# Research Papers — 2025-11 · bundle page 30 of 197 > Full-text markdown of papers #1451–1500 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/30 [← Previous bundle](https://aipod.dev/papers/archive/2025-11/page/29.md) [Next bundle →](https://aipod.dev/papers/archive/2025-11/page/31.md) ### Dynamic Stratified Contrastive Learning with Upstream Augmentation for MILP Branching Authors: Tongkai Lu, Shuai Ma, Chongyang Tao Published: 2025-11-26 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.21107v1 Mixed Integer Linear Programming (MILP) is a fundamental class of NP-hard problems that has garnered significant attention from both academia and industry. The Branch-and-Bound (B\&B) method is the dominant approach for solving MILPs and the branching plays an important role in B\&B methods. Neural-based learning frameworks have recently been developed to enhance branching policies and the efficiency of solving MILPs. However, these methods still struggle with semantic variation across depths, the scarcity of upstream nodes, and the costly collection of strong branching samples. To address these issues, we propose \ours, a Dynamic \underline{\textbf{S}}tratified \underline{\textbf{C}}ontrastive Training Framework for \underline{\textbf{MILP}} Branching. It groups branch-and-bound nodes based on their feature distributions and trains a GCNN-based discriminative model to progressively separate nodes across groups, learning finer-grained node representations throughout the tree. To address data scarcity and imbalance at upstream nodes, we introduce an upstream-augmented MILP derivation procedure that generates both theoretically equivalent and perturbed instances. \ours~effectively models subtle semantic differences between nodes, significantly enhancing branching accuracy and solving efficiency, particularly for upstream nodes. Extensive experiments on standard MILP benchmarks demonstrate that our method enhances branching accuracy, reduces solving time, and generalizes effectively to unseen instances. --- ### EM-KD: Distilling Efficient Multimodal Large Language Model with Unbalanced Vision Tokens Authors: Ze Feng, Sen Yang, Boqiang Duan, Wankou Yang, Jingdong Wang Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21106v1 Efficient Multimodal Large Language Models (MLLMs) compress vision tokens to reduce resource consumption, but the loss of visual information can degrade comprehension capabilities. Although some priors introduce Knowledge Distillation to enhance student models, they overlook the fundamental differences in fine-grained vision comprehension caused by unbalanced vision tokens between the efficient student and vanilla teacher. In this paper, we propose EM-KD, a novel paradigm that enhances the Efficient MLLMs with Knowledge Distillation. To overcome the challenge of unbalanced vision tokens, we first calculate the Manhattan distance between the vision logits of teacher and student, and then align them in the spatial dimension with the Hungarian matching algorithm. After alignment, EM-KD introduces two distillation strategies: 1) Vision-Language Affinity Distillation (VLAD) and 2) Vision Semantic Distillation (VSD). Specifically, VLAD calculates the affinity matrix between text tokens and aligned vision tokens, and minimizes the smooth L1 distance of the student and the teacher affinity matrices. Considering the semantic richness of vision logits in the final layer, VSD employs the reverse KL divergence to measure the discrete probability distributions of the aligned vision logits over the vocabulary space. Comprehensive evaluation on diverse benchmarks demonstrates that EM-KD trained model outperforms prior Efficient MLLMs on both accuracy and efficiency with a large margin, validating its effectiveness. Compared with previous distillation methods, which are equipped with our proposed vision token matching strategy for fair comparison, EM-KD also achieves better performance. --- ### RLM: A Vision-Language Model Approach for Radar Scene Understanding Authors: Pushkal Mishra, Kshitiz Bansal, Dinesh Bharadia Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21105v4 Radar sensors provide reliable perception across adverse weather, lighting, and long-range conditions, yet existing machine learning approaches remain fragmented and task-specific, with each downstream task employing distinct architectures and training objectives. We present RadarVLM, a vision-language framework that learns unified scene-level representations through structured spatial language supervision. Leveraging the CARLA simulator with a realistic radar model, we collect over 800k radar-caption pairs across 110+ hours of simulated driving in diverse scenarios. We make two key contributions: (1) a structured caption framework encoding vehicle distributions in the radar's native coordinate system, and (2) Spatially-Grounded CLIP (SG-CLIP) objective that replaces binary matching with continuous scene similarity, enabling fine-grained spatial reasoning. We further propose localization-aware evaluation metrics that directly assess spatial accuracy beyond traditional linguistic similarity measures. Validated on generative captioning and vehicle segmentation, SG-CLIP achieves up to 50% relative F1-score improvement over vanilla CLIP and a 21% AP gain on segmentation, demonstrating that language grounding produces spatially structured representations. --- ### BRIDGE: Building Representations In Domain Guided Program Synthesis Authors: Robert Joseph George, Carson Eisenach, Udaya Ghai, Dominique Perrault-Joncas, Anima Anandkumar, Dean Foster Published: 2025-11-26 Categories: cs.LG, cs.PL arXiv: https://arxiv.org/abs/2511.21104v4 Large language models can generate plausible code, but remain brittle for formal verification in proof assistants such as Lean. A central scalability challenge is that verified synthesis requires consistent artifacts across several coupled domains: executable code, formal specifications, theorem statements, and proof attempts. Existing approaches often treat these artifacts separately. We present BRIDGE, a structured prompting framework for multi-artifact program synthesis. BRIDGE decomposes generation into three interconnected domains: Code, Specification, and Theorem/Proof, and uses domain-specific intermediate reasoning to connect them. In Lean, BRIDGE often follows a code-first workflow, using the generated implementation as a semantic anchor for downstream specification, theorem statement, and proof-attempt generation. Across 178 algorithmic problems and five LLMs, BRIDGE improves Lean executable correctness by up to nearly 1.5x over direct prompting and can be roughly 2x more sample efficient at comparable generation lengths. We further find that specification-oriented prompting improves Python pass rates by up to 17.5 percentage points. Beyond inference-time prompting, supervised fine-tuning on BRIDGE-style reasoning traces yields nearly 1.5x higher Lean pass success than code-only fine-tuning, suggesting that these intermediate representations provide a learnable inductive bias. BRIDGE provides a practical framework for scaling verified synthesis while highlighting the remaining gap between executable correctness and full formal proof generation. --- ### From Bits to Rounds: Parallel Decoding with Exploration for Diffusion Language Models Authors: Hengyu Fu, Baihe Huang, Virginia Adams, Charles Wang, Venkat Srinivasan, Jiantao Jiao Published: 2025-11-26 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.21103v1 Diffusion Language Models (DLMs) have recently emerged as a strong alternative to autoregressive language models (LMs). DLMs offer comparable accuracy with faster inference speed via parallel decoding. However, standard DLM decoding strategies relying on high-confidence tokens encounter an inherent information-theoretic bottleneck that restricts decoding progress and ultimately slows generation. We demonstrate both theoretically and empirically that prioritizing high-confidence tokens is inherently inefficient. High-probability tokens carry negligible information and strictly relying on them limits the effective progress made in each decoding round. We prove that the number of decoding rounds must grow linearly with the sample's total information (negative log-likelihood) and inversely with the per-round information budget, establishing a bits-to-rounds principle. We also propose Explore-Then-Exploit (ETE), a training-free decoding strategy that maximizes information throughput and decoding efficiency. ETE combines cross-block decoding with targeted exploration of high-uncertainty tokens to reshape the conditional distribution and trigger cascades of confident predictions. Experiments verify our theoretical bounds and demonstrate that ETE consistently reduces the required number of decoding rounds compared to confidence-only baselines without compromising generation quality. --- ### Mortgage Language Model: Domain-Adaptive Pretraining with Residual Instruction, Alignment Tuning, and Task-Specific Routing Authors: Manish Jain, Satheesh Kumar Ponnambalam, Salman Faroz, Chandrakanth Lns, Vinay Sharma Published: 2025-11-26 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2511.21101v2 Large Language Models (LLMs) demonstrate exceptional capabilities across general domains, yet their application to specialized sectors such as mortgage finance requires domain-specific knowledge augmentation while preserving instruction-following fidelity. We present MortgageLLM, a novel domain-specific large language model that addresses this dual challenge. It is developed using a dual-track specialization framework from a single base model (LLaMA-3.1-8B). We opted for this dual-expert approach as a single multi-task model suffers from performance trade-offs, where optimizing for structured tasks (via SFT) degrades conversational fidelity (via DPO). Our dual-track method solves this by creating two specialists, allowing each to be optimally trained for its distinct capability. Our approach applies the instruction residual technique to restore instruction-following capabilities post-domain adaptation without supervised fine-tuning. We contribute: (1) application of this residual technique to the highly specialized mortgage finance domain; (2) a dual-expert architecture combining a conversational Q&A model and a structured task model for classification and summarization; and (3) an intelligent task routing mechanism using few-shot classification performed by one of the expert models itself. We validate our approach on domain-specific benchmarks, where our final model (MLM v2) significantly outperforms the base LLaMA-3.1-8B-Instruct, achieving an LLM-as-a-Judge summarization score of 4.58 (vs. 3.99), a Q&A score of 4.09 (vs. 4.0), and a classification score of 2.6 (vs. 1.2). On semantic similarity, our model achieved a BERTScore of 0.77 for summarization (vs. 0.74), 0.68 for Q&A (vs. 0.58), and 0.75 for classification (vs. 0.73), substantially outperforming baseline approaches. --- ### Pygmalion Effect in Vision: Image-to-Clay Translation for Reflective Geometry Reconstruction Authors: Gayoung Lee, Junho Kim, Jin-Hwa Kim, Junmo Kim Published: 2025-11-26 Categories: cs.CV, cs.AI, cs.GR arXiv: https://arxiv.org/abs/2511.21098v1 Understanding reflection remains a long-standing challenge in 3D reconstruction due to the entanglement of appearance and geometry under view-dependent reflections. In this work, we present the Pygmalion Effect in Vision, a novel framework that metaphorically "sculpts" reflective objects into clay-like forms through image-to-clay translation. Inspired by the myth of Pygmalion, our method learns to suppress specular cues while preserving intrinsic geometric consistency, enabling robust reconstruction from multi-view images containing complex reflections. Specifically, we introduce a dual-branch network in which a BRDF-based reflective branch is complemented by a clay-guided branch that stabilizes geometry and refines surface normals. The two branches are trained jointly using the synthesized clay-like images, which provide a neutral, reflection-free supervision signal that complements the reflective views. Experiments on both synthetic and real datasets demonstrate substantial improvement in normal accuracy and mesh completeness over existing reflection-handling methods. Beyond technical gains, our framework reveals that seeing by unshining, translating radiance into neutrality, can serve as a powerful inductive bias for reflective object geometry learning. --- ### CLRecogEye : Curriculum Learning towards exploiting convolution features for Dynamic Iris Recognition Authors: Geetanjali Sharma, Gaurav Jaswal, Aditya Nigam, Raghavendra Ramachandra Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21097v1 Iris authentication algorithms have achieved impressive recognition performance, making them highly promising for real-world applications such as border control, citizen identification, and both criminal investigations and commercial systems. However, their robustness is still challenged by variations in rotation, scale, specular reflections, and defocus blur. In addition, most existing approaches rely on straightforward point-to-point comparisons, typically using cosine or L2 distance, without effectively leveraging the spatio-spatial-temporal structure of iris patterns. To address these limitations, we propose a novel and generalized matching pipeline that learns rich spatio-spatial-temporal representations of iris features. Our approach first splits each iris image along one dimension, generating a sequence of sub-images that serve as input to a 3D-CNN, enabling the network to capture both spatial and spatio-spatial-temporal cues. To further enhance the modeling of spatio-spatial-temporal feature dynamics, we train the model in curriculum manner. This design allows the network to embed temporal dependencies directly into the feature space, improving discriminability in the deep metric domain. The framework is trained end-to-end with triplet and ArcFace loss in a curriculum manner, enforcing highly discriminative embeddings despite challenges like rotation, scale, reflections, and blur. This design yields a robust and generalizable solution for iris authentication.Github code: https://github.com/GeetanjaliGTZ/CLRecogEye --- ### Target-Aware Early Stage Ranking Authors: Juhee Hong, Meng Liu, Shengzhi Wang, Jin Zhou, Xiaoheng Mao, Zhao Zhu, Ruochen Liu, Huihui Cheng, Leon Gao, Christopher Leung, Chandra Mouli Sekar, Yijia Liu et al. Published: 2025-11-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.21095v2 Early Stage Ranking (ESR) in large-scale recommendation systems is dominated by ''user--item decoupling'' Two Tower architectures, which scale efficiently but cannot capture fine-grained, target-aware user--item interactions directly. We propose Target-Aware Early Stage Ranking (TESR), which augments the Two Tower with a Mixture of Attention (MoA) module trained as a request-level sequence modeling over user history. MoA combines (i) Hard Matching Attention (HMA) to capture explicit categorical-ID level overlap signals between user history and candidate item, (ii) target-aware HSTU attention for implicit affinities conditioned on the candidate, and (iii) target dependent and independent cross-attention for symmetric user-item contextualization. On top of this, a Multi-Logit Parameterized Gating (MLPG) head amplifies these signals at scoring time. To keep latency within ESR budgets, we co-design the architecture with FP8 quantization, custom kernels, and a Torch Inductor compilation path. On a production deployment, TESR delivers consistent offline NE wins and online topline gains, and is, to our knowledge, the first deployment of full target-aware attention sequence modeling in an ESR stage at this scale. --- ### MNM : Multi-level Neuroimaging Meta-analysis with Hyperbolic Brain-Text Representations Authors: Seunghun Baek, Jaejin Lee, Jaeyoon Sim, Minjae Jeong, Won Hwa Kim Published: 2025-11-26 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.21092v1 Various neuroimaging studies suffer from small sample size problem which often limit their reliability. Meta-analysis addresses this challenge by aggregating findings from different studies to identify consistent patterns of brain activity. However, traditional approaches based on keyword retrieval or linear mappings often overlook the rich hierarchical structure in the brain. In this work, we propose a novel framework that leverages hyperbolic geometry to bridge the gap between neuroscience literature and brain activation maps. By embedding text from research articles and corresponding brain images into a shared hyperbolic space via the Lorentz model, our method captures both semantic similarity and hierarchical organization inherent in neuroimaging data. In the hyperbolic space, our method performs multi-level neuroimaging meta-analysis (MNM) by 1) aligning brain and text embeddings for semantic correspondence, 2) guiding hierarchy between text and brain activations, and 3) preserving the hierarchical relationships within brain activation patterns. Experimental results demonstrate that our model outperforms baselines, offering a robust and interpretable paradigm of multi-level neuroimaging meta-analysis via hyperbolic brain-text representation. --- ### MLPMoE: Zero-Shot Architectural Metamorphosis of Dense LLM MLPs into Static Mixture-of-Experts Authors: Ivan Novikov Published: 2025-11-26 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.21089v1 Large Language Models (LLMs) are predominantly deployed as dense transformers, where every parameter in every feed-forward block is activated for every token. While architecturally simple, this is computationally inefficient, since inference costs scale linearly with parameter count. Recent upcycling methods such as MoEfication, CMoE, ToMoE, and MoORE reveal that much of the useful computation lives in sparse, semi-modular substructures inside dense feed-forward networks, but these approaches typically rely on clustering, activation profiling, singular value decomposition, or custom routing that requires calibration data. This paper introduces MLPMoE (MLP Mixture-of-Experts), a training-free, deterministic transformation that restructures the dense MLP in transformer blocks into a static, high-cardinality mixture of experts. The transformation uses simple tensor slicing and summation, reinterpreting the algebra of tensor parallelism as a topological conversion rather than a distributed training pattern. We further introduce Fractal Fade (differential branch sparsity) and Compensated Pruning (variance-preserving branch reduction) as lightweight mechanisms for structured sparsity. On Qwen2.5-0.5B-Instruct and DeepSeek-R1-Distill-Llama-8B, the zero-shot MLPMoE transform changes a proxy perplexity metric by less than 0.05 percent while keeping the parameter count effectively constant. On the 8B model, differential sparsity removes about 20 percent of MLP parameters while keeping perplexity within about 2 percent of the dense baseline. The method operates entirely post hoc on existing checkpoints and does not require gradients, calibration sets, or router training. Code is available at https://gist.github.com/iwallarm/fc2ef1eddf226ca7814f9e5e2ae9bad1 --- ### ASR Error Correction in Low-Resource Burmese with Alignment-Enhanced Transformers using Phonetic Features Authors: Ye Bhone Lin, Thura Aung, Ye Kyaw Thu, Thazin Myint Oo Published: 2025-11-26 Categories: cs.CL, cs.LG, cs.SD arXiv: https://arxiv.org/abs/2511.21088v1 This paper investigates sequence-to-sequence Transformer models for automatic speech recognition (ASR) error correction in low-resource Burmese, focusing on different feature integration strategies including IPA and alignment information. To our knowledge, this is the first study addressing ASR error correction specifically for Burmese. We evaluate five ASR backbones and show that our ASR Error Correction (AEC) approaches consistently improve word- and character-level accuracy over baseline outputs. The proposed AEC model, combining IPA and alignment features, reduced the average WER of ASR models from 51.56 to 39.82 before augmentation (and 51.56 to 43.59 after augmentation) and improving chrF++ scores from 0.5864 to 0.627, demonstrating consistent gains over the baseline ASR outputs without AEC. Our results highlight the robustness of AEC and the importance of feature design for improving ASR outputs in low-resource settings. --- ### MIRA: Multimodal Iterative Reasoning Agent for Image Editing Authors: Ziyun Zeng, Hang Hua, Jiebo Luo Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21087v2 Instruction-guided image editing offers an intuitive way for users to edit images with natural language. However, diffusion-based editing models often struggle to accurately interpret complex user instructions, especially those involving compositional relationships, contextual cues, or referring expressions, leading to edits that drift semantically or fail to reflect the intended changes. We tackle this problem by proposing MIRA (Multimodal Iterative Reasoning Agent), a lightweight, plug-and-play multimodal reasoning agent that performs editing through an iterative perception-reasoning-action loop, effectively simulating multi-turn human-model interaction processes. Instead of issuing a single prompt or static plan, MIRA predicts atomic edit instructions step by step, using visual feedback to make its decisions. Our 150K multimodal tool-use dataset, MIRA-Editing, combined with a two-stage SFT + GRPO training pipeline, enables MIRA to perform reasoning and editing over complex editing instructions. When paired with open-source image editing models such as Flux.1-Kontext, Step1X-Edit, and Qwen-Image-Edit, MIRA significantly improves both semantic consistency and perceptual quality, achieving performance comparable to or exceeding proprietary systems such as GPT-Image and Nano-Banana. --- ### Orthographic Constraint Satisfaction and Human Difficulty Alignment in Large Language Models Authors: Bryan E. Tuck, Rakesh M. Verma Published: 2025-11-26 Categories: cs.CL arXiv: https://arxiv.org/abs/2511.21086v2 Large language models must satisfy hard orthographic constraints during controlled text generation, yet systematic cross-family evaluation remains limited. We evaluate 39 configurations spanning three model families (Qwen3, Claude Haiku 4.5, GPT-5-mini) on 58 word puzzles requiring character-level constraint satisfaction. Cross-family differences produce substantially larger performance gaps (2.0-2.2x, F1 = 0.761 vs. 0.343) than parameter scaling within families (83% gain from 4B to 32B scaling), and a partial-correlation analysis rules out tokenizer design as a confound for within-family scaling. Thinking budget sensitivity proves heterogeneous: high-capacity models show strong returns (+0.102 to +0.136 F1), while mid-sized variants saturate or degrade, showing inconsistent compute benefits. Using difficulty ratings from 10,000 human solvers per puzzle, we establish modest but consistent calibration (\r{ho} = 0.28-0.42) across all families, yet identify systematic failures on common words with unusual orthography ("data", "loll", "acai": 83-91% human success, 94-98% model miss rate). These failures point to over-reliance on distributional plausibility that penalizes orthographically atypical but constraint-valid patterns. --- ### Enhancing Burmese News Classification with Kolmogorov-Arnold Network Head Fine-tuning Authors: Thura Aung, Eaint Kay Khaing Kyaw, Ye Kyaw Thu, Thazin Myint Oo, Thepchai Supnithi Published: 2025-11-26 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.21081v1 In low-resource languages like Burmese, classification tasks often fine-tune only the final classification layer, keeping pre-trained encoder weights frozen. While Multi-Layer Perceptrons (MLPs) are commonly used, their fixed non-linearity can limit expressiveness and increase computational cost. This work explores Kolmogorov-Arnold Networks (KANs) as alternative classification heads, evaluating Fourier-based FourierKAN, Spline-based EfficientKAN, and Grid-based FasterKAN-across diverse embeddings including TF-IDF, fastText, and multilingual transformers (mBERT, Distil-mBERT). Experimental results show that KAN-based heads are competitive with or superior to MLPs. EfficientKAN with fastText achieved the highest F1-score (0.928), while FasterKAN offered the best trade-off between speed and accuracy. On transformer embeddings, EfficientKAN matched or slightly outperformed MLPs with mBERT (0.917 F1). These findings highlight KANs as expressive, efficient alternatives to MLPs for low-resource language classification. --- ### Data-Driven Assessment of Concrete Slab Integrity via Impact-Echo Signals and Neural Networks Authors: Yeswanth Ravichandran, Duoduo Liao, Charan Teja Kurakula Published: 2025-11-26 Categories: eess.SP, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.21080v1 Subsurface defects such as delamination, voids, and honeycombing critically affect the durability of concrete bridge decks but are difficult to detect reliably using visual inspection or manual sounding. This paper presents a machine learning based Impact Echo (IE) framework that automates both defect localization and multi-class classification of common concrete defects. Raw IE signals from Federal Highway Administration (FHWA) laboratory slabs and in-service bridge decks are transformed via Fast Fourier Transform (FFT) into dominant peak-frequency features and interpolated into spatial maps for defect zone visualization. Unsupervised k-means clustering highlights low-frequency, defect-prone regions, while Ground Truth Masks (GTMs) derived from seeded lab defects are used to validate spatial accuracy and generate high-confidence training labels. From these validated regions, spatially ordered peak-frequency sequences are constructed and fed into a stacked Long Short-Term Memory (LSTM) network that classifies four defect types shallow delamination, deep delamination, voids, and honeycombing with 73% overall accuracy. Field validation on the bridge deck demonstrates that models trained on laboratory data generalize under realistic coupling, noise, and environmental variability. The proposed framework enhances the objectivity, scalability, and repeatability of Non-Destructive Evaluation (NDE), supporting intelligent, data-driven bridge health monitoring at a network scale. --- ### Deceptron: Learned Local Inverses for Fast and Stable Physics Inversion Authors: Aaditya L. Kachhadiya Published: 2025-11-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.21076v1 Inverse problems in the physical sciences are often ill-conditioned in input space, making progress step-size sensitive. We propose the Deceptron, a lightweight bidirectional module that learns a local inverse of a differentiable forward surrogate. Training combines a supervised fit, forward-reverse consistency, a lightweight spectral penalty, a soft bias tie, and a Jacobian Composition Penalty (JCP) that encourages $J_g(f(x))\,J_f(x)\!\approx\!I$ via JVP/VJP probes. At solve time, D-IPG (Deceptron Inverse-Preconditioned Gradient) takes a descent step in output space, pulls it back through $g$, and projects under the same backtracking and stopping rules as baselines. On Heat-1D initial-condition recovery and a Damped Oscillator inverse problem, D-IPG reaches a fixed normalized tolerance with $\sim$20$\times$ fewer iterations on Heat and $\sim$2-3$\times$ fewer on Oscillator than projected gradient, competitive in iterations and cost with Gauss-Newton. Diagnostics show JCP reduces a measured composition error and tracks iteration gains. We also preview a single-scale 2D instantiation, DeceptronNet (v0), that learns few-step corrections under a strict fairness protocol and exhibits notably fast convergence. --- ### Aligning LLMs with Biomedical Knowledge using Balanced Fine-Tuning Authors: Zhenchao Tang, Fang Wang, Haohuai He, Jiale Zhou, Tianxu Lv, Jun Zhu, Shouzhi Chen, Minghao Yang, Yu Wang, Jiayang Wu, Yidong Song, Yaokun Li et al. Published: 2025-11-26 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.21075v4 Engineering LLMs to accelerate life sciences research requires a robust alignment with biomedical knowledge. We observe that biomedical text exhibits a fundamentally different uncertainty structure from general text: dense low-confidence runs encode epistemic knowledge gaps (dense causal chains, rare entities) rather than the sparse aleatoric stylistic variation typical of general text. Based on this discovery, we propose Balanced Fine-Tuning (BFT), a dual-scale post-training method that combines group-normalized token reweighting with sequence-level reallocation toward knowledge-dense samples exhibiting dense epistemic uncertainty. Across medical evaluation, biological reasoning, sparse-reward RL, and biological representation tasks, BFT provides more consistent gains than SFT and DFT under a shared training setup. When replacing the default closed-source backbones in GeneAgent (GPT-4o) and VCWorld (Gemini-2.5-Flash), the BFT-aligned 70B model delivers stronger performance across biological process reasoning and chemical perturbation prediction. Critically, all BFT variants further improve after subsequent GRPO with sparse rewards, while SFT and DFT degrade, suggesting that epistemic-aware post-training provides a more robust policy initialization. Beyond text generation, BFT-aligned LLMs produce more accurate and professional biomedical profile texts; after encoding these profiles with a text embedding model, the resulting representations support gene-level, cell-level, and perturbation-response tasks, suggesting that BFT-enhanced generation can facilitate biological representation and, in turn, broader biomedical downstream tasks. --- ### Inference for Similarity and Alignability between Noisy High-Dimensional Datasets Authors: Hongrui Chen, Rong Ma Published: 2025-11-26 Categories: math.ST, stat.ME, stat.ML arXiv: https://arxiv.org/abs/2511.21074v2 The rapid growth of high-dimensional datasets across a wide range of scientific domains has created an urgent need for new statistical methods to compare distributions with underlying low-dimensional structure. Assessing similarity between high-dimensional datasets whose observations concentrate near low-dimensional manifolds is particularly challenging due to the nontrivial effects of noise in high dimensions. We propose a principled framework for statistical inference on the similarity and alignability of high-dimensional datasets with low-dimensional smooth signals under heterogeneous noise. The key idea is to link the spectral properties of the observed data matrices to the geometry of their underlying signal distributions. Under a manifold signal-plus-noise model, we build on the principal variances associated to the underlying signals and develop a scale- and rotation-invariant dissimilarity measure between two datasets that may differ in sample size and noise structure. We further construct an estimator of the dissimilarity and a statistical test of dataset alignability, namely, whether the dissimilarity vanishes. The proposed methodology and its theoretical guarantees under high-dimensional asymptotic settings draw on recent advances in random matrix theory (RMT). The proposed framework accommodates heterogeneous noise across datasets and provides a fast, theoretically grounded approach to comparing high-dimensional datasets with low-dimensional structures. Through extensive simulations and analyses of multiple single-cell datasets, we demonstrate that the proposed method substantially outperforms existing approaches. --- ### Context-Aware Pragmatic Metacognitive Prompting for Sarcasm Detection Authors: Michael Iskandardinata, William Christian, Derwin Suhartono Published: 2025-11-26 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2511.21066v1 Detecting sarcasm remains a challenging task in the areas of Natural Language Processing (NLP) despite recent advances in neural network approaches. Currently, Pre-trained Language Models (PLMs) and Large Language Models (LLMs) are the preferred approach for sarcasm detection. However, the complexity of sarcastic text, combined with linguistic diversity and cultural variation across communities, has made the task more difficult even for PLMs and LLMs. Beyond that, those models also exhibit unreliable detection of words or tokens that require extra grounding for analysis. Building on a state-of-the-art prompting method in LLMs for sarcasm detection called Pragmatic Metacognitive Prompting (PMP), we introduce a retrieval-aware approach that incorporates retrieved contextual information for each target text. Our pipeline explores two complementary ways to provide context: adding non-parametric knowledge using web-based retrieval when the model lacks necessary background, and eliciting the model's own internal knowledge for a self-knowledge awareness strategy. We evaluated our approach with three datasets, such as Twitter Indonesia Sarcastic, SemEval-2018 Task 3, and MUStARD. Non-parametric retrieval resulted in a significant 9.87% macro-F1 improvement on Twitter Indonesia Sarcastic compared to the original PMP method. Self-knowledge retrieval improves macro-F1 by 3.29% on Semeval and by 4.08% on MUStARD. These findings highlight the importance of context in enhancing LLMs performance in sarcasm detection task, particularly the involvement of culturally specific slang, references, or unknown terms to the LLMs. Future work will focus on optimizing the retrieval of relevant contextual information and examining how retrieval quality affects performance. The experiment code is available at: https://github.com/wllchrst/sarcasm-detection_pmp_knowledge-base. --- ### OVOD-Agent: A Markov-Bandit Framework for Proactive Visual Reasoning and Self-Evolving Detection Authors: Chujie Wang, Jianyu Lu, Zhiyuan Luo, Xi Chen, Chu He Published: 2025-11-26 Categories: cs.AI, cs.CV arXiv: https://arxiv.org/abs/2511.21064v2 Open-Vocabulary Object Detection (OVOD) aims to enable detectors to generalize across categories by leveraging semantic information. Although existing methods are pretrained on large vision-language datasets, their inference is still limited to fixed category names, creating a gap between multimodal training and unimodal inference. Previous work has shown that improving textual representation can significantly enhance OVOD performance, indicating that the textual space is still underexplored. To this end, we propose OVOD-Agent, which transforms passive category matching into proactive visual reasoning and self-evolving detection. Inspired by the Chain-of-Thought (CoT) paradigm, OVOD-Agent extends the textual optimization process into an interpretable Visual-CoT with explicit actions. OVOD's lightweight nature makes LLM-based management unsuitable; instead, we model visual context transitions as a Weakly Markovian Decision Process (w-MDP) over eight state spaces, which naturally represents the agent's state, memory, and interaction dynamics. A Bandit module generates exploration signals under limited supervision, helping the agent focus on uncertain regions and adapt its detection policy. We further integrate Markov transition matrices with Bandit trajectories for self-supervised Reward Model (RM) optimization, forming a closed loop from Bandit exploration to RM learning. Experiments on COCO and LVIS show that OVOD-Agent provides consistent improvements across OVOD backbones, particularly on rare categories, confirming the effectiveness of the proposed framework. --- ### G-Net: A Provably Easy Construction of High-Accuracy Random Binary Neural Networks Authors: Alireza Aghasi, Nicholas Marshall, Saeid Pourmand, Wyatt Whiting Published: 2025-11-26 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2511.21063v1 We propose a novel randomized algorithm for constructing binary neural networks with tunable accuracy. This approach is motivated by hyperdimensional computing (HDC), which is a brain-inspired paradigm that leverages high-dimensional vector representations, offering efficient hardware implementation and robustness to model corruptions. Unlike traditional low-precision methods that use quantization, we consider binary embeddings of data as points in the hypercube equipped with the Hamming distance. We propose a novel family of floating-point neural networks, G-Nets, which are general enough to mimic standard network layers. Each floating-point G-Net has a randomized binary embedding, an embedded hyperdimensional (EHD) G-Net, that retains the accuracy of its floating-point counterparts, with theoretical guarantees, due to the concentration of measure. Empirically, our binary models match convolutional neural network accuracies and outperform prior HDC models by large margins, for example, we achieve almost 30\% higher accuracy on CIFAR-10 compared to prior HDC models. G-Nets are a theoretically justified bridge between neural networks and randomized binary neural networks, opening a new direction for constructing robust binary/quantized deep learning models. Our implementation is available at https://github.com/GNet2025/GNet. --- ### Zipf Distributions from Two-Stage Symbolic Processes: Stability Under Stochastic Lexical Filtering Authors: Vladimir Berman Published: 2025-11-26 Categories: stat.ME, cs.CL, stat.ML arXiv: https://arxiv.org/abs/2511.21060v1 Zipf's law in language lacks a definitive origin, debated across fields. This study explains Zipf-like behavior using geometric mechanisms without linguistic elements. The Full Combinatorial Word Model (FCWM) forms words from a finite alphabet, generating a geometric distribution of word lengths. Interacting exponential forces yield a power-law rank-frequency curve, determined by alphabet size and blank symbol probability. Simulations support predictions, matching English, Russian, and mixed-genre data. The symbolic model suggests Zipf-type laws arise from geometric constraints, not communicative efficiency. --- ### Long-Term Alzheimers Disease Prediction: A Novel Image Generation Method Using Temporal Parameter Estimation with Normal Inverse Gamma Distribution on Uneven Time Series Authors: Xin Hong, Xinze Sun, Yinhao Li, Yen-Wei Chen Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21057v1 Image generation can provide physicians with an imaging diagnosis basis in the prediction of Alzheimer's Disease (AD). Recent research has shown that long-term AD predictions by image generation often face difficulties maintaining disease-related characteristics when dealing with irregular time intervals in sequential data. Considering that the time-related aspects of the distribution can reflect changes in disease-related characteristics when images are distributed unevenly, this research proposes a model to estimate the temporal parameter within the Normal Inverse Gamma Distribution (T-NIG) to assist in generating images over the long term. The T-NIG model employs brain images from two different time points to create intermediate brain images, forecast future images, and predict the disease. T-NIG is designed by identifying features using coordinate neighborhoods. It incorporates a time parameter into the normal inverse gamma distribution to understand how features change in brain imaging sequences that have varying time intervals. Additionally, T-NIG utilizes uncertainty estimation to reduce both epistemic and aleatoric uncertainties in the model, which arise from insufficient temporal data. In particular, the T-NIG model demonstrates state-of-the-art performance in both short-term and long-term prediction tasks within the dataset. Experimental results indicate that T-NIG is proficient in forecasting disease progression while maintaining disease-related characteristics, even when faced with an irregular temporal data distribution. --- ### Bilevel Data Curation for LLM Fine-tuning: Offline Selection and Online Self-Refining Generation Authors: Quan Xiao, Yutong Xuan, Gaowen Liu, Ramana Rao Kompella, Tianyi Chen Published: 2025-11-26 Categories: cs.LG, cs.CL, math.OC arXiv: https://arxiv.org/abs/2511.21056v2 Supervised fine-tuning (SFT) datasets are critical to the downstream performance of large language models, yet they often contain low-quality or harmful question-response pairs. To improve SFT data quality, we develop a unified bilevel framework that combines offline data selection with the online self-refining generation. In the offline setting, bilevel data selection (BDS) selects question-response pairs from the offline SFT dataset to maximize the validation performance. We theoretically show that the optimal model given by BDS outperforms direct data mixing approach in useful data coverage. Moreover, we provide a global convergence analysis for gradient-based BDS approach for one-layer Transformer, showing that the epsilon-global optimum of offline BDS is achievable in finite time. Although efficient, offline BDS discards potentially harmful questions together with responses, thereby reducing question diversity. We address this limitation by refining the responses to selected questions using online self-refining generation framework. However, BDS is inefficient to update the response weights when responses are regenerated online. To address this issue, we introduce bilevel multi-objective optimization (BMO) for response-level weighting. We show that BMO recovers the same validation-aligned solution as BDS, but admits a closed-form importance-ratio weight that adapts to regenerated responses. Experiments on LLM quality enhancement and safety-aware fine-tuning demonstrate that the proposed framework consistently improves both data quality and downstream fine-tuning performance. --- ### Efficient Diffusion Planning with Temporal Diffusion Authors: Jiaming Guo, Rui Zhang, Zerun Li, Yunkai Gao, Shaohui Peng, Siming Lan, Xing Hu, Zidong Du, Xishan Zhang, Ling Li Published: 2025-11-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.21054v1 Diffusion planning is a promising method for learning high-performance policies from offline data. To avoid the impact of discrepancies between planning and reality on performance, previous works generate new plans at each time step. However, this incurs significant computational overhead and leads to lower decision frequencies, and frequent plan switching may also affect performance. In contrast, humans might create detailed short-term plans and more general, sometimes vague, long-term plans, and adjust them over time. Inspired by this, we propose the Temporal Diffusion Planner (TDP) which improves decision efficiency by distributing the denoising steps across the time dimension. TDP begins by generating an initial plan that becomes progressively more vague over time. At each subsequent time step, rather than generating an entirely new plan, TDP updates the previous one with a small number of denoising steps. This reduces the average number of denoising steps, improving decision efficiency. Additionally, we introduce an automated replanning mechanism to prevent significant deviations between the plan and reality. Experiments on D4RL show that, compared to previous works that generate new plans every time step, TDP improves the decision-making frequency by 11-24.8 times while achieving higher or comparable performance. --- ### AerialMind: Towards Referring Multi-Object Tracking in UAV Scenarios Authors: Chenglizhao Chen, Shaofeng Liang, Runwei Guan, Xiaolou Sun, Haocheng Zhao, Haiyun Jiang, Tao Huang, Henghui Ding, Qing-Long Han Published: 2025-11-26 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2511.21053v2 Referring Multi-Object Tracking (RMOT) aims to achieve precise object detection and tracking through natural language instructions, representing a fundamental capability for intelligent robotic systems. However, current RMOT research remains mostly confined to ground-level scenarios, which constrains their ability to capture broad-scale scene contexts and perform comprehensive tracking and path planning. In contrast, Unmanned Aerial Vehicles (UAVs) leverage their expansive aerial perspectives and superior maneuverability to enable wide-area surveillance. Moreover, UAVs have emerged as critical platforms for Embodied Intelligence, which has given rise to an unprecedented demand for intelligent aerial systems capable of natural language interaction. To this end, we introduce AerialMind, the first large-scale RMOT benchmark in UAV scenarios, which aims to bridge this research gap. To facilitate its construction, we develop an innovative semi-automated collaborative agent-based labeling assistant (COALA) framework that significantly reduces labor costs while maintaining annotation quality. Furthermore, we propose HawkEyeTrack (HETrack), a novel method that collaboratively enhances vision-language representation learning and improves the perception of UAV scenarios. Comprehensive experiments validated the challenging nature of our dataset and the effectiveness of our method. --- ### MUSE: Manipulating Unified Framework for Synthesizing Emotions in Images via Test-Time Optimization Authors: Yingjie Xia, Xi Wang, Jinglei Shi, Vicky Kalogeiton, Jian Yang Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21051v1 Images evoke emotions that profoundly influence perception, often prioritized over content. Current Image Emotional Synthesis (IES) approaches artificially separate generation and editing tasks, creating inefficiencies and limiting applications where these tasks naturally intertwine, such as therapeutic interventions or storytelling. In this work, we introduce MUSE, the first unified framework capable of both emotional generation and editing. By adopting a strategy conceptually aligned with Test-Time Scaling (TTS) that widely used in both LLM and diffusion model communities, it avoids the requirement for additional updating diffusion model and specialized emotional synthesis datasets. More specifically, MUSE addresses three key questions in emotional synthesis: (1) HOW to stably guide synthesis by leveraging an off-the-shelf emotion classifier with gradient-based optimization of emotional tokens; (2) WHEN to introduce emotional guidance by identifying the optimal timing using semantic similarity as a supervisory signal; and (3) WHICH emotion to guide synthesis through a multi-emotion loss that reduces interference from inherent and similar emotions. Experimental results show that MUSE performs favorably against all methods for both generation and editing, improving emotional accuracy and semantic diversity while maintaining an optimal balance between desired content, adherence to text prompts, and realistic emotional expression. It establishes a new paradigm for emotion synthesis. --- ### Breaking the Safety-Capability Tradeoff: Reinforcement Learning with Verifiable Rewards Maintains Safety Guardrails in LLMs Authors: Dongkyu Derek Cho, Huan Song, Arijit Ghosh Chowdhury, Haotian An, Yawei Wang, Rohit Thekkanal, Negin Sokhandan, Sharlina Keshava, Hannah Marlowe Published: 2025-11-26 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2511.21050v1 Fine-tuning large language models (LLMs) for downstream tasks typically exhibit a fundamental safety-capability tradeoff, where improving task performance degrades safety alignment even on benign datasets. This degradation persists across standard approaches including supervised finetuning (SFT) and reinforcement learning from human feedback (RLHF). While reinforcement learning with verifiable rewards (RLVR) has emerged as a promising alternative that optimizes models on objectively measurable tasks, its safety implications remain unexplored. We present the first comprehensive theoretical and empirical analysis of safety properties in RLVR. Theoretically, we derive upper bounds on safety drift under KL-constrained optimization and prove conditions under which safety degradation is eliminated. Empirically, we conduct extensive experiments across five adversarial safety benchmarks, demonstrating that RLVR can simultaneously enhance reasoning capabilities while maintaining or improving safety guardrails. Our comprehensive ablation studies examine the effects of optimization algorithms, model scale, and task domains. Our findings challenge the prevailing assumption of an inevitable safety capability trade-off, and establish that a specific training methodology can achieve both objectives simultaneously, providing insights for the safe deployment of reasoning-capable LLMs. --- ### FedAPA: Federated Learning with Adaptive Prototype Aggregation Toward Heterogeneous Wi-Fi CSI-based Crowd Counting Authors: Jingtao Guo, Yuyi Mao, Ivan Wang-Hei Ho Published: 2025-11-26 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.21048v1 Wi-Fi channel state information (CSI)-based sensing provides a non-invasive, device-free approach for tasks such as human activity recognition and crowd counting, but large-scale deployment is hindered by the need for extensive site-specific training data. Federated learning (FL) offers a way to avoid raw data sharing but is challenged by heterogeneous sensing data and device resources. This paper proposes FedAPA, a collaborative Wi-Fi CSI-based sensing algorithm that uses adaptive prototype aggregation (APA) strategy to assign similarity-based weights to peer prototypes, enabling adaptive client contributions and yielding a personalized global prototype for each client instead of a fixed-weight aggregation. During local training, we adopt a hybrid objective that combines classification learning with representation contrastive learning to align local and global knowledge. We provide a convergence analysis of FedAPA and evaluate it in a real-world distributed Wi-Fi crowd counting scenario with six environments and up to 20 people. The results show that our method outperform multiple baselines in terms of accuracy, F1 score, mean absolute error (MAE), and communication overhead, with FedAPA achieving at least a 9.65% increase in accuracy, a 9% gain in F1 score, a 0.29 reduction in MAE, and a 95.94% reduction in communication overhead. --- ### PG-ControlNet: A Physics-Guided ControlNet for Generative Spatially Varying Image Deblurring Authors: Hakki Motorcu, Mujdat Cetin Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21043v1 Spatially varying image deblurring remains a fundamentally ill-posed problem, especially when degradations arise from complex mixtures of motion and other forms of blur under significant noise. State-of-the-art learning-based approaches generally fall into two paradigms: model-based deep unrolling methods that enforce physical constraints by modeling the degradations, but often produce over-smoothed, artifact-laden textures, and generative models that achieve superior perceptual quality yet hallucinate details due to weak physical constraints. In this paper, we propose a novel framework that uniquely reconciles these paradigms by taming a powerful generative prior with explicit, dense physical constraints. Rather than oversimplifying the degradation field, we model it as a dense continuum of high-dimensional compressed kernels, ensuring that minute variations in motion and other degradation patterns are captured. We leverage this rich descriptor field to condition a ControlNet architecture, strongly guiding the diffusion sampling process. Extensive experiments demonstrate that our method effectively bridges the gap between physical accuracy and perceptual realism, outperforming state-of-the-art model-based methods as well as generative baselines in challenging, severely blurred scenarios. --- ### LungNoduleAgent: A Collaborative Multi-Agent System for Precision Diagnosis of Lung Nodules Authors: Cheng Yang, Hui Jin, Xinlei Yu, Zhipeng Wang, Yaoqun Liu, Fenglei Fan, Dajiang Lei, Gangyong Jia, Changmiao Wang, Ruiquan Ge Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21042v1 Diagnosing lung cancer typically involves physicians identifying lung nodules in Computed tomography (CT) scans and generating diagnostic reports based on their morphological features and medical expertise. Although advancements have been made in using multimodal large language models for analyzing lung CT scans, challenges remain in accurately describing nodule morphology and incorporating medical expertise. These limitations affect the reliability and effectiveness of these models in clinical settings. Collaborative multi-agent systems offer a promising strategy for achieving a balance between generality and precision in medical applications, yet their potential in pathology has not been thoroughly explored. To bridge these gaps, we introduce LungNoduleAgent, an innovative collaborative multi-agent system specifically designed for analyzing lung CT scans. LungNoduleAgent streamlines the diagnostic process into sequential components, improving precision in describing nodules and grading malignancy through three primary modules. The first module, the Nodule Spotter, coordinates clinical detection models to accurately identify nodules. The second module, the Radiologist, integrates localized image description techniques to produce comprehensive CT reports. Finally, the Doctor Agent System performs malignancy reasoning by using images and CT reports, supported by a pathology knowledge base and a multi-agent system framework. Extensive testing on two private datasets and the public LIDC-IDRI dataset indicates that LungNoduleAgent surpasses mainstream vision-language models, agent systems, and advanced expert models. These results highlight the importance of region-level semantic alignment and multi-agent collaboration in diagnosing nodules. LungNoduleAgent stands out as a promising foundational tool for supporting clinical analyses of lung nodules. --- ### CNN-LSTM Hybrid Architecture for Over-the-Air Automatic Modulation Classification Using SDR Authors: Dinanath Padhya, Krishna Acharya, Bipul Kumar Dahal, Dinesh Baniya Kshatri Published: 2025-11-26 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2511.21040v2 Automatic Modulation Classification (AMC) is a core technology for future wireless communication systems, enabling the identification of modulation schemes without prior knowledge. This capability is essential for applications in cognitive radio, spectrum monitoring, and intelligent communication networks. We propose an AMC system based on a hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) architecture, integrated with a Software Defined Radio (SDR) platform. The proposed architecture leverages CNNs for spatial feature extraction and LSTMs for capturing temporal dependencies, enabling efficient handling of complex, time-varying communication signals. The system's practical ability was demonstrated by identifying over-the-air (OTA) signals from a custom-built FM transmitter alongside other modulation schemes. The system was trained on a hybrid dataset combining the RadioML2018 dataset with a custom-generated dataset, featuring samples at Signal-to-Noise Ratios (SNRs) from 0 to 30dB. System performance was evaluated using accuracy, precision, recall, F1 score, and the Area Under the Receiver Operating Characteristic Curve (AUC-ROC). The optimized model achieved 93.48% accuracy, 93.53% precision, 93.48% recall, and an F1 score of 93.45%. The AUC-ROC analysis confirmed the model's discriminative power, even in noisy conditions. This paper's experimental results validate the effectiveness of the hybrid CNN-LSTM architecture for AMC, suggesting its potential application in adaptive spectrum management and advanced cognitive radio systems. --- ### Semantic Anchors in In-Context Learning: Why Small LLMs Cannot Flip Their Labels Authors: Anantha Padmanaban Krishna Kumar Published: 2025-11-26 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2511.21038v1 Can in-context learning (ICL) override pre-trained label semantics, or does it merely refine an existing semantic backbone? We address this question by treating LLMs as prompt-induced classifiers and contrasting their behavior under \emph{natural} demonstrations (with correct labels) and \emph{inverted} demonstrations (systematically flipping label meanings). We decompose ICL behavior into three alignment metrics (truth, prior, and prompt alignment) and introduce a semantic override rate, defined as correctness under flipped semantics. Across eight classification tasks and eight open-source LLMs (1--12B parameters), we find consistent evidence for a semantic anchor view. With natural demonstrations, ICL improves accuracy while maintaining strong prior alignment; most correct predictions coincide with zero-shot behavior, even when the prior is weak. With inverted demonstrations, models cannot learn coherent anti-semantic classifiers: prompt alignment increases only by sacrificing accuracy, and semantic override rates remain exactly zero in our few-shot 1--12B setting. Rather than flexibly remapping label meanings, ICL primarily adjusts how inputs project onto stable semantic directions learned during pre-training, clarifying fundamental limits of few-shot prompting and suggesting that overriding label semantics at these scales requires interventions beyond ICL. All code is available at: https://github.com/AnanthaPadmanaban-KrishnaKumar/semantic-anchors-icl. --- ### RAVQ-HoloNet: Rate-Adaptive Vector-Quantized Hologram Compression Authors: Shima Rafiei, Zahra Nabizadeh Shahr-Babak, Soroush Khoubyarian, Alexandre Cooper, Shadrokh Samavi, Shahram Shirani Published: 2025-11-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.21035v2 Holography offers significant potential for AR/VR applications. However, its adoption is limited by the high demand for data compression. Existing deep learning approaches generally lack rate adaptivity within a single network and often require multiple models to cover different bandwidth requirements. We present RAVQ-HoloNet, a rate-adaptive vector quantization framework that integrates the rate-adaptive compression with the transformation of image data into phase-only hologram. RAVQ-HoloNet achieves high-fidelity reconstructions, outperforming current state-of-the-art methods implemented via two distinct architectural configurations: a standard model optimized for low bit rates and a deeper, extended variant tailored for ultra low bit rate setting. To evaluate these models, we utilized the DIV2K dataset as a benchmark for high-fidelity holographic reconstruction. Quantitative analysis in the simulation reveals that our approach significantly surpasses current benchmarks. Specifically, in the low bit rate domain, our method achieves a BD-Rate reduction of -33.91% and a BD-PSNR gain of 1.02dB relative to the state-of-the-art method. Additionally, experimental results on the SLM device show that our method achieves higher contrast and improved quality. --- ### Prediction of Herd Life in Dairy Cows Using Multi-Head Attention Transformers Authors: Mahdi Saki, Justin Lipman Published: 2025-11-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.21034v1 Dairy farmers should decide to keep or cull a cow based on an objective assessment of her likely performance in the herd. For this purpose, farmers need to identify more resilient cows, which can cope better with farm conditions and complete more lactations. This decision-making process is inherently complex, with significant environmental and economic implications. In this study, we develop an AI-driven model to predict cow longevity using historical multivariate time-series data recorded from birth. Leveraging advanced AI techniques, specifically Multi-Head Attention Transformers, we analysed approximately 780,000 records from 19,000 unique cows across 7 farms in Australia. The results demonstrate that our model achieves an overall determination coefficient of 83% in predicting herd life across the studied farms, highlighting its potential for practical application in dairy herd management. --- ### Towards Trustworthy Legal AI through LLM Agents and Formal Reasoning Authors: Linze Chen, Yufan Cai, Zhe Hou, Jin Song Dong Published: 2025-11-26 Categories: cs.AI arXiv: https://arxiv.org/abs/2511.21033v2 Legal decisions should be logical and based on statutory laws. While large language models(LLMs) are good at understanding legal text, they cannot provide verifiable justifications. We present L4L, a solver-centric framework that enforces formal alignment between LLM-based legal reasoning and statutory laws. The framework integrates role-differentiated LLM agents with SMT-backed verification, combining the flexibility of natural language with the rigor of symbolic reasoning. Our approach operates in four stages: (1) Statute Knowledge Building, where LLMs autoformalize legal provisions into logical constraints and validate them through case-level testing; (2) Dual Fact-and-Statute Extraction, in which the prosecutor-and defense-aligned agents independently map case narratives to argument tuples; (3) Solver-Centric Adjudication, where SMT solvers check the legal admissibility and consistency of the arguments against the formalized statute knowledge; (4) Judicial Rendering, in which a judge agent integrates solver-validated reasoning with statutory interpretation and similar precedents to produce a legally grounded verdict. Experiments on public legal benchmarks show that L4L consistently outperforms baselines, while providing auditable justifications that enable trustworthy legal AI. --- ### A Probabilistic Framework for Temporal Distribution Generalization in Industry-Scale Recommender Systems Authors: Yuxuan Zhu, Cong Fu, Yabo Ni, Anxiang Zeng, Yuan Fang Published: 2025-11-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.21032v1 Temporal distribution shift (TDS) erodes the long-term accuracy of recommender systems, yet industrial practice still relies on periodic incremental training, which struggles to capture both stable and transient patterns. Existing approaches such as invariant learning and self-supervised learning offer partial solutions but often suffer from unstable temporal generalization, representation collapse, or inefficient data utilization. To address these limitations, we propose ELBO$_\text{TDS}$, a probabilistic framework that integrates seamlessly into industry-scale incremental learning pipelines. First, we identify key shifting factors through statistical analysis of real-world production data and design a simple yet effective data augmentation strategy that resamples these time-varying factors to extend the training support. Second, to harness the benefits of this extended distribution while preventing representation collapse, we model the temporal recommendation scenario using a causal graph and derive a self-supervised variational objective, ELBO$_\text{TDS}$, grounded in the causal structure. Extensive experiments supported by both theoretical and empirical analysis demonstrate that our method achieves superior temporal generalization, yielding a 2.33\% uplift in GMV per user and has been successfully deployed in Shopee Product Search. Code is available at https://github.com/FuCongResearchSquad/ELBO4TDS. --- ### FlowerDance: MeanFlow for Efficient and Refined 3D Dance Generation Authors: Kaixing Yang, Xulong Tang, Ziqiao Peng, Xiangyue Zhang, Puwei Wang, Jun He, Hongyan Liu Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21029v3 Music-to-dance generation aims to translate auditory signals into expressive human motion, with broad applications in virtual reality, choreography, and digital entertainment. Despite promising progress, the limited generation efficiency of existing methods leaves insufficient computational headroom for high-fidelity 3D rendering, thereby constraining the expressiveness of 3D characters during real-world applications. Thus, we propose FlowerDance, which not only generates refined motion with physical plausibility and artistic expressiveness, but also achieves significant generation efficiency on inference speed and memory utilization. Specifically, FlowerDance combines MeanFlow with Physical Consistency Constraints, which enables high-quality motion generation with only a few sampling steps. Moreover, FlowerDance leverages a simple but efficient model architecture with BiMamba-based backbone and Channel-Level Cross-Modal Fusion, which generates dance with efficient non-autoregressive manner. Meanwhile, FlowerDance supports motion editing, enabling users to interactively refine dance sequences. Extensive experiments on AIST++ and FineDance show that FlowerDance achieves state-of-the-art results in both motion quality and generation efficiency. Code will be released upon acceptance. --- ### Deep Parameter Interpolation for Scalar Conditioning Authors: Chicago Y. Park, Michael T. McCann, Cristina Garcia-Cardona, Brendt Wohlberg, Ulugbek S. Kamilov Published: 2025-11-26 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2511.21028v1 We propose deep parameter interpolation (DPI), a general-purpose method for transforming an existing deep neural network architecture into one that accepts an additional scalar input. Recent deep generative models, including diffusion models and flow matching, employ a single neural network to learn a time- or noise level-dependent vector field. Designing a network architecture to accurately represent this vector field is challenging because the network must integrate information from two different sources: a high-dimensional vector (usually an image) and a scalar. Common approaches either encode the scalar as an additional image input or combine scalar and vector information in specific network components, which restricts architecture choices. Instead, we propose to maintain two learnable parameter sets within a single network and to introduce the scalar dependency by dynamically interpolating between the parameter sets based on the scalar value during training and sampling. DPI is a simple, architecture-agnostic method for adding scalar dependence to a neural network. We demonstrate that our method improves denoising performance and enhances sample quality for both diffusion and flow matching models, while achieving computational efficiency comparable to standard scalar conditioning techniques. Code is available at https://github.com/wustl-cig/parameter_interpolation. --- ### CaptionQA: Is Your Caption as Useful as the Image Itself? Authors: Shijia Yang, Yunong Liu, Bohan Zhai, Ximeng Sun, Zicheng Liu, Emad Barsoum, Manling Li, Chenfeng Xu Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21025v2 Image captions serve as efficient surrogates for visual content in multimodal systems such as retrieval, recommendation, and multi-step agentic inference pipelines. Yet current evaluation practices miss a fundamental question: Can captions stand-in for images in real downstream tasks? We propose a utility-based benchmark, CaptionQA, to evaluate model-generated captions, where caption quality is measured by how well it supports downstream tasks. CaptionQA is an extensible domain-dependent benchmark covering 4 domains--Natural, Document, E-commerce, and Embodied AI--each with fine-grained taxonomies (25 top-level and 69 subcategories) that identify useful information for domain-specific tasks. CaptionQA builds 33,027 densely annotated multiple-choice questions (50.3 per image on average) that explicitly require visual information to answer, providing a comprehensive probe of caption utility. In our evaluation protocol, an LLM answers these questions using captions alone, directly measuring whether captions preserve image-level utility and are utilizable by a downstream LLM. Evaluating state-of-the-art MLLMs reveals substantial gaps between the image and its caption utility. Notably, models nearly identical on traditional image-QA benchmarks lower by up to 32% in caption utility. We release CaptionQA along with an open-source pipeline for extension to new domains. The code is available at https://github.com/bronyayang/CaptionQA. --- ### CameraMaster: Unified Camera Semantic-Parameter Control for Photography Retouching Authors: Qirui Yang, Yang Yang, Ying Zeng, Xiaobin Hu, Bo Li, Huanjing Yue, Jingyu Yang, Peng-Tao Jiang Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21024v1 Text-guided diffusion models have greatly advanced image editing and generation. However, achieving physically consistent image retouching with precise parameter control (e.g., exposure, white balance, zoom) remains challenging. Existing methods either rely solely on ambiguous and entangled text prompts, which hinders precise camera control, or train separate heads/weights for parameter adjustment, which compromises scalability, multi-parameter composition, and sensitivity to subtle variations. To address these limitations, we propose CameraMaster, a unified camera-aware framework for image retouching. The key idea is to explicitly decouple the camera directive and then coherently integrate two critical information streams: a directive representation that captures the photographer's intent, and a parameter embedding that encodes precise camera settings. CameraMaster first uses the camera parameter embedding to modulate both the camera directive and the content semantics. The modulated directive is then injected into the content features via cross-attention, yielding a strongly camera-sensitive semantic context. In addition, the directive and camera embeddings are injected as conditioning and gating signals into the time embedding, enabling unified, layer-wise modulation throughout the denoising process and enforcing tight semantic-parameter alignment. To train and evaluate CameraMaster, we construct a large-scale dataset of 78K image-prompt pairs annotated with camera parameters. Extensive experiments show that CameraMaster produces monotonic and near-linear responses to parameter variations, supports seamless multi-parameter composition, and significantly outperforms existing methods. --- ### Structure-Aware Prototype Guided Trusted Multi-View Classification Authors: Haojian Huang, Jiahao Shi, Zhe Liu, Harold Haodong Chen, Han Fang, Hao Sun, Zhongjiang He Published: 2025-11-26 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2511.21021v1 Trustworthy multi-view classification (TMVC) addresses the challenge of achieving reliable decision-making in complex scenarios where multi-source information is heterogeneous, inconsistent, or even conflicting. Existing TMVC approaches predominantly rely on globally dense neighbor relationships to model intra-view dependencies, leading to high computational costs and an inability to directly ensure consistency across inter-view relationships. Furthermore, these methods typically aggregate evidence from different views through manually assigned weights, lacking guarantees that the learned multi-view neighbor structures are consistent within the class space, thus undermining the trustworthiness of classification outcomes. To overcome these limitations, we propose a novel TMVC framework that introduces prototypes to represent the neighbor structures of each view. By simplifying the learning of intra-view neighbor relations and enabling dynamic alignment of intra- and inter-view structure, our approach facilitates more efficient and consistent discovery of cross-view consensus. Extensive experiments on multiple public multi-view datasets demonstrate that our method achieves competitive downstream performance and robustness compared to prevalent TMVC methods. --- ### Probabilistic Wildfire Spread Prediction Using an Autoregressive Conditional Generative Adversarial Network Authors: Taehoon Kang, Taeyong Kim Published: 2025-11-26 Categories: cs.LG, cs.AI, cs.CE, cs.CV arXiv: https://arxiv.org/abs/2511.21019v1 Climate change has intensified the frequency and severity of wildfires, making rapid and accurate prediction of fire spread essential for effective mitigation and response. Physics-based simulators such as FARSITE offer high-fidelity predictions but are computationally intensive, limiting their applicability in real-time decision-making, while existing deep learning models often yield overly smooth predictions that fail to capture the complex, nonlinear dynamics of wildfire propagation. This study proposes an autoregressive conditional generative adversarial network (CGAN) for probabilistic wildfire spread prediction. By formulating the prediction task as an autoregressive problem, the model learns sequential state transitions, ensuring long-term prediction stability. Experimental results demonstrate that the proposed CGAN-based model outperforms conventional deep learning models in both overall predictive accuracy and boundary delineation of fire perimeters. These results demonstrate that adversarial learning allows the model to capture the strong nonlinearity and uncertainty of wildfire spread, instead of simply fitting the pixel average. Furthermore, the autoregressive framework facilitates systematic temporal forecasting of wildfire evolution. The proposed CGAN-based autoregressive framework enhances both the accuracy and physical interpretability of wildfire spread prediction, offering a promising foundation for time-sensitive response and evacuation planning. --- ### Gated KalmaNet: A Fading Memory Layer Through Test-Time Ridge Regression Authors: Liangzu Peng, Aditya Chattopadhyay, Luca Zancato, Elvis Nunez, Wei Xia, Stefano Soatto Published: 2025-11-26 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2511.21016v3 Linear State-Space Models (SSMs) offer an efficient alternative to softmax Attention with constant memory and linear compute, but their lossy, fading summary of the past hurts recall-oriented tasks. We propose Gated KalmaNet (GKA, pronounced "gee-ka"), a layer that accounts for the full past while retaining SSM-style efficiency. We ground our approach in the Kalman Filter (KF), and show that several existing SSM layers (DeltaNet, Gated DeltaNet, Kimi Delta Attention) are approximations to the KF recurrence under an identity error covariance assumption, which ignores how past keys and values should optimally influence state updates. In contrast, GKA maintains the full error covariance and computes the exact Kalman gain. Under a steady-state assumption that enables parallelization, this reduces to an online ridge regression with constant memory and linear compute. The standard KF equations are numerically unstable in low-precision settings (e.g., bfloat16) and hard to parallelize on GPUs. We address this with (1) adaptive regularization via input-dependent gating to control the ridge regression's condition number, and (2) Chebyshev Iteration, which we show is more stable than conventional iterative solvers in low precision. We further develop hardware-aware chunk-wise kernels for efficient training. Empirically, GKA outperforms existing SSM layers (e.g., Mamba2, Gated DeltaNet) on short-context tasks and achieves more than 10\% relative improvement on long-context RAG and LongQA up to 128k tokens. We further show GKA outperforms Mamba when extended to ImageNet classification. Our code, including Triton kernels for training and inference (vLLM), along with a model zoo of GKA-based Hybrid models at 8B and 32B scale on HuggingFace, is released under Apache 2.0. --- ### Staggered Environment Resets Improve Massively Parallel On-Policy Reinforcement Learning Authors: Sid Bharthulwar, Stone Tao, Hao Su Published: 2025-11-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.21011v1 Massively parallel GPU simulation environments have accelerated reinforcement learning (RL) research by enabling fast data collection for on-policy RL algorithms like Proximal Policy Optimization (PPO). To maximize throughput, it is common to use short rollouts per policy update, increasing the update-to-data (UTD) ra- tio. However, we find that, in this setting, standard synchronous resets introduce harmful nonstationarity, skewing the learning signal and destabilizing training. We introduce staggered resets, a simple yet effective technique where environments are initialized and reset at varied points within the task horizon. This yields training batches with greater temporal diversity, reducing the nonstationarity induced by synchronized rollouts. We characterize dimensions along which RL environments can benefit significantly from staggered resets through illustrative toy environ- ments. We then apply this technique to challenging high-dimensional robotics environments, achieving significantly higher sample efficiency, faster wall-clock convergence, and stronger final performance. Finally, this technique scales better with more parallel environments compared to naive synchronized rollouts. --- ### ChatGpt Content detection: A new approach using xlm-roberta alignment Authors: Md Tasnin Tanvir, Dr Santanu Kumar Dash, Ishan Shahnan, Nafis Fuad, Tanvir Rahman, Abdullah Al Faisal, Asadullah Al Mamun Published: 2025-11-26 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.21009v1 The challenge of separating AI-generated text from human-authored content is becoming more urgent as generative AI technologies like ChatGPT become more widely available. In this work, we address this issue by looking at both the detection of content that has been entirely generated by AI and the identification of human text that has been reworded by AI. In our work, a comprehensive methodology to detect AI- generated text using XLM-RoBERTa, a state-of-the-art multilingual transformer model. Our approach includes rigorous preprocessing, and feature extraction involving perplexity, semantic, and readability features. We fine-tuned the XLM-RoBERTa model on a balanced dataset of human and AI-generated texts and evaluated its performance. The model demonstrated high accuracy and robust performance across various text genres. Additionally, we conducted feature analysis to understand the model's decision-making process, revealing that perplexity and attention-based features are critical in differentiating between human and AI-generated texts. Our findings offer a valuable tool for maintaining academic integrity and contribute to the broader field of AI ethics by promoting transparency and accountability in AI systems. Future research directions include exploring other advanced models and expanding the dataset to enhance the model's generalizability. --- ### Estimating Ising Models in Total Variation Distance Authors: Constantinos Daskalakis, Vardis Kandiros, Rui Yao Published: 2025-11-26 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2511.21008v1 We consider the problem of estimating Ising models over $n$ variables in Total Variation (TV) distance, given $l$ independent samples from the model. While the statistical complexity of the problem is well-understood (DMR20), identifying computationally and statistically efficient algorithms has been challenging. In particular, remarkable progress has occurred in several settings, such as when the underlying graph is a tree (DP21, BGPV21), when the entries of the interaction matrix follow a Gaussian distribution (GM24, CK24), or when the bulk of its eigenvalues lie in a small interval (AJK+24, KLV24), but no unified framework for polynomial-time estimation in TV exists so far. Our main contribution is a unified analysis of the Maximum Pseudo-Likelihood Estimator (MPLE) for two general classes of Ising models. The first class includes models that have bounded operator norm and satisfy the Modified Log-Sobolev Inequality (MLSI), a functional inequality that was introduced to study the convergence of the associated Glauber dynamics to stationarity. In the second class of models, the interaction matrix has bounded infinity norm (or bounded width), which is the most common assumption in the literature for structure learning of Ising models. We show how our general results for these classes yield polynomial-time algorithms and optimal or near-optimal sample complexity guarantees in a variety of settings. Our proofs employ a variety of tools from tensorization inequalities to measure decompositions and concentration bounds. --- ### MetaRank: Task-Aware Metric Selection for Model Transferability Estimation Authors: Yuhang Liu, Wenjie Zhao, Xin Wang, Yunhui Guo Published: 2025-11-26 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.21007v2 Selecting an appropriate pre-trained source model is a critical, yet computationally expensive, task in transfer learning. Model Transferability Estimation (MTE) methods address this by providing efficient proxy metrics to rank models without full fine-tuning. In practice, the choice of which MTE metric to use is often ad hoc or guided simply by a metric's average historical performance. However, we observe that the effectiveness of MTE metrics is highly task-dependent and no single metric is universally optimal across all target datasets. To address this gap, we introduce MetaRank, a meta-learning framework for automatic, task-aware MTE metric selection. MetaRank adopts a retrieve-and-rerank cascade. A lightweight retrieval stage first narrows the metric pool using performance observed on similar meta-training datasets. To address the heterogeneous definitions and scales of MTE scores, the reranking stage represents each retrieved metric through the pairwise source-model ordering it induces on the target dataset. A product-kernel regressor combines dataset and ordering similarities to refine the ranking, enabling cross-metric transfer from historical metrics with similar ordering behavior. MetaRank then ranks the retrieved metrics and selects the most appropriate one to guide source-model selection on an unseen target dataset. Extensive experiments across 11 pre-trained models, 13 candidate MTE metrics, and 14 target datasets demonstrate that MetaRank significantly outperforms all compared baselines. --- ### TrackList: Tracing Back Query Linguistic Diversity for Head and Tail Knowledge in Open Large Language Models Authors: Ioana Buhnila, Aman Sinha, Mathieu Constant Published: 2025-11-26 Categories: cs.CL arXiv: https://arxiv.org/abs/2511.21006v2 Large Language Models (LLMs) have proven efficient in giving definition-type answers to user input queries. While for humans giving various types of answers, such as examples and paraphrases, is an easy task, LLMs struggle to provide correct answers for other than definition-type queries. In this study, we evaluated this drop in performance using TrackList, a fine-grained linguistic and statistical analysis pipeline to investigate the impact of the pre-training data on LLMs answers to diverse linguistic queries. We also introduce RefoMed-EN, an English dataset consisting of 6170 human-annotated medical terms alongside their corresponding definitions, denominations, exemplifications, explanations, or paraphrases. We studied whether the high frequency of a concept (head) or low frequency (tail) impacts the language model's performance. We evaluated the quality of the LLM's output using syntactic and semantic similarity metrics, statistical correlations and embeddings. Results showed that the LLM's task performance for definition type questions is the highest, while for the exemplification type it is the lowest. Additionally, we showed that for definition-type questions, large language models are prone to paraphrase more on popular and frequent knowledge and less on tail and technical knowledge, especially in the expert texts. ---