mitigation. We study four bias mitigation techniques applied across ten models from seven model families, and we explore racial, religious, profession- and gender-related biases. We measure the impact of debiasing on model coherence and stereotypical preference using the StereoSet benchmark. Our results consistently show that while targeted mitigation can sometimes reduce bias in the intended dimension, it frequently leads to unintended and often negative consequences in others, such as increasing model bias and decreasing general coherence. These findings underscore the critical need for robust, multi-dimensional evaluation tools when examining and developing bias mitigation strategies to avoid inadvertently shifting or worsening bias along untargeted axes. --- ### Bridging Philosophy and Machine Learning: A Structuralist Framework for Classifying Neural Network Representations Authors: Yildiz Culcu Published: 2025-11-23 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.18633v1 Machine learning models increasingly function as representational systems, yet the philosoph- ical assumptions underlying their internal structures remain largely unexamined. This paper develops a structuralist decision framework for classifying the implicit ontological commitments made in machine learning research on neural network representations. Using a modified PRISMA protocol, a systematic review of the last two decades of literature on representation learning and interpretability is conducted. Five influential papers are analysed through three hierarchical criteria derived from structuralist philosophy of science: entity elimination, source of structure, and mode of existence. The results reveal a pronounced tendency toward structural idealism, where learned representations are treated as model-dependent constructions shaped by architec- ture, data priors, and training dynamics. Eliminative and non-eliminative structuralist stances appear selectively, while structural realism is notably absent. The proposed framework clarifies conceptual tensions in debates on interpretability, emergence, and epistemic trust in machine learning, and offers a rigorous foundation for future interdisciplinary work between philosophy of science and machine learning. --- ### The Locally Deployable Virtual Doctor: LLM Based Human Interface for Automated Anamnesis and Database Conversion Authors: Jan Benedikt Ruhland, Doguhan Bahcivan, Jan-Peter Sowa, Ali Canbay, Dominik Heider Published: 2025-11-23 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.18632v1 Recent advances in large language models made it possible to achieve high conversational performance with substantially reduced computational demands, enabling practical on-site deployment in clinical environments. Such progress allows for local integration of AI systems that uphold strict data protection and patient privacy requirements, yet their secure implementation in medicine necessitates careful consideration of ethical, regulatory, and technical constraints. In this study, we introduce MedChat, a locally deployable virtual physician framework that integrates an LLM-based medical chatbot with a diffusion-driven avatar for automated and structured anamnesis. The chatbot was fine-tuned using a hybrid corpus of real and synthetically generated medical dialogues, while model efficiency was optimized via Low-Rank Adaptation. A secure and isolated database interface was implemented to ensure complete separation between patient data and the inference process. The avatar component was realized through a conditional diffusion model operating in latent space, trained on researcher video datasets and synchronized with mel-frequency audio features for realistic speech and facial animation. Unlike existing cloud-based systems, this work demonstrates the feasibility of a fully offline, locally deployable LLM-diffusion framework for clinical anamnesis. The autoencoder and diffusion networks exhibited smooth convergence, and MedChat achieved stable fine-tuning with strong generalization to unseen data. The proposed system thus provides a privacy-preserving, resource-efficient foundation for AI-assisted clinical anamnesis, also in low-cost settings. --- ### FOS: A Large-Scale Temporal Graph Benchmark for Scientific Interdisciplinary Link Prediction Authors: Kiyan Rezaee, Morteza Ziabakhsh, Niloofar Nikfarjam, Mohammad M. Ghassemi, Yazdan Rezaee Jouryabi, Sadegh Eskandari, Reza Lashgari Published: 2025-11-23 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.18631v1 Interdisciplinary scientific breakthroughs mostly emerge unexpectedly, and forecasting the formation of novel research fields remains a major challenge. We introduce FOS (Future Of Science), a comprehensive time-aware graph-based benchmark that reconstructs annual co-occurrence graphs of 65,027 research sub-fields (spanning 19 general domains) over the period 1827-2024. In these graphs, edges denote the co-occurrence of two fields in a single publication and are timestamped with the corresponding publication year. Nodes are enriched with semantic embeddings, and edges are characterized by temporal and topological descriptors. We formulate the prediction of new field-pair linkages as a temporal link-prediction task, emphasizing the "first-time" connections that signify pioneering interdisciplinary directions. Through extensive experiments, we evaluate a suite of state-of-the-art temporal graph architectures under multiple negative-sampling regimes and show that (i) embedding long-form textual descriptions of fields significantly boosts prediction accuracy, and (ii) distinct model classes excel under different evaluation settings. Case analyses show that top-ranked link predictions on FOS align with field pairings that emerge in subsequent years of academic publications. We publicly release FOS, along with its temporal data splits and evaluation code, to establish a reproducible benchmark for advancing research in predicting scientific frontiers. --- ### Majority of the Bests: Improving Best-of-N via Bootstrapping Authors: Amin Rakhsha, Kanika Madan, Tianyu Zhang, Amir-massoud Farahmand, Amir Khasahmadi Published: 2025-11-23 Categories: cs.LG, cs.AI, cs.CL, stat.ME, stat.ML arXiv: https://arxiv.org/abs/2511.18630v1 Sampling multiple outputs from a Large Language Model (LLM) and selecting the most frequent (Self-consistency) or highest-scoring (Best-of-N) candidate is a popular approach to achieve higher accuracy in tasks with discrete final answers. Best-of-N (BoN) selects the output with the highest reward, and with perfect rewards, it often achieves near-perfect accuracy. With imperfect rewards from reward models, however, BoN fails to reliably find the correct answer and its performance degrades drastically. We consider the distribution of BoN's outputs and highlight that, although the correct answer does not usually have a probability close to one under imperfect rewards, it is often the most likely outcome. This suggests that the mode of this distribution can be more reliably correct than a sample from it. Based on this idea, we propose Majority-of-the-Bests (MoB), a novel selection mechanism that estimates the output distribution of BoN via bootstrapping and selects its mode. Experimental results across five benchmarks, three different base LLMs, and two reward models demonstrate consistent improvements over BoN in 25 out of 30 setups. We also provide theoretical results for the consistency of the bootstrapping. MoB serves as a simple, yet strong alternative to BoN and self-consistency, and more broadly, motivates further research in more nuanced selection mechanisms. --- ### Functional Localization Enforced Deep Anomaly Detection Using Fundus Images Authors: Jan Benedikt Ruhland, Thorsten Papenbrock, Jan-Peter Sowa, Ali Canbay, Nicole Eter, Bernd Freisleben, Dominik Heider Published: 2025-11-23 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2511.18627v1 Reliable detection of retinal diseases from fundus images is challenged by the variability in imaging quality, subtle early-stage manifestations, and domain shift across datasets. In this study, we systematically evaluated a Vision Transformer (ViT) classifier under multiple augmentation and enhancement strategies across several heterogeneous public datasets, as well as the AEyeDB dataset, a high-quality fundus dataset created in-house and made available for the research community. The ViT demonstrated consistently strong performance, with accuracies ranging from 0.789 to 0.843 across datasets and diseases. Diabetic retinopathy and age-related macular degeneration were detected reliably, whereas glaucoma remained the most frequently misclassified disease. Geometric and color augmentations provided the most stable improvements, while histogram equalization benefited datasets dominated by structural subtlety. Laplacian enhancement reduced performance across different settings. On the Papila dataset, the ViT with geometric augmentation achieved an AUC of 0.91, outperforming previously reported convolutional ensemble baselines (AUC of 0.87), underscoring the advantages of transformer architectures and multi-dataset training. To complement the classifier, we developed a GANomaly-based anomaly detector, achieving an AUC of 0.76 while providing inherent reconstruction-based explainability and robust generalization to unseen data. Probabilistic calibration using GUESS enabled threshold-independent decision support for future clinical implementation. --- ### OpenGloss: A Synthetic Encyclopedic Dictionary and Semantic Knowledge Graph Authors: Michael J. Bommarito Published: 2025-11-23 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2511.18622v1 We present OpenGloss, a synthetic encyclopedic dictionary and semantic knowledge graph for English that integrates lexicographic definitions, encyclopedic context, etymological histories, and semantic relationships in a unified resource. OpenGloss contains 537K senses across 150K lexemes, on par with WordNet 3.1 and Open English WordNet, while providing more than four times as many sense definitions. These lexemes include 9.1M semantic edges, 1M usage examples, 3M collocations, and 60M words of encyclopedic content. Generated through a multi-agent procedural generation pipeline with schema-validated LLM outputs and automated quality assurance, the entire resource was produced in under one week for under $1,000. This demonstrates that structured generation can create comprehensive lexical resources at cost and time scales impractical for manual curation, enabling rapid iteration as foundation models improve. The resource addresses gaps in pedagogical applications by providing integrated content -- definitions, examples, collocations, encyclopedias, etymology -- that supports both vocabulary learning and natural language processing tasks. As a synthetically generated resource, OpenGloss reflects both the capabilities and limitations of current foundation models. The dataset is publicly available on Hugging Face under CC-BY 4.0, enabling researchers and educators to build upon and adapt this resource. --- ### Prompt Optimization as a State-Space Search Problem Authors: Maanas Taneja Published: 2025-11-23 Categories: cs.CL arXiv: https://arxiv.org/abs/2511.18619v1 Language Models are extremely susceptible to performance collapse with even small changes to input prompt strings. Libraries such as DSpy (from Stanford NLP) avoid this problem through demonstration-based prompt optimisation. Inspired by this, I propose an alternative approach that treats prompt optimisation as a classical state-space search problem. I model the prompt space as a graph where nodes represent prompt states and edges correspond to deliberate transformations such as shortening, adding examples, or re- ordering content. Using beam search and random walk algorithms, I systematically explore this space, evaluating candidates on development sets and pruning unpromising branches. Across five NLP tasks (sentiment classification, question answering, summarisation, reason- ing, and natural language inference), I find that even shallow search configurations (beam width=2, depth=2) improve upon seed prompts on development sets. For instance, beam search achieves development accuracy gains from 0.40 to 0.80 on reasoning tasks, though test set improvements are more modest (0.20 to 0.50), indicating overfitting to the develop- ment heuristic. Analysis of successful optimisation paths reveals that transformations that make prompts concise appear most frequently, while verbosity operators are never selected. My results validate prompt optimization as a search problem and suggest that with greater computational resources and improved evaluation metrics, deeper exploration could yield more robust prompts that generalize beyond development sets. Code and implementation are available at (https://github.com/MaanasTaneja/PromptOptimiser). --- ### A Unified BERT-CNN-BiLSTM Framework for Simultaneous Headline Classification and Sentiment Analysis of Bangla News Authors: Mirza Raquib, Munazer Montasir Akash, Tawhid Ahmed, Saydul Akbar Murad, Farida Siddiqi Prity, Mohammad Amzad Hossain, Asif Pervez Polok, Nick Rahimi Published: 2025-11-23 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2511.18618v2 In our daily lives, newspapers are an essential information source that impacts how the public talks about present-day issues. However, effectively navigating the vast amount of news content from different newspapers and online news portals can be challenging. Newspaper headlines with sentiment analysis tell us what the news is about (e.g., politics, sports) and how the news makes us feel (positive, negative, neutral). This helps us quickly understand the emotional tone of the news. This research presents a state-of-the-art approach to Bangla news headline classification combined with sentiment analysis applying Natural Language Processing (NLP) techniques, particularly the hybrid transfer learning model BERT-CNN-BiLSTM. We have explored a dataset called BAN-ABSA of 9014 news headlines, which is the first time that has been experimented with simultaneously in the headline and sentiment categorization in Bengali newspapers. Over this imbalanced dataset, we applied two experimental strategies: technique-1, where undersampling and oversampling are applied before splitting, and technique-2, where undersampling and oversampling are applied after splitting on the In technique-1 oversampling provided the strongest performance, both headline and sentiment, that is 78.57\% and 73.43\% respectively, while technique-2 delivered the highest result when trained directly on the original imbalanced dataset, both headline and sentiment, that is 81.37\% and 64.46\% respectively. The proposed model BERT-CNN-BiLSTM significantly outperforms all baseline models in classification tasks, and achieves new state-of-the-art results for Bangla news headline classification and sentiment analysis. These results demonstrate the importance of leveraging both the headline and sentiment datasets, and provide a strong baseline for Bangla text classification in low-resource. --- ### AutoFocus-IL: VLM-based Saliency Maps for Data-Efficient Visual Imitation Learning without Extra Human Annotations Authors: Litian Gong, Fatemeh Bahrani, Yutai Zhou, Amin Banayeeanzade, Jiachen Li, Erdem Bıyık Published: 2025-11-23 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2511.18617v2 AutoFocus-IL is a simple yet effective method to improve data efficiency and generalization in visual imitation learning by guiding policies to attend to task-relevant features rather than distractors and spurious correlations. Although saliency regularization has emerged as a promising way to achieve this, existing approaches typically require costly supervision such as human gaze data or manual saliency annotations. In contrast, AutoFocus-IL leverages vision-language models (VLMs) to automatically identify and track key objects in demonstrations, generating temporal saliency maps that highlight causal visual signals while suppressing distractors. These maps are then used to regularize behavior cloning policies, yielding stronger alignment between visual attention and task-relevant cues. Experiments in both the CARLA simulator and real-robot manipulation tasks demonstrate that AutoFocus-IL not only outperforms standard behavior cloning but also surpasses state-of-the-art baselines that assume privileged access to human supervision, such as gaze data. Code, datasets, and trained policy videos are available at https://AutoFocus-IL.github.io/. --- ### What Helps Language Models Predict Human Beliefs: Demographics or Prior Stances? Authors: Joseph Malone, Rachith Aiyappa, Byunghwee Lee, Haewoon Kwak, Jisun An, Yong-Yeol Ahn Published: 2025-11-23 Categories: cs.CL arXiv: https://arxiv.org/abs/2511.18616v2 Beliefs shape how people reason, communicate, and behave. Rather than existing in isolation, they exhibit a rich correlational structure--some connected through logical dependencies, others through indirect associations or social processes. As usage of large language models (LLMs) becomes more ubiquitous in our society, LLMs' ability to understand and reason through human beliefs has many implications from privacy issues to personalized persuasion and the potential for stereotyping. Yet how LLMs capture this interrelated landscape of beliefs remains unclear. For instance, when predicting someone's beliefs, what information affects the prediction most--who they are (demographics), what else they believe (prior stances), or a combination of both? We address these questions using data from an online debate platform, evaluating the ability of off-the-shelf open-weight LLMs to predict individuals' stance under four conditions: no context, demographics only, prior beliefs only, and both combined. We find that both types of information improve predictions over a blind baseline, with their combination yielding the best performance in most cases. However, the relative value of each varies substantially across belief domains. These findings reveal how current LLMs leverage different types of social information when reasoning about human beliefs, highlighting both their capabilities and limitations. --- ### Bayesian-based Online Label Shift Estimation with Dynamic Dirichlet Priors Authors: Jiawei Hu, Javier A. Barria Published: 2025-11-23 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2511.18615v1 Label shift, a prevalent challenge in supervised learning, arises when the class prior distribution of test data differs from that of training data, leading to significant degradation in classifier performance. To accurately estimate the test priors and enhance classification accuracy, we propose a Bayesian framework for label shift estimation, termed Full Maximum A Posterior Label Shift (FMAPLS), along with its online version, online-FMAPLS. Leveraging batch and online Expectation-Maximization (EM) algorithms, these methods jointly and dynamically optimize Dirichlet hyperparameters $\boldsymbolα$ and class priors $\boldsymbolπ$, thereby overcoming the rigid constraints of the existing Maximum A Posterior Label Shift (MAPLS) approach. Moreover, we introduce a linear surrogate function (LSF) to replace gradient-based hyperparameter updates, yielding closed-form solutions that reduce computational complexity while retaining asymptotic equivalence. The online variant substitutes the batch E-step with a stochastic approximation, enabling real-time adaptation to streaming data. Furthermore, our theoretical analysis reveals a fundamental trade-off between online convergence rate and estimation accuracy. Extensive experiments on CIFAR100 and ImageNet datasets under shuffled long-tail and Dirichlet test priors demonstrate that FMAPLS and online-FMAPLS respectively achieve up to 40% and 12% lower KL divergence and substantial improvements in post-shift accuracy over state-of-the-art baselines, particularly under severe class imbalance and distributional uncertainty. These results confirm the robustness, scalability, and suitability of the proposed methods for large-scale and dynamic learning scenarios. --- ### KAN vs LSTM Performance in Time Series Forecasting Authors: Tabish Ali Rather, S M Mahmudul Hasan Joy, Nadezda Sukhorukova, Federico Frascoli Published: 2025-11-23 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.18613v2 This study presents a controlled comparison of baseline Kolmogorov-Arnold Networks (KAN), implemented via PyKAN, and Long Short-Term Memory (LSTM) networks for the forecasting of stochastic, non-stationary financial time series. The two architectures are assessed in terms of predictive accuracy, computational efficiency, and interpretability, with accuracy measured by the Root Mean Square Error (RMSE) in normalised feature space. Under a direct multi-output forecasting protocol, LSTM attains clearly superior accuracy across all tested prediction horizons, consistent with its well-established effectiveness for sequential data modelling. Baseline KAN, although offering theoretical interpretability through the Kolmogorov-Arnold representation theorem, exhibits substantially higher error rates and limited practical applicability for time series forecasting in its standard form. Several specialised temporal variants -- including Temporal KAN and Time-Frequency KAN -- have since been proposed to address these sequential modelling limitations, but they lie outside the scope of the present study. KAN is observed to converge faster during training under the configurations tested, although direct runtime comparisons are constrained by methodological factors. These findings support the adoption of LSTM for accuracy-critical financial forecasting and establish an empirical baseline for standard KAN on stochastic sequential data, motivating further investigation of temporally-aware KAN architectures. The study benchmarks baseline KAN against baseline LSTM only; the results do not extend to specialised KAN variants designed for sequential data, nor to the broader family of temporal models. ---