# Research Papers — 2026-06 · bundle page 84 of 260 > Full-text markdown of papers #4151–4200 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-06/page/84 [← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/83.md) [Next bundle →](https://aipod.dev/papers/archive/2026-06/page/85.md) ### BAC-JEPA: Label-Efficient Breast Arterial Calcification Segmentation via Synthetic Mammography-Guided Supervision Authors: Scott Chase Waggener, Lakshman Tamil Published: 2026-06-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.22089v1 Breast arterial calcification (BAC) on screening mammograms is an emerging cardiovascular risk biomarker, but quantitative use requires reproducible segmentation and expert pixel-level labels are costly. We present BAC-JEPA, a label-efficient segmentation framework trained on procedurally generated arterial calcification inserted into real mammographic backgrounds with exact masks. Candidate backgrounds were selected from model-screened mammograms with low predicted BAC response; the generator samples arterial structure, disease burden, radiographic appearance, and hard-negative distractors including nonarterial calcifications and metallic objects. Synthetic masks are paired with mammography self-supervised Vision Transformer encoders and a high-resolution convolutional decoder to produce full-resolution segmentation maps. The study used 75,472 mammography studies from 34,956 patients for background selection and representation learning, trained on synthetic images from 10,000 backgrounds, selected checkpoints with 1,000 development backgrounds, and evaluated transfer on all 1,000 human-labeled BacSeg synthetic 2D mammograms. On held-out synthetic validation data, the larger backbone achieved IoU 0.5325 and Dice 0.6357. On BacSeg, image-level classification from segmentation probability maps reached AUROC 0.8719, with 0.8547 for the smaller backbone. Four-view inference required 110.68--213.63 ms on an RTX 5090 GPU, and severe-preset synthetic image generation averaged 2.7071 s per image on a multicore workstation. These results indicate that BAC-specific synthetic supervision can produce useful image-level transfer without human pixel-level training masks, while expert-reviewed real-mammogram segmentation remains necessary for clinical validation and calibration. --- ### Can Reasoning Models Detect Changes to their Chains of Thought? Authors: Sathvik Napa, Utkarsh Singh, Chengyuan Xue, Miriam Wanner, William Walden Published: 2026-06-20 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2606.22085v1 There are many reasons one may want to edit a model's chain of thought (CoT) -- e.g., to prefill it with reasoning from a stronger model or to remove steps that may yield unsafe outputs. The success of these interventions plausibly depends on a model's inability to notice them, as the model may alter its behavior if it suspects tampering. In this work, we study whether recent reasoning models are able to detect such interventions on their CoTs under a variety of conditions: both during reasoning and after it, and when prefilled both with their own CoTs and with those of other models. Broadly, we find that (i) models exhibit only very modest detection accuracy; (ii) models struggle to identify *how* their CoT was modified; and (iii) models are about as good at detecting changes to their own CoTs as to those of other models. --- ### Patched Flow Matching: Generative Wall-Pressure Reconstruction Beyond Training-Domain Scales from Sparse Sensors Authors: Meet Hemant Parikh, Yi Liu, Jian-Xun Wang Published: 2026-06-20 Categories: physics.flu-dyn, cs.LG arXiv: https://arxiv.org/abs/2606.22084v1 Characterizing the complete wall-pressure spectrum in turbulent wall-bounded flows requires simultaneous access to the viscous-scale high-wavenumber content and the outer-layer low-wavenumber content -- a requirement that neither short-domain direct numerical simulation (DNS) nor sparse experimental measurements alone can satisfy. We propose Patched Flow Matching (Patched FM), a generative framework that fuses these two complementary sources by learning a patch-local prior over inner-scaled wall-pressure statistics from short-domain DNS and assimilating sparse sensor measurements at inference time through training-free posterior sampling. The patch-additive decomposition of the flow matching vector field decouples the generative prior from the global domain size, enabling reconstruction on domains arbitrarily larger than the training configuration. By expressing the patch prior in inner-scaled coordinates, where high-wavenumber wall-pressure statistics are approximately Reynolds-number invariant, the framework extends to higher Reynolds numbers through hierarchical transfer learning with as few as $500$ short-domain snapshots ($2.5\%$ of the base training data) at a fraction of the scratch-training cost. Applied to compressible channel-flow DNS at $Re_τ= 180$, $500$, and $1000$, Patched FM reconstructs full-resolution wall-pressure fields on a domain four times larger than the training configuration ($L_x^L = 16πδ$ versus $L_x^S = 4πδ$) from sensor coverage as low as $0.25\%$, recovering the low-wavenumber spectral content inaccessible to short-domain DNS with high fidelity in both streamwise and spanwise directions. Zero-shot generalization to unseen Reynolds numbers and ablation studies further confirm the role of inner scaling as a physical prerequisite for data-efficient Reynolds-number transfer. --- ### CodeTeam: An LLM-Powered Multi-Agent Framework for Repository-Level Code Generation Authors: Yifei Wang, Ruiyin Li, Peng Liang, Qiong Feng, Zengyang Li, Mojtaba Shahin, Arif Ali Khan Published: 2026-06-20 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2606.22082v1 Natural language to repository generation (NL2Repo) requires a system to construct an entire software repository from a natural-language requirements document. Compared with function-level code generation, this task demands longer planning horizons, stable interfaces across files, and iterative debugging of cross-file inconsistencies. To address these challenges, we propose CodeTeam, an LLM-based multi-agent framework that separates planning, decision making, and implementation into distinct, coordinated stages. In the planning stage, multiple Architect agents draft competing software design sketches (SDS), optionally grounded by retrieved design references. A CTO agent then evaluates, selects, and normalizes the most promising SDS into a machine-checkable contract that specifies file ownership, public interfaces, and dependency constraints. In the implementation stage, Developer agents generate code under a dependency-aware scheduler with bounded context and lightweight Git-based coordination, while a QA agent runs tests and drives iterative repairs. On the synthesis-based SketchEval benchmark, we explicitly compare CodeTeam's prompt-engineering (PE) and supervised fine-tuning (SFT) variants with the corresponding CodeS variants, where CodeTeam improves the overall SketchBLEU by 4.1 and 2.9 absolute points, respectively. On the execution-based NL2Repo-Bench benchmark, used as an external validation protocol, CodeTeam achieves the highest average test pass rate in both settings (34.6% PE, 42.3% SFT), confirming that the sketch-improvements extend to functional correctness under upstream test suites. Ablation results show that project-specific developer allocation and retrieval-augmented planning each contribute substantially to the SketchBLEU improvement (9.9% and 8.1% relative, respectively). CodeTeam and the experimental results are available at https://github.com/WhitenWhiten/CodeTeam --- ### Where Does the Signal Live? A Web Data Recipe for Medical Encoder Pretraining Authors: Bofeng Huang, Jacques Sun, Diane Bouchacourt, Nicolas Barascud, Fajwel Fogel Published: 2026-06-20 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2606.22079v2 Web data curation has been widely studied for decoder Large Language Model (LLM) pretraining. Encoders for dense-terminology domains such as medicine, by contrast, are pretrained on small, manually-curated corpora that limit scalability and writing style diversity, a bottleneck even more severe in non-English clinical settings. Whether web-scale data curation also benefits encoder Masked Language Modeling (MLM) in a dense-terminology domain remains an open question. To address this, we introduce two complementary levers. Medical-term density filtering selects documents rich in medical terms. Signal-amplifying rephrasing uses an LLM to rewrite documents into denser variants with broader entity contexts. We instantiate the recipe on French medical NLP. The medical-term density filter outperforms the widely-used educational quality filter on downstream medical tasks, and the two complement each other. Signal-amplifying rephrasing alone improves on raw web data, and mixing it with filtered web data produces the largest gain. The recipe yields FineMed, a French medical pretraining corpus, and DoctoBERT, a state-of-the-art French medical encoder family evaluated on both the public benchmark DrBenchmark and a proprietary clinical Named Entity Recognition (NER) task. --- ### Where Does the Signal Live? A Web Data Recipe for Medical Encoder Pretraining Authors: Bofeng Huang, Jacques Sun, Diane Bouchacourt, Nicolas Barascud, Fajwel Fogel Published: 2026-06-20 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2606.22079v1 Web data curation has been widely studied for decoder Large Language Model (LLM) pretraining. Encoders for dense-terminology domains such as medicine, by contrast, are pretrained on small, manually-curated corpora that limit scalability and writing style diversity, a bottleneck even more severe in non-English clinical settings. Whether web-scale data curation also benefits encoder Masked Language Modeling (MLM) in a dense-terminology domain remains an open question. To address this, we introduce two complementary levers. Medical-term density filtering selects documents rich in medical terms. Signal-amplifying rephrasing uses an LLM to rewrite documents into denser variants with broader entity contexts. We instantiate the recipe on French medical NLP. The medical-term density filter outperforms the widely-used educational quality filter on downstream medical tasks, and the two complement each other. Signal-amplifying rephrasing alone improves on raw web data, and mixing it with filtered web data produces the largest gain. The recipe yields FineMed, a French medical pretraining corpus, and DoctoBERT, a state-of-the-art French medical encoder family evaluated on both the public benchmark DrBenchmark and a proprietary clinical Named Entity Recognition (NER) task. --- ### Morphology-Aware Multimodal Representation Learning for Insect Phylogenetic Reconstruction Authors: Zixuan Liu, Kaijie Yu, Chun He, Xiaoxu Cai, Xinhai Ye, Haishuai Wang, Gongyin Ye, Jiajun Bu Published: 2026-06-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.22077v1 Morphological traits provide important evidence for phylogenetic reconstruction and evolutionary relationship analysis. Recent image-based approaches have introduced deep learning, particularly convolutional models, to derive morphological features from specimen images, but these methods generally rely on single-modality visual representations and do not explicitly incorporate morphological semantics. This study proposes a morphology-aware multimodal alignment framework for insect phylogenetic reconstruction. The framework combines specimen images with curated morphological descriptions by adapting a vision transformer through parameter-efficient fine-tuning and supervised contrastive learning, followed by image-text alignment in a shared latent space. The learned image embeddings are then used as continuous traits for Bayesian phylogenetic reconstruction. On the public Rove-Tree-11 dataset, comparative and ablation experiments across multiple visual backbones and feature adaptation strategies demonstrate that multimodal alignment improves topological agreement with the reference phylogeny. The results indicate that the proposed framework can derive morphology-aware visual traits for computational phylogenetic reconstruction. --- ### Learning Cross-View Semantic Priors for Single-Reference Unseen Object Pose Estimation Authors: Jiahong Chen, Jinghao Wang, Ziwen Wang, Zi Wang, Banglei Guan, Qifeng Yu Published: 2026-06-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.22076v2 Single-reference unseen object 6D pose estimation reduces object onboarding by estimating poses of arbitrary novel objects from only one reference view. Recent correspondence-based pipelines have achieved robust performance with vision foundation model (VFM) features. However, they typically treat these features as intra-view descriptors, leaving dense visual-semantic cues, including appearance, structure, and context, insufficiently exchanged across views before geometric decoding. Consequently, the decoded point features may lack joint semantic and geometric discriminability, making correspondence estimation still difficult in challenging cases. Instead of processing features independently, we build the correspondence pipeline around an early cross-view semantic prior. Specifically, cross-view semantic interaction (CVSI) enables dense query and reference VFM tokens to exchange semantic context and form a cross-view prior. Nevertheless, direct CVSI may disturb the VFM token structure, while the resulting semantic prior still needs 3D representation consistency for rigid correspondence. To make this CVSI prior reliable for 3D correspondence learning, we introduce two complementary training-time constraints: the intra-view structure preservation (IVSP) loss preserves the original intra-view token affinity structure during interaction, while the reference-anchored geometric consistency (RAGC) loss enforces spatial representation consistency of decoded point features. The final pose is recovered from learned correspondences through weighted SVD. We further construct a challenging view-pair protocol from the BOP Challenge datasets YCB-V and TUD-L to evaluate robustness in difficult matching scenarios. Extensive experiments on six benchmarks under different view-pair settings show that our method achieves state-of-the-art performance while maintaining comparable inference speed. --- ### Frequency-Domain Neural ODEs for Modeling Non-Linear Dynamical Systems Authors: Mohammed Ashraf, Ayman A. El-Badawy Published: 2026-06-20 Categories: cs.LG, math.DS arXiv: https://arxiv.org/abs/2606.22075v1 Standard continuous-depth models, such as Neural Ordinary Differential Equations (NODEs), offer significant advantages in modeling physical systems by learning continuous vector fields rather than discrete temporal steps. However, when applied to complex dynamical systems, standard NODEs frequently struggle with highly nonlinear dynamics. This paper investigates the Frequency-domain Neural ODE (FNODE), an architecture that projects continuous temporal dynamics into the frequency domain using the Fast Fourier Transform (FFT). By operating in the frequency domain, the model provides better generalization to the dynamical system. The architecture is empirically evaluated against discrete models, specifically Gated Recurrent Units (GRUs) and Long Short-Term Memory (LSTMs), and other continuous-depth variants, including Augmented Neural ODE (ANODE), across four distinct dynamical systems: the Lotka-Volterra model, the forced Duffing oscillator, the Van der Pol oscillator, and the Lorenz system. To rigorously assess generalization and robustness, curriculum and ensemble learning are used to evaluate the model's convergence by estimating confidence intervals across different ensemble models. The empirical results demonstrate that the FNODE architecture achieves better generalization while exhibiting remarkable convergence stability. --- ### A Controlled Study of CLIP-Based Body-Scene Fusion for Emotion Recognition in Context Authors: Zubair Abbas, Muhammad Umair, Muqaddas Hameed Published: 2026-06-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.22072v2 Apparent emotion in natural images is often not visible from the face alone. The face may be small, hidden, or neutral, while posture and scene context carry much of the evidence. This work studies context-aware emotion recognition on EMOTIC with an image-only two-stream model. A ResNet-18 body stream encodes the target-person crop, and a CLIP ViT-B/16 scene stream encodes the full image. The fused feature predicts 26 categorical emotion labels and the continuous valence, arousal, and dominance values. This study examines whether small context-debiasing or rare-class training changes still help after adding a CLIP scene encoder. The clean two-stream model is compared with simplified CCIM-style intervention, CLEF-lite context-bias subtraction, ASL tuning, and class-balanced sampling under the same implementation pipeline. No tested variant improves over the clean two-stream model, which achieves 34.52% mAP on the EMOTIC test split. CLIP gives the model broad scene semantics, but the simplified causal, counterfactual, and rare-class changes do not automatically improve performance. Most remaining errors are in rare and subtle emotion categories, so the next step should focus on label relationships and finer subject-context interaction. --- ### Alternate loss functions and regression models that achieve robustness to outliers by modulating the learning rate Authors: Mathew Mithra Noel, Arindam Banerjee, Yug D. Oswal, Geraldine Bessie Amali D, Venkataraman Muthiah-Nakarajan Published: 2026-06-20 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.22068v2 Most real-world datasets used for training supervised learning models are contaminated with noisy data and outliers leading to large prediction errors. This paper proposes a new approach for achieving robustness where the learning rate is modulated by a factor that is sensitive to outliers. In this approach a reduction of the learning rate is shown to be achieved by using alternate loss functions that are infinitely differentiable, strictly convex or quasiconvex and more closely approximate the absolute error than Huber and log-cosh losses. A comparison of the performance of regression models trained with different loss functions on a wide variety of benchmarks and datasets is presented to demonstrate the superior performance of the Square Root Loss (SRL) and Smooth Mean Absolute Error (SMAE) losses proposed in this paper. Two new robust linear regression models are presented. Highly vectorized robust parameter update formulae that take advantage of modern GPUs for both stochastic and batch gradient descent are presented. --- ### How Should a Simulation-to-Reality Transfer Budget Be Spent? Authors: Syed Hamzah Rizvi, Yash Vardhan Tomar Published: 2026-06-20 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2606.22062v2 Simulation-to-reality transfer, often called sim-to-real transfer, is a central challenge in robot learning. Yet, the tradeoff between measuring a system more accurately and training over a broader range of simulated dynamics is still poorly understood. In this work, we focused on the allocation of real-robot measurement time between system identification and domain randomization. We studied this tradeoff in a controlled sim-to-sim pendulum setting, where a hidden-parameter model stands in for the physical robot, and the experiment sweeps identification rollouts against the width of the randomization distribution. Across the reality gaps and noise levels we tested, the measurement budget did most of the work. A small number of identification rollouts closed most of the transfer gap, and once any real data was available, policies performed best when trained at the estimated parameters rather than over a widened randomization band. Broad randomization that contained the true system still did not substitute for measurement. These results hold in a benign regime where the dynamics are identifiable and only two parameters are unknown, so structural model mismatch remains the setting where randomization breadth may become more valuable. Overall, our results suggest that sim-to-real pipelines should first measure the parameters they can and reserve randomization for the uncertainty that remains. --- ### NL2Scratch: An Executable Benchmark and Evaluation for Block-Based Programming Authors: Heejin Do, Alexandre Ballenghien, Yang Wu, April Yi Wang Published: 2026-06-20 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.22061v1 Block-based programming environments such as Scratch are widely used in early programming education, yet natural-language-to-code (NL2Code) research has focused primarily on text-based languages. Scratch programs are event-driven, visually compositional, and distributed across concurrent scripts, making conventional NL2Code assumptions and evaluation insufficient. We introduce NL2Scratch, an executable benchmark for natural-language-to-Scratch generation comprising 311,648 parser-valid NL--program pairs, whose program side is extracted from real Scratch projects and paired with semantically aligned NL descriptions. For reliable evaluation beyond surface overlap, we propose Semantic Alignment Consistency (SAC), an interpretable slot-level metric for measuring semantic agreement between descriptions and programs. With SAC, we construct a semantically validated pool of 23,594 examples, and a slot-balanced 800 diagnostic benchmark. Experiments across instruction-tuned and fine-tuned LLMs reveal a notable gap between lexical similarity and semantic alignment: models achieving token-level F1 above 0.93 often fail to attain perfect SAC, particularly on longer examples. Errors concentrate on operational slots like actions, conditions, and numeric arguments, exposing failure modes largely invisible under conventional metrics. --- ### Provably Efficient Policy-Reward Co-Pretraining for Adversarial Imitation Learning Authors: Tian Xu, Zexuan Chen, Zhilong Zhang, Yi-Chen Li, Chenyang Wang, Lei Yuan, Yang Yu Published: 2026-06-20 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.22056v1 Adversarial imitation learning (AIL) achieves high-quality imitation compared to behavioral cloning (BC), but demands substantial online environment interaction. Recent empirical work has explored initializing AIL algorithms with BC pretrained policies to address this limitation, yet a rigorous theoretical understanding of pretraining's role in AIL remains elusive. This paper provides a systematic theoretical analysis and introduces principled pretraining algorithms for accelerating AIL. We begin by analyzing AIL with policy pretraining alone, identifying reward error as the dominant source of suboptimality. This reveals a critical and previously overlooked gap: the absence of reward pretraining. Motivated by this finding, we develop a principled policy-reward co-pretraining approach grounded in a reward shaping analysis. Our analysis uncovers a fundamental connection between expert policies and shaping rewards, which naturally gives rise to CoPT-AIL, an approach that jointly pretrains both policy and reward through a single BC procedure. We prove that CoPT-AIL achieves an improved imitation gap bound over standard AIL, establishing the first theoretical guarantee for the benefits of pretraining in AIL. Experimental results confirm CoPT-AIL's superior performance over existing AIL methods. --- ### Anticipating the Optimism Gap: Predicting Distribution-Shift Degradation of RF-Impairment Detectors from In-Distribution Statistics Authors: Chakshu Baweja Published: 2026-06-20 Categories: eess.SP, cs.LG arXiv: https://arxiv.org/abs/2606.22054v2 Detectors for GNSS radio-frequency impairments (jamming, spoofing, multipath) are usually reported with a single AUC measured on the distribution they were tuned on. That number falls once conditions move, and the size of the drop is rarely known in advance because labelled field data is scarce. We ask whether this optimism can be predicted before any out-of-distribution data is seen. On an open, parameter-grounded synthetic testbed with a tunable severity shift, we evaluate thirteen detectors (five physics baselines, full-feature logistic regression and multilayer perceptrons, and single-feature learned controls) across four impairment classes. The optimism gap, the difference between in-distribution and shifted AUC, grows monotonically as the shift deepens (mean Spearman correlation 0.50). It is driven by how many observables a detector uses rather than by whether it is learned, and it varies systematically by class. Centrally, a ridge model built only from in-distribution score statistics predicts the gap for a detector it has never seen (R^2 = 0.47) and for an impairment class it has never seen (R^2 = 0.46); both are significant against a 2000-fold permutation null (p < 0.001) and survive removing the feature that is, by construction, part of the target. The headline findings are synthetic. We then run the pre-registered protocol on three open field corpora: on Jammertest 2024 the cross-detector prediction holds (R^2 = 0.11, p = 0.009), and on SatGrid, whose spoofer power sweep gives a calibrated severity axis, in-distribution AUC overstates higher-severity AUC by up to 0.22 and to the point of sign inversion, with in-distribution AUC and realised gap perfectly rank-correlated (Spearman rho = 1.0). The mechanism survives contact with real data, at smaller magnitude than in simulation. We release the testbed, a software-receiver front end, the ingest adapters and the protocol. --- ### Gradient-Descent Steps to Success over Mean Accuracy: A Paradigm Shift for ML Authors: Riccardo Poli, Ahmet Yilmaz Published: 2026-06-20 Categories: cs.LG, cs.AI, cs.NE arXiv: https://arxiv.org/abs/2606.22053v1 Traditional evaluation of machine learning (ML) models typically focuses on achieving the maximum possible accuracy irrespective of the computational cost. In this article, we propose a paradigm shift towards evaluating performance based on computational effort-explicitly defined here as the total number of gradient descent steps required to reach an acceptable level of accuracy with high probability. Building upon the concept of computational effort originally introduced by Koza for Genetic Programming, we extend this metric to any ML model trained via gradient descent. Furthermore, we demonstrate that minimising this effort acts as a novel form of Automatic Machine Learning (AutoML). By evaluating it across 11 diverse ML models and five standard classification datasets, we uncover significant insights into the dynamics of gradient-based learning. Our findings reveal that optimal hyper-parameters consistently favour unusually large learning rates. Crucially, we demonstrate that the rapid, aggressive landscape traversal enabled by these large rates not only promotes generalisation-as seen in phenomena like superconvergence-but also statistically minimises the expected computational effort for training. Furthermore, we identify distinct phase transitions in the optimal search strategy: while a single training run suffices for lower accuracy targets, reaching a model's performance limit requires a dramatic shift towards conducting numerous independent, short restarts. Finally, we illustrate how this effort-based paradigm provides a robust framework for model selection, allowing practitioners to choose optimal algorithms based on the difficulty of a problem as perceived by different models for a given target accuracy, or to maximise the achievable accuracy for a fixed budget of gradient descent steps. --- ### When Does a Video-Language Model Stop Watching? Reward Strength Controls the Formation and Reversal of Visual Shortcuts in Multimodal RLVR Authors: Zekun Xu Published: 2026-06-20 Categories: cs.AI, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2606.22043v1 Reinforcement learning with verifiable rewards (RLVR) is increasingly applied to large vision-language models (LVLMs), yet outcome-only optimization can drive a model to stop attending to the video and instead exploit linguistic priors -- a failure we call a visual shortcut. While the existence of such perception bypass is by now documented, how it forms, whether it can be undone, and when intervention still helps remain open. We treat the strength of a grounding penalty, lambda, as a control knob and characterize the formation-reversal dynamics of visual shortcuts along the training time axis. On a held-out, out-of-distribution diagnostic set, we find: (i) a sharp onset -- shortcut reliance emerges abruptly over a narrow window of optimization steps and is robust across random seeds; (ii) a monotone dose-response -- increasing lambda progressively suppresses the shortcut, and at an intermediate dose the trajectory first forms and then reverses the shortcut, exposing a hysteresis-like asymmetry between acquiring and removing it; and (iii) a critical intervention window -- applying the penalty before onset arrests shortcut formation, whereas the same penalty applied after consolidation is markedly less effective. Together these results recast visual-shortcut collapse not as a binary defect but as a controllable, time-dependent, and asymmetric process, with direct implications for when and how strongly to regularize multimodal RLVR. --- ### IDAG-Edit: Multi-Object Video Editing via Instance-Decoupled Attention and Guidance Authors: Yuan-Zhih Lin, Huu-Thang Nguyen, Huu-Phu Do, Hong-Han Shuai, Ching-Chun Huang Published: 2026-06-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.22042v1 Diffusion-based video editing has made significant progress; however, achieving precise and temporally consistent object-level control, especially in multi-object scenarios, remains challenging due to attention leakage, identity drift, and unstable temporal dynamics. In this work, we propose IDAGEdit, a training-free framework for fine-grained multi-object video editing with strong temporal consistency. The framework adopts Layout-guided Attention Modulation to facilitate coherent multi-object editing, while Instance-level Masks are introduced to preserve individual object identity and enforce localized attention within each object region, thereby enabling fine-grained, object-level editing. Extensive qualitative and quantitative evaluations demonstrate that our method improves temporal stability and multi-object controllability over state-of-the-art video editing approaches. --- ### Attractor Domain Theory: A Mathematical Framework for Cardiovascular Attractor Analysis with Wearable Photoplethysmography (PPG) Validation Authors: Timothy Oladunni, Farouk Ganiyu Adewumi Published: 2026-06-20 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.22039v2 The cardiovascular system evolves along a bounded trajectory in physiological state space that converges to a compact geometric object: the cardiac attractor. A wearable photoplethysmograph (PPG) or electrocardiograph (ECG) observes a one-dimensional projection of this attractor; by Takens' embedding theorem, delay coordinates reconstruct its full geometry. Three decades of nonlinear cardiac dynamics have extracted Lyapunov exponents, recurrence statistics, and sample entropy from reconstructed attractors, yet no principled account exists of which attractor properties capture which cardiovascular quantities, or why, leaving feature selection as a search problem and negative results uninterpretable. We introduce Attractor Domain Theory (ADT), which proves that the reconstructed attractor's information partitions into three mutually non-redundant domains: the Geometry Domain G (delay embedding; native capability: artifact rejection), the Ergodic Domain S (asymptotic statistical invariants; native capability: stability estimation), and the Variational Domain V (finite-time Lyapunov exponent field; native capability: hemodynamic inference). We prove a Domain Sufficiency Theorem (the Parseval analog for attractor information) and establish that three domains are necessary and sufficient. Geometry Domain validation via the SCSI framework across 176,742 PPG segments from four datasets yields AUC = 0.757 (0.686-0.828) and NPV = 0.966 after correcting three systematic evaluation artifacts (+0.179 net inflation). Ablation confirms C_NL as the dominant Geometry Domain component (Delta AUC = -0.413) and intra-domain redundancy across five components. --- ### A Completion-Aware Framework for Impactful Counterfactual Explainability in Graph Neural Networks Authors: Maria Myrto Villia, Filippos Gouidis, Theodore Patkos, Panos Trahanias Published: 2026-06-20 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.22033v1 In this study, we propose a novel pipeline for generic, model-agnostic, local-level counterfactual explainability in graph neural networks (GNNs). Although counterfactual explainers capable of both adding and removing edges have emerged in recent years, the need for generic and efficient solutions remains unmet, particularly concerning qualitative explanation generation. Our approach couples progress in factual explainability with missing edge prediction models rooted in link prediction research, in order to enhance the quality, robustness and intuitiveness of explanations. A multi-faceted experimental analysis conducted on real-world and synthetic graph classification benchmarks, both binary and multi-label, demonstrates the advancements in comparison to state-of-the-art baselines across diverse metrics. --- ### When Does Belief-Based Agent Memory Help? Reliability-Conditional Updating and Provenance-Capped Poisoning Defense Authors: Pranav Singh Published: 2026-06-20 Categories: cs.AI, cs.CL, cs.IR, cs.LG arXiv: https://arxiv.org/abs/2606.22030v2 We investigate when belief-based memory actually improves large language model (LLM) agents. Our vehicle is Nous, a long-term memory architecture that represents each entity-attribute pair as a categorical probability distribution updated through closed-form Bayesian inference, with information-theoretic surprise driving belief revision and entropy-based forgetting. A controlled ablation on the LoCoMo benchmark shows that Bayesian belief updating alone provides little benefit over naive last-write-wins because existing conversational memory benchmarks rarely contain contradictory or differently reliable evidence. We then introduce reliability-conditioned updating, estimating per-observation reliability from epistemic language, and show on a controlled contradiction benchmark that belief updating substantially outperforms last-write-wins and raw-memory retrieval when observations differ in trustworthiness. Because content-derived reliability is itself vulnerable to manipulation, we further propose provenance-capped belief updating, where trust is bounded by source provenance rather than textual confidence. Under controlled memory-poisoning experiments, this approach resists volumetric poisoning attacks while revealing the utility costs and implementation requirements of provenance-aware memory. Finally, we quantify a 27.5-point discrepancy between strict token-F1 and LLM-as-judge evaluation on identical outputs, highlighting important reproducibility concerns for long-term memory benchmarks. Our results suggest that probabilistic belief-based memory is most beneficial in environments requiring reasoning over conflicting and differently trustworthy evidence, rather than conventional conversational recall alone. --- ### Topological summaries of fingerprint ridge patterns carry identity information Authors: Chad M. Topaz, Niny Arcila-Maya, Elizabeth Munch, Zofia Stanley, Lori Ziegelmeier Published: 2026-06-20 Categories: cs.CV, stat.AP arXiv: https://arxiv.org/abs/2606.22029v1 Fingerprints are the most widely deployed biometric. Verifying whether two impressions come from the same finger typically relies on minutiae, small landmarks such as skin ridge endings and bifurcations. These landmarks are extracted through a multi-stage pipeline of image enhancement, skeletonization, minutiae detection, and alignment. We investigate an alternative: using topological data analysis to represent the full pattern of skin ridges and valleys directly, bypassing minutiae detection and the downstream matching pipeline. We apply persistent homology, a topological tool that tracks how loops in the ridge pattern form and fill in across spatial scales, producing multi-scale summaries of ridge geometry. We develop and compare a range of verification methods on a standard benchmark dataset, FVC2000 DB1. Even the simplest topological summaries, with no trained parameters, substantially outperform geometry-only baselines. A trained method achieves an AUC of 0.91, while an optimal-transport method excels at the strictest false-accept thresholds, suggesting they capture different aspects of the ridge pattern. Fusing these two approaches yields the best performance at every low false-accept threshold we examine. Our results establish that these topological summaries capture substantial fingerprint identity information, far more effective for verification than raw pixel-level geometry. Because the entire pipeline is openly specified, it offers a transparent complement to minutiae-based systems, and we provide a modular framework for constructing, evaluating, and combining topological verification methods. --- ### RARM: Confidence-Gated Progress Reward Modeling for RL in Manipulation Authors: Pengzhi Yang, Xinyu Wang, Pengyu Jing, Kehan Wen, Yiduo Qu, Zhenhao Huang, Minghao Fu, Xin Liu, Yaheng Shen, Fan Shi Published: 2026-06-20 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2606.22027v4 Reinforcement learning for robot manipulation is often bottlenecked by reward design, especially in long-horizon tasks: sparse success rewards provide weak supervision, while hand-crafted dense rewards are tedious to design and generalize poorly across tasks. Progress-based reward models offer a promising alternative by estimating how far an observation has advanced toward task completion, but existing approaches often require task-specific demonstrations or progress labels, and can assign high rewards to visually plausible but physically incorrect states. We introduce the Reference-Anchored Reward Model (RARM), a lightweight visual comparator that converts a single successful demonstration into a dense, progress-aware reward. RARM is trained once on general-purpose videos with a contrastive temporal objective, requiring no robot-specific data, task-specific reward labels, or per-task reward engineering. At deployment, RARM matches rollout clips to reference clips and rewards only confident forward progress, suppressing uncertain matches that may otherwise produce false-positive rewards. Across 9 simulated manipulation tasks from LIBERO and MetaWorld and 4 real-world tasks, RARM achieves the best overall success rates in subsequent RL training, with particularly large gains on long-horizon tasks such as cloth folding, where unreliable progress estimates are especially harmful. --- ### RARM: Confidence-Gated Progress Reward Modeling for RL in Manipulation Authors: Pengzhi Yang, Xinyu Wang, Pengyu Jing, Kehan Wen, Yiduo Qu, Zhenhao Huang, Minghao Fu, Xin Liu, Yaheng Shen, Fan Shi Published: 2026-06-20 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2606.22027v2 Reinforcement learning for robot manipulation is often bottlenecked by reward design, especially in long-horizon tasks: sparse success rewards provide weak supervision, while hand-crafted dense rewards are tedious to design and generalize poorly across tasks. Progress-based reward models offer a promising alternative by estimating how far an observation has advanced toward task completion, but existing approaches often require task-specific demonstrations or progress labels, and can assign high rewards to visually plausible but physically incorrect states. We introduce the Reference-Anchored Reward Model (RARM), a lightweight visual comparator that converts a single successful demonstration into a dense, progress-aware reward. RARM is trained once on general-purpose videos with a contrastive temporal objective, requiring no robot-specific data, task-specific reward labels, or per-task reward engineering. At deployment, RARM matches rollout clips to reference clips and rewards only confident forward progress, suppressing uncertain matches that may otherwise produce false-positive rewards. Across 9 simulated manipulation tasks from LIBERO and MetaWorld and 4 real-world tasks, RARM achieves the best overall success rates in subsequent RL training, with particularly large gains on long-horizon tasks such as cloth folding, where unreliable progress estimates are especially harmful. --- ### Cluster-Specific Localized Drift Detection for Efficient Batch Model Adaptation under Controlled Distribution Shift Authors: Ignacio Cabrera Martin, Marcello Trovati, Almas Baimagambetov, Nikolaos Polatidis Published: 2026-06-20 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.22026v1 Machine learning systems deployed in dynamic environments frequently operate under nonstationary data distributions, where controlled distribution shift can progressively degrade predictive performance. However, many widely used tabular benchmark datasets lack explicit temporal structure, limiting reproducible evaluation of drift adaptation methods. This work proposes a cluster-induced distribution shift simulation framework that transforms static tabular datasets into controlled evolving data streams through structured perturbations across featurespace partitions. Using this framework, six adaptation strategies are systematically evaluated: static learning, sliding-window retraining, global ADWIN retraining, cluster-local ADWIN retraining, random subspace drift detection, and feature-partitioned drift detection. Experiments are conducted on five benchmark datasets covering both classification and regression tasks using diverse predictive model families, including linear models, k-Nearest Neighbours, tree ensembles, boosting methods, and adaptive online learners. --- ### What Do Neural Networks Learn for TDOA Estimation? A Cross-Architecture Probing Study Authors: Yaozhong Kang, Jiang Wang, Runwu Shi, Takeshi Ashizawa, Benjamin Yen, Kazuhiro Nakadai Published: 2026-06-20 Categories: cs.SD, cs.LG arXiv: https://arxiv.org/abs/2606.22020v1 Neural networks outperform classical GCC-PHAT for Time-Difference-of-Arrival (TDOA) estimation in noise and reverberation, yet their internal strategy remains unexplored. To uncover it, we turn GCC-PHAT's mathematical steps into diagnostic targets, probing hidden layers of three architectures (MLP, CNN, Transformer) and complementing with gradient attribution and causal frequency masking. We find that cross-power computation consistently emerges across all architectures and conditions, while PHAT whitening, the defining step of GCC-PHAT, fails to emerge. Instead, networks learn a magnitude-aware frequency weighting that preserves per-frequency reliability information discarded by PHAT. This makes PHAT an information bottleneck: removing it from both classical and neural GCC pipelines improves performance under additive noise. On real-world reverberant data, PHAT remains the best classical weighting, but end-to-end networks achieve lower error by learning data-adaptive weighting. --- ### Channel Location Constrains the Auditability of Subliminal Learning Authors: Tamas Madl Published: 2026-06-20 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.22019v1 Subliminal learning lets a student inherit a teacher's hidden trait from distillation data that never names it. We ask when such transfer can be audited before training. The answer is not model identity or scale alone, but channel location: the carrier through which the trait reaches the student. We find three regimes. In a controlled initialization-dependent body channel, a pre-training screen works. Coverage, the cosine between the student's initial distillation update and the teacher's fine-tuning displacement, predicts held-out transfer (Spearman $ρ\approx 0.95$; AUROC 0.997). In pretrained language models, masked single-token traits instead ride convergent vocabulary geometry. This channel is initialization-independent, so initialization-alignment screens, including coverage, are not mechanistic; the useful handles are post-hoc detection and targeted mitigation. Even when a single-token named entity is removed from the loss, the student's held-out probability for that entity rises to 0.40 on average ($\sim 2500\times$), and a related semantic class transfers. In an untied-head model, orthogonalizing the trait's output row against entangled neighbours collapses leakage, while equal-size random-subspace edits do not. Thus removing a target string from distillation labels does not remove the corresponding preference: neighbouring tokens can carry it. Finally, conditional behaviours can route through the network body. For sycophancy, with agreement and correction markers masked from the loss, transfer reaches about 0.63 of the teacher's effect, localizes to body computation, and evades four audits across two model families. We scope this as masked transfer of a condition-present policy. Channel location is necessary for deciding which audits can be sound. It is not a deployment-ready screen: an audit used outside its carrier regime can give false assurance. --- ### Load Testing for Machine Learning Model Serving Systems at Scale Authors: Amr S. Abdelfattah, Nakul Tirumalai, Indu Mohanan, Xiao Li, Pengchao Wang, Dinakar Dhurjati, Eric Sung Published: 2026-06-20 Categories: cs.LG, cs.PF arXiv: https://arxiv.org/abs/2606.22013v1 Machine learning (ML) model serving has become a dominant consumer of GPU infrastructure, yet capacity planning in these systems remains largely ad hoc. Under-provisioning leads to service-level objective (SLO) violations and production incidents, while over-provisioning results in substantial resource waste. This paper presents \sys, an industrial load testing framework for ML serving systems that systematically estimates serving capacity through an adaptive, feedback-driven search strategy. The approach leverages real-time performance signals, incorporating dampening, spike tolerance, and convergence detection to efficiently identify maximum sustainable throughput under SLO constraints. We evaluate \sys through a longitudinal analysis of 14 industrial case studies spanning four ML architecture classes: recommendation, ranking, vision, and NLP. This study demonstrates that systematic load testing leads to substantial improvements in GPU resource efficiency and operational reliability. Prior to adopting \sys, a significant fraction of model launches were under-provisioned, resulting in recurring incidents; these issues were substantially reduced after deployment. Our results show that ML-specific design decisions are critical to accurate capacity estimation: workload calibration using recorded traffic reduces estimation error from approximately 30\% to 2--6\%, while proper warmup handling yields a 22.2\% improvement in accuracy. Further analysis reveals key factors influencing prediction error, including model size and co-location effects. This paper distills six lessons and derive architectural guidelines for ML load testing, offering actionable insights for building reliable and efficient ML serving systems. --- ### Benchmarking Large Language Models for Grapheme-to-Phoneme Conversion: A Japanese Case Study Authors: Tomoki Koriyama Published: 2026-06-20 Categories: cs.CL, eess.AS arXiv: https://arxiv.org/abs/2606.22009v1 Grapheme-to-phoneme (G2P) conversion is essential for controllable and robust text-to-speech, and large language models (LLMs), with broad linguistic knowledge, offer a promising approach. We benchmarked over 30 LLMs on Japanese G2P, comparing them with conventional morphological analyzers on 3000 manually annotated sentences. We evaluated two prompting strategies: a parse mode, where the LLM performs morphological analysis followed by rule-based kana conversion, and a direct mode, where the LLM directly predicts kana readings. The results show that model size, version, and Japanese-specialized training are key factors, with the best LLMs achieving kana character error rate below 0.52\% vs. the best conventional tool (1.03\%). Parse mode outperforms direct mode for most models, as rule-based post-processing relieves the LLM of handling complex pronunciation rules. We also show that feeding LLM-predicted kana into a kana-input TTS yields better pronunciation than end-to-end TTS. --- ### One-Shot Data Selection for Medical Image Classification via Graph Coverage Authors: Zahiriddin Rustamov, Nadia Badawi, Rafat Damseh, Nazar Zaki Published: 2026-06-20 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2606.22002v1 Training medical image classifiers on entire datasets is wasteful when annotation budgets are limited: not all samples contribute equally, yet acquiring expert labels is expensive. Active learning reduces annotation cost through iterative querying, but assumes repeated access to an oracle and requires multiple rounds of model training. One-shot geometry-based methods such as facility location avoid retraining but operate on pairwise distances that ignore the local structure of the data manifold. We propose a graph-based one-shot selection method that operates entirely on frozen foundation model embeddings. Given embeddings from a pretrained encoder, we construct a k-nearest neighbor graph over all training samples and derive a two-term coverage kernel from the heat diffusion kernel, capturing both direct and two-hop neighborhood relationships. Greedy facility location on this kernel selects class-balanced subsets that maximize coverage of the data manifold. The two-term kernel matches the full spectral heat kernel in selection behavior while reducing computation to sparse matrix operations with a single hyperparameter. We evaluate on five MedMNIST datasets spanning histopathology, radiology, and microscopy, comparing against both training-dynamics and geometry-based baselines. Our method achieves the highest balanced accuracy on nine of ten dataset-ratio conditions, with the largest gains on class-imbalanced datasets where global graph construction captures cross-class structure that per-class methods miss, all without any model training during selection. Code is available at https://github.com/zahiriddin-rustamov/graph-coverage-selection. --- ### CFAgentBench: A Reproducible Environment and Benchmark for Autonomous Construction-Finance Agents Authors: Rishi Srivastava Published: 2026-06-20 Categories: cs.AI, cs.CL, cs.LG, cs.SE arXiv: https://arxiv.org/abs/2606.22000v1 We introduce CFAgentBench, a reproducible, self-hostable environment and benchmark for autonomous construction-finance agents: a CFO/controller-class agent operating across the real software stack a US construction finance team runs - ERP, project management, email, documents, pay applications, payroll, certified payroll, lien waivers, and bank/treasury portals. It contains 1,014 machine-gradeable task specifications across 8 domains and 77 families, every family grounded in a real source; a self-validated subset of 40 tasks (54 with a project-management extension) is compiled into oracle-validated executable evaluators, the runnable suite reported here. Following WebArena, the benchmark runs on an executable environment rather than static traces: 35 mock applications (31 reconciled to one company book, plus 4 PM platforms) over 9 archetypes, each implementing a uniform self-hostable app contract, so every task is graded by functional correctness - a state diff plus forbidden-side-effect checks plus required-output regexes - with an LLM judge used only for reply quality, never as reward. A distinguishing principle is a money-movement guard: 278 instances embed a payment, payroll, e-signature, or e-filing step where the correct behavior is to stop and stage for human approval, and executing even the correct transaction fails the task. The public split (n=711) is sized for a 95% Wilson half-width of +/-4.1%; a private, contamination-protected split (n=303) is reserved for remote scoring. In a first three-model open-weight sweep (k=5), the strongest agent reaches pass^1 = 0.67 but only pass^5 = 0.38 - losing 43% of its successes when required to repeat them under temperature-0 decoding. The within-model pass^1 to pass^5 collapse and sharp per-domain heterogeneity are clear evidence that single-attempt accuracy overstates deployable construction-finance competence. --- ### Prefix-Guided On-Policy Distillation: Mining Golden Trajectories from Rollouts Authors: Qingfei Zhao, Huan Song, Shuyu Tian, Jiawei Shao, Xuelong Li Published: 2026-06-20 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.21994v2 On-policy distillation (OPD) improves reasoning models by applying dense teacher supervision on student-sampled trajectories. However, scaling OPD to long-horizon reasoning exposes a reliability and efficiency problem: standard OPD assigns every candidate the same long rollout budget, even though some trajectories may quickly become weakly aligned with the teacher and provide less useful supervision. Prior analyses suggest that teacher--student compatibility is important for OPD success, motivating early-prefix top-k overlap as a proxy for continuation value. Continuing low-overlap trajectories to the full rollout length may cause them to drift further from the teacher as generation proceeds, increasing computational cost while providing limited distillation benefit. To address this, we introduce Prefix-Guided On-Policy Distillation (PG-OPD), a rollout-allocation framework that estimates continuation value from fixed-length prefixes. PG-OPD computes teacher-student top-k overlap in an early probe window and allocates long rollouts only to high-overlap candidates, while stopping the rest at the prefix length. Across teacher--student combinations on AMC, AIME, and HMMT benchmarks, PG-OPD achieves up to a 4.80-point accuracy gain and up to a 2.46x wall-clock speedup across configurations. These results demonstrate that using early-prefix compatibility to guide candidate pruning can improve both training efficiency and reasoning performance. --- ### From Driving Videos to Simulatable Scenarios Authors: Alexandre Levy, Ernest Valveny Llobet, Antonio Manuel López Published: 2026-06-20 Categories: cs.SE, cs.CV arXiv: https://arxiv.org/abs/2606.21993v1 Autonomous vehicles (AVs) face driving scenarios ranging from routine traffic to rare events. To assess safety it is crucial to reproduce these scenarios in a controllable, repeatable, and scalable manner, with simulation playing a key role. This paper introduces D-V2S, a novel framework that automatically generates simulatable driving scenarios from driving videos. D-V2S operates in two stages: a Driving Record Analyzer (DRA) uses a vision language model (VLM) with our designed prompt to produce natural-language descriptions from input videos, capturing road layouts and dynamic traffic interactions; subsequently, a Scenario Generator (SG) uses a large language model (LLM) and our conditioning context to translate these descriptions into executable scenarios. Using simulations, we show that D-V2S generates scenarios where 90% of the relevant semantic elements of the videos are present. We also provide qualitative results demonstrating D-V2S's capability to transform real-world driving videos into simulatable scenarios. Moreover, we provide both semantic and human driven ablative analyses of D-V2S's modules. In particular, we show how the VLM choice matters for DRA, and how our SG achieves a 75% preference rate over other state-of-the-art methods. --- ### Adding Robust Code-Switching Capabilities to High Performance Multilingual ASR Authors: Enes Yavuz Ugan, Alexander Waibel Published: 2026-06-20 Categories: cs.CL, eess.AS arXiv: https://arxiv.org/abs/2606.21990v1 Code-switching (CSW) remains challenging for large multi-lingual ASR systems in real-world deployment. While fine-tuning on synthetic CSW data is possible, it generally degrades strong monolingual baselines. Our goal is to preserve these capabilities while extending models to handle complex code-switching, including morphological variations across languages. We propose Bayesian factorized adaptation, which learns to efficiently integrate switching-relevant knowledge into strong pretrained models without overwriting existing capabilities. Requiring only a small amount of synthetic data, our approach reduces transcription errors by 32.87% on code-switched words while improving overall WER by 5.31%, all while maintaining mono-lingual performance. Our results demonstrate that effective CSW adaptation depends more on knowledge integration than data complexity. --- ### CoDMD: Copula-aware Distribution Matching Distillation for Fast Video Generation Authors: Wenhu Zhang, Kun Cheng, Changyuan Wang, Shiyao Li, Yuechen Zhang, Wenbo Li, Jiajun Zha, Jingyi Zhang, Kang Zhao, Jiaya Jia Published: 2026-06-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.21982v1 Few-step distillation for video diffusion models has attracted significant attention, driven by the urgent demand for efficient deployment in real-world scenarios. However, Distribution Matching Distillation (DMD), a leading paradigm, tends to degrade under limited NFE budgets, manifesting in video generation as layout instability, oversaturation, and broken motion dynamics. We trace this failure to a structural limitation: standard DMD is an intra-sample distribution-matching objective with coordinate-wise gradients, and thus imposes no explicit constraint on the relational geometry across batch elements or temporal frames, leaving the underlying copula largely unregulated. Combined with the mode-seeking tendency of its reverse-KL objective, this absence of relational guidance makes DMD prone to collapsing into local optima in the few-step regime. Motivated by this insight, we propose Copula-aware DMD (CoDMD), a lightweight relational regularizer that reuses score estimates already produced by the frozen teacher and the online fake model to construct pairwise relation matrices across samples and frames. These are matched through a supplementary distributional objective that requires no additional networks, datasets, or sampling trajectories. On the Wan-2.1-T2V model series at 1.3B & 14B scales, CoDMD distills 50-step teachers into 4-step students, achieving an approximate 25$\times$ speed-up while attaining VBench scores of 84.46 & 84.87, outperforming prior trajectory-based (rCM 82.81 & 84.05) and distribution-based (DMD 83.38 & 83.81) methods. --- ### Can LLMs Control Readability? A Multi-Dimensional Evaluation Framework for CEFR-Controlled Arabic Generation Authors: Nour Rabih, Chatrine Qwaider, Ted Briscoe Published: 2026-06-20 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2606.21981v1 While Large Language Models (LLMs) can generate fluent Arabic text, their ability to reliably control readability levels remains unclear. We propose a multi-dimensional evaluation framework for Common European Framework of Reference for Language (CEFR)-controlled Arabic text generation, assessing whether instruction-following LLMs can serve as reliable generators for adaptive language learning. Our framework integrates controlled prompting, automatic readability prediction using a validated Taha-19 model, lexical constraint validation, and syntactic complexity profiling. Results show that structured prompting substantially improves CEFR alignment. In particular, CEFR-guided prompting with lexical constraints achieves the highest conformity to reference linguistic profiles (0.91 cosine similarity) and near-perfect agreement with predicted readability levels (0.99), while unconstrained prompting exhibits weak control. These findings establish an empirical foundation for integrating readability-aware Arabic text generation into adaptive educational systems. --- ### Old Fictions, New Skins: Evaluating the Manipulative Capabilities of LLMs in Healthcare Authors: Gathoni Ireri, Roger D. Odipo Published: 2026-06-20 Categories: cs.CY, cs.AI, cs.HC arXiv: https://arxiv.org/abs/2606.21977v1 Large language models (LLMs) are increasingly piloted in African healthcare contexts, raising concerns about their potential to manipulate users in high-stakes settings. In a randomised experiment, we examined the manipulative capabilities of two publicly available models, ChatGPT 5.2 and DeepSeek V3.2, among Kenyan participants (N = 303). Participants interacted with either a manipulative variant or a non-manipulative variant before making a treatment decision within a hypothetical clinical scenario. The manipulative variant was prompted to covertly steer participants towards an incorrect treatment option while the non-manipulative variant served as the control condition. Manipulation success rates were higher in the manipulative condition (59.5%) than in the control condition (44.0%), with the effect reaching significance (OR = 2.11, 95% CI (1.12, 4.00), p = .021). These findings highlight the need for improved safety infrastructure specifically targeting manipulation, particularly given the integration of AI into healthcare systems across Africa. --- ### IRumAI: Reinforcement Learning for Indian Rummy Authors: Vignesh Mohan Published: 2026-06-20 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.21975v1 Despite its massive player base and complex hidden-information dynamics, Indian Rummy has received no reinforcement learning attention. Existing agents rely on combinatorial search, which is tactically strong but slow at inference. We present IRumAI, the first RL agent for the domain. IRumAI integrates Proximal Policy Optimization (PPO), meld-aware observation encoding, deadwood-driven reward shaping, and a dual-branch convolutional architecture. IRumAI is RL-trained solely against weak heuristics, after a one-time behaviour-cloning warm-start on stronger demonstration data. It generalises to defeat the entire baseline hierarchy, including a 53.9% win rate against the strongest search-based opponent unseen during RL training. Bypassing explicit search, IRumAI requires just 0.33 ms per action, which is over 7,000x faster than the state-of-the-art heuristic. Ablations validate our architectural choices, and linear probing reveals that the network implicitly models the opponent's hidden hand from public interactions. --- ### Fine-Tuning Large Language Models for Quantum Reasoning Authors: Katherine Ip, Casey R. Myers, Udaya Parampalli, James Quach, Peiyong Wang Published: 2026-06-20 Categories: quant-ph, cs.AI arXiv: https://arxiv.org/abs/2606.21974v1 Large language models (LLMs) exhibit abilities beyond natural language modelling and text generation. Recent advances in their reasoning capabilities have spurred interest in applying LLMs to complex scientific tasks requiring deep domain expertise and sophisticated reasoning. Quantum computing, as a highly specialised field with significant knowledge barriers and hardware constraints, could greatly benefit from such advancements. However, a key open question that first must be answered is: How can we develop fine-tuning pipelines that instil genuine quantum reasoning in LLMs, rather than task-specific pattern matching? We study this question through quantum circuit simulation as a training objective, where the model must predict the measurement probability distribution resulting from a sequence of quantum gate operations. We propose and compare two fine-tuning pipelines: (1) Supervised Fine-Tuning (SFT) on explicit gate-by-gate state-vector simulation traces, and (2) a two-stage SFT+Group Relative Policy Optimisation (GRPO) approach that sequentially applies SFT followed by GRPO with verifiable rewards. Our findings show that SFT achieves near-perfect in-distribution and gate-count extrapolation accuracy, significantly outperforming both the base model and the GPT-OSS-120B baseline. SFT+GRPO trades some in-distribution precision for better generalisation to larger qubit systems that SFT alone cannot handle. Both pipelines significantly outperform the baselines, demonstrating that targeted fine-tuning on explicit reasoning traces is an effective strategy for advancing quantum reasoning in LLMs. --- ### SPOTR: Spatio-temporal Pooling One-Token Reconstruction for Universal Physiological Signal Self-supervised Learning Authors: Yiyu Gui, Mingzhi Chen, Yuesheng Zhu, Guibo Luo, Yuchao Yang Published: 2026-06-20 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.21973v1 Physiological signals such as EEG, ECG, and PPG are widely used in clinical monitoring. Recent self-supervised learning (SSL) methods offer an attractive way to leverage unlabeled recordings, yet they still fall short in practice. In particular, current SSL methods struggle across heterogeneous datasets, often distorting clinically meaningful structures or learning shortcuts from temporal and cross-channel redundancy. Consequently, existing SSL methods often deliver limited performance under linear probing, a lightweight adaptation setting that better matches real-world medical scenarios. Moreover, most Transformer-based SSL models encode a flattened spatiotemporal token sequence, incurring high computation and memory cost, and are typically developed within a single modality. To address these limitations, we present SPOTR (Spatio-temporal Pooling One-Token Reconstruction), a compress-reconstruct pretraining framework that introduces a single-token global bottleneck for physiological signals. SPOTR compresses each waveform into a single-token representation and reconstructs the signal conditioned only on this representation. Meanwhile, SPOTR introduces an efficient spatio-temporal compaction module to reduce computation and memory cost. Pretrained on 20 datasets spanning EEG, iEEG, ECG, and PPG, SPOTR consistently outperforms the strongest baseline under linear probing, improving average AUC by 18.49%, 21.71%, 17.86%, and 4.64%, respectively. Compared with a representative general-purpose time-series foundation model, SPOTR achieves around 78% lower latency and 52% lower peak GPU memory on average. The code can be found at https://github.com/5GYYYYY/SPOTR. --- ### Human vs Machine Mathematical Difficulty on Project Euler: An Experimental Analysis Authors: David Holmes, Johannes Schmitt Published: 2026-06-20 Categories: cs.AI, math.HO arXiv: https://arxiv.org/abs/2606.21972v1 We study how the effort and success probability of frontier AI systems scale with human difficulty on problems from Project Euler, an online platform of computational mathematics problems. Our dataset, from the MathArena benchmark, consists of 3840 attempts across 50 problems and 26 model configurations, with problem difficulty measured by the site's public human solve times. Motivated by a proposal of Timothy Gowers, we test a power-law relation $t_{\text{machine}} = a \cdot t_{\text{human}}^b$ between generated-token cost per successful answer and human time, and find $b < 1$ for 20 of the 25 models with usable fits, including the strongest base models; this operationalization therefore does not support an earlier prediction that machines scale worse than humans with difficulty. We also investigate whether success probability on the tested problems can be modeled by a simple exponential decay $p_{\text{success}} = e^{c t_{\text{human}}}$, predicting a linear relation between $\log p_{\text{success}}$ and $t_{\text{human}}$. Using a binning approach for data aggregation we find moderate empirical support (median bin-level $R^2 = 0.92$ across the 22 best-covered configurations) for this model. Following METR, we also fit logistic success curves and extract 50\% task-length horizons $h_{50}$; the strongest configurations in our 20 April 2026 snapshot reach roughly $2.5$--$4.3$ hours on our fastest-five human baseline, with a log-linear fit through the state-of-the-art frontier giving a descriptive doubling time of about $75$~days for the SOTA $h_{50}$. --- ### REBA: A Revealed Belief Automaton Framework for Online Planning in Continuous POMDPs Authors: Xiangwei Chen, Lingling Fang, Andreas Holzinger, Liming Chen Published: 2026-06-20 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.21971v1 Online planning in continuous partially observable Markov decision processes (POMDPs) using $ω$-regular specifications requires handling continuous belief dynamics within the finite symbolic memory in order to track temporal progress. Existing methods based on either direct search in belief space or predefined discrete abstractions suffer from drawbacks, e.g., lack of symbolic memory for long-horizon logical progress or difficult to certify from noisy online beliefs. As such, obtaining reliable symbolic states online from continuous observations remains a challenge. To address this issue, we introduce the Revealed Belief Automaton (REBA), an event-driven framework that advances the research from global belief-space discretization to a fundamental new way of thinking, namely online certification of revelation events. Specifically, we propose an online revelation method that, through information-theoretic gates, can dynamically analyse and establish belief abstraction from the continuous belief space by discovering reliable anchors among noisy beliefs. We then develop an incremental topology adaptation mechanism over the certified anchors to realise the online finite Belief Automaton. By combining with the $ω$-regular specification, REBA is able to support formal parity policy synthesis without a predefined discrete abstraction, which in turn can guide the Monte Carlo Tree Search process to perform online search beyond its local horizon. In addition, we design an error decomposition analysis which can assess the effectiveness and reliability of this discrete guidance for the underlying continuous POMDP. Empirical evaluations in patrolling and navigation scenarios show that REBA matches or exceeds all evaluated baselines, with primary metric gains of +17.0\% to +47.4\% over state-of-the-art approaches. --- ### Integrating Facial Generation into Full-Duplex Spoken Dialogue Systems Authors: Jingjing Jiang, Atsumoto Ohashi, Ryuichiro Higashinaka Published: 2026-06-20 Categories: cs.HC, cs.CL, cs.CV, eess.AS arXiv: https://arxiv.org/abs/2606.21970v1 Full-duplex spoken dialogue models, such as Moshi, enable natural, low-latency voice conversations. However, they remain limited to the audio modality, lacking the facial expressions that are integral to human communication. We present Moshi-Face, the first full-duplex dialogue model that jointly processes the user's audio and facial input while simultaneously generating speech and facial motion. We first construct a vector-quantized variational autoencoder (VQ-VAE) as a face codec that encodes 3D head meshes extracted from facial videos into compact discrete tokens, referred to as face tokens, and conversely reconstructs 3D meshes from these tokens. We then extend Moshi with a Face Transformer module that generates face tokens non-autoregressively, enabling Moshi-Face to produce synchronized audio and face tokens in real time. Experiments show that Moshi-Face achieves audiovisual alignment at low latency while preserving the dialogue quality of the original audio-only model. --- ### Look Before You Zoom: Adaptive Routing for the Resolution-Context Trade-off in Visual RAG Authors: Oanh N. Tran, Thanh Quoc Hung Le, Oscar Chew, Kuan-Hao Huang, Khoa D. Doan Published: 2026-06-20 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2606.21968v1 Vision-Language Models (VLMs) struggle as query-relevant objects become smaller. To address this, recent training-free approaches dynamically retrieve and zoom into local image regions. However, we show that indiscriminately applying retrieval ignores a critical vulnerability: the resolution-context trade-off. Patch-based zooming recovers details for small targets, but can split large objects and destroy global spatial context; attention-based retrieval better preserves large objects, but remains less reliable on tiny details; and global perception is often fastest when retrieval is unnecessary. Motivated by these failure modes, we introduce ViRGo (Visual Retrieval or Global Perception), a lightweight framework that formulates visual retrieval as an adaptive routing problem. ViRGo estimates object scale from the VLM's intrinsic localization heads during the initial forward pass and combines it with semantic token confidence to select between global perception, patch-based retrieval, and attention-based retrieval with minimal additional computation. Experiments across multiple VQA benchmarks and object-size groups show that ViRGo improves the accuracy-efficiency trade-off: it matches patch retrieval on small details, leverages attention-based retrieval for larger objects, and reduces inference time by routing to the global baseline when zooming is unnecessary. --- ### Holmes: Multimodal Agentic Diagnosis for Mixed-Language Mobile Crashes at Industrial Scale Authors: Jia Li, Wenyuan Ma, Ting Peng, Haibin Zheng, Yuetang Deng Published: 2026-06-20 Categories: cs.AI, cs.SE arXiv: https://arxiv.org/abs/2606.21963v1 Diagnosing mobile crashes in ultra-large-scale industrial applications is a formidable challenge due to the sheer volume of code, the complexity of mixed-language environments, and the inability to reproduce failures locally. Traditional static analysis struggles with scalability, while existing LLM-based agents often rely on reproducible environments unavailable in post-mortem scenarios. We present Holmes, a multi-agent system that automates root cause analysis by synthesizing multimodal runtime signals--stack traces, logs, and thread states--to reconstruct failure contexts without reproduction. Holmes introduces a hierarchical Retrieve-Explore-Reason architecture that leverages low-level artifacts (e.g., registers, assembly) to bridge the semantic gap between open-source business logic and closed-source system frameworks. By dynamically compressing the search space using runtime clues, Holmes precisely navigates 70-million-line codebases to identify non-local defects. Evaluated on real-world crashes from WeChat, Holmes achieves 87.6% accuracy in function-level fault localization and reduces average investigation time by over 98% (to ~77 seconds), demonstrating its effectiveness in transforming labor-intensive debugging into an efficient verification workflow. --- ### VegSim: A Geospatial World Model for Scenario-Conditioned Vegetation Simulation Authors: Irene Iele, Elena Mulero Ayllón, Paolo Soda, Matteo Tortora Published: 2026-06-20 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.21961v2 Vegetation monitoring under climate stress requires answering not only how it will evolve given the expected weather, but how it would respond to alternative meteorological conditions. Forecasting models return the expected vegetation state for the observed weather and cannot answer these scenario-conditioned questions, because future weather is fixed to the recorded trajectory. We present VegSim, a geospatial world model for scenario-conditioned vegetation simulation. VegSim infers a latent vegetation state from sparse satellite-derived NDVI histories, past meteorological covariates, and static spatial context, propagates it forward under future weather forcing through recurrent latent dynamics, and decodes predictive NDVI quantiles at each lead time. Because future forcing enters as a controllable input, the same trained model supports probabilistic forecasting under observed weather and conditional simulation under user-defined meteorological forcing, without supervision on scenario responses. We evaluate VegSim on GreenEarthNet across in-distribution data and spatial, temporal, and joint spatio-temporal shifts, where it achieves strong point and probabilistic accuracy against time series and Earth Observation forecasting baselines while using a compact architecture. We then simulate vegetation responses across Europe under two contrasting meteorological scenarios, obtaining a weak but broadly positive response under winter wet warming and a widespread negative response under summer drought. The code is available at https://github.com/arco-group/vegsim. --- ### VegSim: A Geospatial World Model for Scenario-Conditioned Vegetation Simulation Authors: Irene Iele, Elena Mulero Ayllón, Paolo Soda, Matteo Tortora Published: 2026-06-20 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.21961v1 Vegetation monitoring under climate stress requires answering not only how it will evolve given the expected weather, but how it would respond to alternative meteorological conditions. Forecasting models return the expected vegetation state for the observed weather and cannot answer these scenario-conditioned questions, because future weather is fixed to the recorded trajectory. We present VegSim, a geospatial world model for scenario-conditioned vegetation simulation. VegSim infers a latent vegetation state from sparse satellite-derived NDVI histories, past meteorological covariates, and static spatial context, propagates it forward under future weather forcing through recurrent latent dynamics, and decodes predictive NDVI quantiles at each lead time. Because future forcing enters as a controllable input, the same trained model supports probabilistic forecasting under observed weather and conditional simulation under user-defined meteorological forcing, without supervision on scenario responses. We evaluate VegSim on GreenEarthNet across in-distribution data and spatial, temporal, and joint spatial-temporal shift, where it achieves strong point and probabilistic accuracy against time series and Earth observation forecasting baselines while using a compact architecture. We then simulate vegetation responses across Europe under four meteorological scenarios, and in a France summer 2022 case study, obtaining spatially coherent patterns consistent with known sensitivity to temperature and precipitation. The code is available at https://github.com/arco-group/vegsim. --- ### A Standard Processing Pipeline for High-accuracy Measurement of Few-shot Regression on Laser Induced Breakdown Spectroscopy Authors: Hao Li Published: 2026-06-20 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.21960v1 Laser-induced breakdown spectroscopy (LIBS) faces challenges in high-accuracy quantitative measurement under few-shot scenarios due to spectral noise and data scarcity. Traditional preprocessing methods often fail to preserve subtle spectral features or capture nonlinear correlations. This work proposes a standardized processing pipeline integrating diffusion-based denoising, attention-based autoencoder for dimensionality reduction, group shuffling data augmentation, and ordinary least squares regression. The diffusion module employs a 3D UNet architecture to remove spectral noise while preserving essential emission features. The attention-autoencoder captures nonlinear spectral correlations, effectively reducing high-dimensional spectral data to compact latent representations. Group shuffling data augmentation enhances model robustness by creating synthetic samples through feature group permutation. Experimental results on multiple elemental concentrations demonstrate that our Diffusion-DA-AE pipeline achieves superior performance with a mean RMAE of 0.2847, representing 37.7\% and 37.6\% improvements over baseline autoencoder and traditional PCA-PLS regression, respectively. The framework's effectiveness validates its generalizability and establishes a new benchmark for few-shot LIBS regression. --- ### OpenBioRQ: Unsolved Biomedical Research Questions for Agents Authors: Minbyul Jeong Published: 2026-06-20 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.21959v1 A working citation looks like proof -- but the fact that a link resolves does not mean the cited paper supports the claim. I find that current agentic models rarely fabricate citations (over $99\%$ resolve), yet roughly $15.9\%$ link to the wrong paper. Existing benchmarks miss this failure mode: when a question has a fixed answer key, a model can reproduce the expected source from that key rather than independently verifying that the source supports the claim. I introduce \textbf{\openbiorq{}}, a retrieval-grounded agentic benchmark of $12{,}553$ unsolved biomedical research questions across $12$ domains that treats open questions as a faithfulness-and-abstention probe. To my knowledge, this is the first biomedical benchmark to combine an agentic setting -- where the model must issue multiple tool calls -- with unsolved questions that have no answer key. Openness is verified against real follow-up evidence rather than a model's parametric knowledge. Difficulty is empirical: I anchor it on questions that three open-weight reference models fail to answer, rather than on subjective hardness labels. On this hardest subset, held-out models from the same lineage as the difficulty anchors solve only ~17%, while three independent frontier agents (Gemini-3-Pro, Opus-4.7, GPT-5.5) span a wide 29-60% range. The benchmark is thus hard, non-saturating (the best agent still leaves ~33-40\% unsolved), and discriminating across capability tiers. Beyond difficulty, I observe agentic collapse on the hardest questions, where agents stop using their tools. For the most collapse-prone model, blocking tool access entirely barely changes its score -- so tools stop paying off exactly where they are needed most. A frozen per-question checklist raises inter-judge agreement from Spearman 0.35 to 0.82. --- ### Learning by Shifting: Temporal View Construction for Time Series Contrastive Learning Authors: Abdul-Kazeem Shamba, Kerstin Bach, Gavin Taylor Published: 2026-06-20 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.21957v1 Supervised learning demands large quantities of labeled data, a bottleneck that is expensive and reliant on domain-specific expertise. Self-supervised learning, particularly contrastive learning, has emerged as a compelling alternative, enabling rich representation learning directly from unlabeled data. Yet its success hinges critically on the design of positive and negative sample pairs. Existing approaches for time series rely on hand-crafted augmentations and masking heuristics that embed strong domain assumptions, often limiting generalization across diverse temporal patterns and potentially introducing spurious correlations. In this work, we challenge this paradigm by demonstrating that explicitly encoding temporal shift invariance through a simple, deterministic view construction is sufficient to learn strong representations for time series classification. By exploiting temporal structure, our method, Shift Invariant Feature Training (ShiFT), achieves state-of-the-art performance on six diverse real-world time series benchmark datasets, as well as the UCR and UEA archives, while reducing training time. Beyond empirical performance, we present a systematic analysis of contrastive learning dynamics in time series settings, examining the effects of batch size and the number of negatives on downstream performance. Our findings provide practical insights for designing efficient contrastive learning frameworks for time series representation learning. The source code is publicly available at https://github.com/sfi-norwai/ShiFT. ---