arXiv:2604.01741v2 Announce Type: replace Abstract: Modeling users' cognitive states (e.g., cognitive load and decision confidence) is essential for building adaptive AI in high-stakes decision-making. While eye tracking provides non-invasive behavioral signals correlated with cognitive effort, prior work has not systematically examined how AI assistance contexts, specifically varying advice reliability and user heterogeneity, can alter the mapping between gaze signals and cognitive states. We conducted a within-subject lab eye-tracking study (N=54) on factual verification tasks under three conditions: No-AI, Correct-AI advice, and Incorrect-AI advice. We analyze condition-dependent changes in self-reports and eye-tracking patterns and evaluate the robustness of eye-tracking-based user modeling. Results show that AI advice increases decision confidence compared to No-AI, while Correct-AI is associated with lower perceived cognitive load and more efficient gaze behavior. Crucially, predictive modeling is context-sensitive: the relationship between eye-tracking signals and cognitive states shifts across AI conditions. Finally, fusing eye-tracking features with user priors (demographics, AI literacy/experience, and propensity to trust technology) improves cross-participant generalization. These findings support condition-aware and personalized user modeling for cognitively aligned adaptive AI systems.
Science Journals
arXiv:2604.01958v2 Announce Type: replace Abstract: Infrared and visible video fusion combines the object saliency from infrared images with the texture details from visible images to produce semantically rich fusion results. However, most existing methods are designed for static image fusion and cannot effectively handle frame-to-frame motion in videos. Current video fusion methods improve temporal consistency by introducing interactions across frames, but they often require high computational cost. To mitigate these challenges, we propose MAVFusion, an end-to-end video fusion framework featuring a motion-aware sparse interaction mechanism that enhances efficiency while maintaining superior fusion quality. Specifically, we leverage optical flow to identify dynamic regions in multi-modal sequences, adaptively allocating computationally intensive cross-modal attention to these sparse areas to capture salient transitions and facilitate inter-modal information exchange. For static background regions, a lightweight weak interaction module is employed to maintain structural and appearance integrity. By decoupling the processing of dynamic and static regions, MAVFusion simultaneously preserves temporal consistency and fine-grained details while significantly accelerating inference. Extensive experiments demonstrate that MAVFusion achieves state-of-the-art performance on multiple infrared and visible video benchmarks, achieving a speed of 14.16 FPS at $640 \times 480$ resolution. The source code will be available at https://github.com/ixilai/MAVFusion.
arXiv:2604.03107v2 Announce Type: replace Abstract: We propose a new scintillator-based tracker concept based on a monolithic plastic scintillator plate with embedded scatterers and wavelength-shifting fiber readout. The embedded scatterers localize scintillation light so that channels closer to the charged-particle crossing point collect more light. The particle crossing position is reconstructed from the channel-to-channel light yield distribution with a position resolution well below the readout pitch. We performed a positron beam test with prototypes to validate the reconstruction principle and to evaluate the detection efficiency and position resolution. The beam test validated the position reconstruction principle, and demonstrated a near-100% detection efficiency and a position resolution of 1.47 mm for normal incidence and 1.85 mm for an incidence angle of 45{\deg}, with a 10-mm readout pitch. In this paper, we describe the detector concept, the reconstruction method, and the results of the beam test.
arXiv:2604.04535v2 Announce Type: replace Abstract: Modern machine learning systems, such as generative models and recommendation systems, often evolve through a cycle of deployment, user interaction, and periodic model updates. This differs from standard supervised learning frameworks, which focus on loss or regret minimization over a fixed sequence of prediction tasks. Motivated by this setting, we revisit the classical model of learning from equivalence queries, introduced by Angluin (1988). In this model, a learner repeatedly proposes hypotheses and, when a deployed hypothesis is inadequate, receives a counterexample. Under fully adversarial counterexample generation, however, the model can be overly pessimistic. In addition, most prior work assumes a \emph{full-information} setting, where the learner also observes the correct label of the counterexample, an assumption that is not always natural. We address these issues by restricting the environment to a broad class of less adversarial counterexample generators, which we call \emph{symmetric}. Informally, such generators choose counterexamples based only on the symmetric difference between the hypothesis and the target. This class captures natural mechanisms such as random counterexamples (Angluin and Dohrn, 2017; Bhatia, 2021; Chase, Freitag, and Reyzin, 2024), as well as generators that return the simplest counterexample according to a prescribed complexity measure. Within this framework, we study learning from equivalence queries under both full-information and bandit feedback. We obtain tight bounds on the number of learning rounds in both settings and highlight directions for future work. Our analysis combines a game-theoretic view of symmetric adversaries with adaptive weighting methods and minimax arguments.
arXiv:2604.08116v2 Announce Type: replace Abstract: In the last decades, energy-based models (EBMs) have become an important class of probabilistic models in which a component of the likelihood is intractable and therefore cannot be evaluated explicitly. Consequently, parameter estimation in EBMs is challenging for conventional inference methods. In this work, we provide a unified framework that connects noise contrastive estimation (NCE), reverse logistic regression (RLR), multiple importance sampling (MIS), and bridge sampling within the context of EBMs. We further show that these methods are equivalent under specific conditions. This unified perspective clarifies relationships among existing methods and enables the development of new estimators, with the potential to improve statistical and computational efficiency. Furthermore, this study helps elucidate the success of NCE in terms of its flexibility and robustness, while also identifying scenarios in which its performance can be further improved. Hence, rather than being a purely descriptive review, this work offers a unifying perspective and additional methodological contributions. The MATLAB code used in the numerical experiments is also made freely available to support the reproducibility of the results.
arXiv:2604.08350v2 Announce Type: replace Abstract: We introduce a systematic hierarchy of one-body Green's function methods derived from the $GW$ approximation, constructed by progressively reducing the dynamical content of the self-energy. Starting from the fully dynamical Dyson formulation, we generate a family of approximations that interpolates between the standard $GW$ approximation to purely static effective single-particle Hamiltonians. This framework enables a controlled investigation of the role of dynamical effects and particle-hole coupling in the description of ionization potentials. Within this unified formalism, the hole and particle branches can be selectively decoupled through downfolding strategies into reduced one-particle spaces. By benchmarking the different members of this hierarchy on molecular ionization energies, we assess their accuracy, numerical robustness, and algorithmic complexity. We demonstrate that consistently derived partially static schemes can yield reliable quasiparticle energies while significantly simplifying the underlying eigenvalue problem. We further introduce a novel static Hermitian self-energy obtained as the static limit of this hierarchy. Despite its conceptually distinct origin, it produces results remarkably close to those of qs$GW$, thereby providing an alternative static route toward partial self-consistency.
arXiv:2604.09237v2 Announce Type: replace Abstract: Many disciplines pose natural-language research questions over large document collections whose answers typically require structured evidence, traditionally obtained by manually designing an annotation schema and exhaustively labeling the corpus, a slow and error-prone process. We introduce ScheMatiQ, which leverages calls to a backbone LLM to take a question and a corpus to produce a schema and a grounded database, with a web interface that lets steer and revise the extraction. In collaboration with domain experts, we show that ScheMatiQ yields outputs that support real-world analysis in law and computational biology. We release ScheMatiQ as open source with a public web interface, and invite experts across disciplines to use it with their own data. All resources, including the website, source code, and demonstration video, are available at: www.ScheMatiQ-ai.com
arXiv:2604.15877v2 Announce Type: replace Abstract: As LLM agents scale to long-horizon, multi-session deployments, efficiently managing accumulated experience becomes a critical bottleneck. Agent memory systems and agent skill discovery both address this challenge, extracting reusable knowledge from interaction traces, yet a citation analysis of 1{,}136 references across 22 primary papers reveals a cross-community citation rate below 1\%. We propose the \emph{Experience Compression Spectrum}, a unifying framework that positions memory, skills, and rules as points along a single axis of increasing compression (5--20$\times$ for episodic memory, 50--500$\times$ for procedural skills, 1{,}000$\times$+ for declarative rules), directly reducing context consumption, retrieval latency, and compute overhead. Mapping 20+ systems onto this spectrum reveals that every system operates at a fixed, predetermined compression level: none supports adaptive cross-level compression, a gap we term the \emph{missing diagonal}. We further show that specialization alone is insufficient (both communities independently solve shared sub-problems without exchanging solutions), that evaluation methods are tightly coupled to compression levels, that transferability increases with compression at the cost of specificity, and that knowledge lifecycle management remains largely neglected. We articulate open problems and design principles for scalable, full-spectrum agent learning systems.
arXiv:2604.19227v2 Announce Type: replace Abstract: We introduce SignatureTensors.jl, a new package for computing signature tensors of paths in julia. We present its core functionality and demonstrate its use through illustrative examples. The package is compatible with the computer algebra system OSCAR, enabling both exact and numerical computations with signatures.
arXiv:2605.10152v4 Announce Type: replace Abstract: In this paper, we present a learning-based control for a class of nonlinear systems that guarantees exponential stability as well as bounded output errors. The control is based on the Gaussian Process Submodel Online Learning (GPSOL) algorithm and the Disturbance Error Rate Limiting (DERL) algorithm, both of which were developed in previous work. The GPSOL algorithm provides a method to learn Gaussian Process (GP) models for subsystems online, whereas the DERL algorithm allows to limit the rate of the prediction error of these GP models. The focus of this paper is the utilization of the GP model within an adaptive controller and the derivation of corresponding stability conditions and system peak-to-peak gains by means of linear matrix inequalities (LMIs). These peak-to-peak gains are then used to prescribe a desired prediction error rate for the DERL algorithm to achieve user-defined output error bounds. The gains and the related bounds were successfully verified using a simulation model. Furthermore, results form a successful experimental validation of the bounds and the overall control structure on a pneumatic test rig are presented. While the control scheme and error bounds proposed in this paper are limited to first-order single-input-single-output systems, an extension to certain classes of higher-order and multiple-input-multiple-output systems is expected to be forthcoming.
arXiv:2605.19874v3 Announce Type: replace Abstract: Contributions from connected quintuple excitations in coupled cluster theory can reach the 0.5 kcal/mol range, important enough to matter in accurate computational thermochemistry, yet the very steep $\propto N^{12}$ CPU time scaling impedes routine evaluation. We show that for the differential contribution of quintuples, convergence of a frozen natural orbital (FNO) expansion with respect to the NO cutoff is rapid enough to make FNO-CCSDTQ(5)$_\Lambda$ with cutoffs of 0.0025 or 0.001 viable alternatives. A naive extrapolation to zero cutoff from \{0.005,0.0025\} works surprisingly well as a low-cost option. Interestingly, FNO convergence is definitely slower for second-row than for first-row compounds.
arXiv:2605.20919v3 Announce Type: replace Abstract: Sutra is a typed, purely functional programming language whose compiled forward pass is a PyTorch neural network. The compiler beta-reduces the whole program -- primitives, control flow, string I/O -- to one fused tensor-op graph over a frozen embedding substrate. Rotation binding, unbind, bundle, polynomial Kleene three-valued logic, and tail-recursive loops all lower to tensor operations; the Kleene connectives are Lagrange-interpolated polynomials exact on the {-1, 0, +1} truth grid. Validation is one fact tested two ways. (1) The same program runs on four frozen embeddings spanning two modalities -- three text encoders (nomic-embed-text, all-minilm, mxbai-embed-large) and one protein language model (ESM-2) -- and decodes bundles at 100% accuracy through width k=8 on every substrate, where the textbook Hadamard product has already collapsed (2.5% on mxbai-embed-large, 7.5% on all-minilm). (2) PyTorch autograd flows through the actually compiled graph: a fuzzy-rule classifier written in .su trains from random init (18.7 +/- 9.5%; chance = 20%, five classes) to 100.0 +/- 0.0% (three seeds) by backpropagating through the emitted graph, the symbolic source unmodified. A weighted variant additionally trains a scalar cosine gain and writes it back into the .su source as a numeric literal; recompiling reproduces the trained behaviour to ~2e-7 per logit, so the trained model is itself legible, recompilable code. The same artifact is therefore both a logic program and a trainable neural network.
arXiv:2605.24316v2 Announce Type: replace Abstract: Scaling laws provide compact descriptions of how prediction error varies with compute, model size, and data, but existing theory mainly treats single-sample SGD or full data reuse, leaving the role of mini-batching unclear. We study batch scaling laws for sketched linear regression under a power-law covariance spectrum and a source condition on the target parameter. We analyze one-pass batch SGD, multi-pass batch SGD with replacement, and multi-pass batch SGD without replacement. Our first result is a risk decomposition: all three procedures share the same irreducible and approximation terms, while their stochastic terms depend on the sampling protocol. One-pass batch SGD splits into bias and variance, whereas the two multi-pass methods split into GD bias, GD variance, and a fluctuation term around a common GD reference trajectory. We then prove source-condition scaling laws for one-pass and multi-pass mini-batch methods. For one-pass batch SGD, mini-batching preserves the approximation and optimization-bias exponents, while the variance scales as $O(\min(M,(T_{\mathrm{eff}}\gamma)^{1/a})/(B T_{\mathrm{eff}}))$. Thus the usual $1/B$ covariance reduction holds at fixed update count $T$, but in the one-pass regime $T=N/B$ it is partly offset by the shorter optimization horizon. For multi-pass batch SGD, with- and without-replacement sampling have identical approximation and GD bias/variance terms; they differ only in the fluctuation covariance prefactor, which is $1/B$ with replacement and $\rho_{N,B}=(N-B)/(B(N-1))$ without replacement. Hence without-replacement sampling is less noisy for $B>1$, and when $B=N$ the fluctuation vanishes, recovering deterministic gradient descent. These results place batch size on the same theoretical footing as compute, data, and model dimension in sketched linear regression.
arXiv:2605.24417v2 Announce Type: replace Abstract: Supervised classification on tabular data remains a central machine learning task, but its dependence on large labeled datasets limits its applicability in data-scarce settings. Few-shot methods such as TabPFN achieve strong performance through large-scale synthetic pretraining, yet still require labeled context examples. Large Language Models (LLMs) offer a more flexible alternative through zero- and few-shot in-context learning from task descriptions, but their behavior on tabular data remains inconsistent. We introduce LLMTabBench, a benchmark for evaluating LLMs on tabular classification under low-data conditions. The benchmark studies how LLM prior knowledge interacts with task descriptions and few-shot examples, and how performance changes with increasing data complexity across real-world and controlled synthetic datasets. We find that LLMs can be highly competitive in zero-shot settings, sometimes outperforming models given few-shot examples. However, additional examples may conflict with prior knowledge, thereby degrading performance. We also observe a complexity threshold at which LLM performance declines and few-shot examples become less useful. These results clarify key limits of in-context learning for tabular data and inform the deployment of LLMs in low-data regimes.
arXiv:2605.25613v2 Announce Type: replace Abstract: This paper presents a Jacobi-type iteration for computing a given specified eigenpair of a symmetric matrix. For a certain class of diagonally dominant matrices, the procedure is shown to converge at a linear rate depending on how the matrix is significantly dominated. The cost per iteration is generally quadratic. Therefore, the proposed procedure can compute an approximation of the desired eigenpair in quadratic time.
arXiv:2605.28091v2 Announce Type: replace Abstract: Text-to-Image generation has evolved from basic image synthesis into a frequently used core capability in professional creative workflows, where simple text-image alignment can no longer satisfy users' pressing demands for faithful real-world reconstruction and genuine creative expression. Existing benchmarks, however, remain anchored in these foundational criteria and do not yet capture the nuanced capabilities that matter in authentic artistic practice, making it difficult to reliably distinguish state-of-the-art T2I models. To address the gap, we introduce Qwen-Image-Bench, a creator-centric benchmark co-designed with professional artists and grounded in real-world creation scenarios. Qwen-Image-Bench enriches conventional evaluation with two application-driven dimensions: Real-world Fidelity and Creative Generation. Drawing on the staged reasoning inherent in professional artistic workflows, we organize these five pillars into a top-down hierarchical taxonomy that further decomposes into 23 second-level sub-capabilities and 56 third-level verifiable rubrics. To ensure broad coverage, we curate 1000 stratified prompts with each prompt jointly exercising more than four fine-grained facets across multiple pillars. We train a unified judge model Q-Judger based on Qwen3.6-27B, supervised by 80 professional annotators from global art academies under blind labeling and triple-review protocols, that scores every image across all 56 verifiable facets, producing fine-grained, rubric-grounded, and fully attributable diagnostics rather than a single opaque score. Empirically, Qwen-Image-Bench reliably distinguishes leading T2I models, achieving the greatest separation on the two application-driven dimensions of Real-world Fidelity and Creative Generation where existing benchmarks provide little insight, while also providing a trustworthy optimization signal for production-level T2I development.
arXiv:2606.12716v2 Announce Type: replace Abstract: The integration of Large Language Models (LLMs) and Multimodal LLMs (MLLMs) into scientific peer-review workflows introduces novel and significant risks for adversarial manipulation, especially given the multimodal nature of scientific papers where figures, not just text, convey core evidence. This creates a significant gap: current robustness studies on AI peer-review are overwhelmingly text-only. Moreover, the problem is distinct from standard jailbreaking, as a peer-review attack seeks to induce a domain-specific, targeted failure (e.g., "inflate this score") rather than a general safety policy violation, for which no practical defenses exist. To address this, we introduce PaperGuard, the first comprehensive benchmark designed to systematically evaluate and defend AI-generated peer-review against these domain-specific, cross-modal attacks. Our framework is built on three pillars: (1) a new multimodal peer-review dataset spanning multiple scientific domains; (2) a unified suite of attacks, including black-box prompt injections and white-box perturbations, specifically designed to target both text (GCG) and figures (PGD); and (3) a practical defense, motivated by the long-context challenge of academic papers, that uses chunk-based embedding search to efficiently localize and mitigate harmful instructions. Our extensive experiments, conducted across state-of-the-art models, confirm that AI reviewers are pervasively vulnerable. PaperGuard establishes the foundational benchmark, protocols, and actionable defense necessary to pioneer trustworthy, attack-resilient AI-assisted scholarly reviewing.
arXiv:2606.13042v2 Announce Type: replace Abstract: In intelligent video surveillance, cameras record image sequences during day and night. Commonly, this demands different sensors. To achieve a better performance it is not unusual to combine them. We focus on the case that a long-wave infrared camera records continuously and in addition to this, another camera records in the visible spectral range during daytime and an intelligent algorithm supervises the picked up imagery. More accurate, our task is multispectral CNN-based object detection. At first glance, images originating from the visible spectral range differ between thermal infrared ones in the presence of color and distinct texture information on the one hand and in not containing information about thermal radiation that emits from objects on the other hand. Although color can provide valuable information for classification tasks, effects such as varying illumination and specialties of different sensors still represent significant problems. Anyway, obtaining sufficient and practical thermal infrared datasets for training a deep neural network poses still a challenge. That is the reason why training with the help of data from the visible spectral range could be advantageous, particularly if the data, which has to be evaluated contains both visible and infrared data. However, there is no clear evidence of how strongly variations in thermal radiation, shape, or color information influence classification accuracy. To gain deeper insight into how Convolutional Neural Networks make decisions and what they learn from different sensor input data, we investigate the suitability and robustness of different augmentation techniques...
arXiv:2606.13755v2 Announce Type: replace Abstract: We argue that aligning AI to aggregated human preferences is the wrong target. With current technology, one can train AIs to share the values of a Silicon Valley techno-optimist, a degrowth environmentalist, a national-conservative culture warrior, a single-party state cadre, or a devout religious traditionalist. We should not. Human values produce societies that thrive or fail on the merits of those values - from failed states and extreme inequality to declining happiness, political polarization, and government dysfunction in the world's wealthiest democracies. The pluralistic-alignment program correctly diagnoses that there is no single "humanity" to align with, but is dangerous if taken as the main directive. We argue that AI should be trained to a non-negotiable floor of objective alignment goals - competence, bounded by the constraints of factual accuracy, honesty, and lawfulness and that pluralism belongs at the surface (language, register, conventions, missing-context defaults) and across the wide band of legitimate value tradeoffs that respect the floor, but not at the level of values that violate it. We highlight the empirical reality of unfiltered pluralistic values, propose four commitments as a constructive alternative, and engage six credible objections: commercial pressure and practical feasibility, democratic legitimacy, regulatory compliance, over-reliance on institutionalist explanations, the charge that the floor itself is culturally laden, and the limits of Coherent Extrapolated Volition.
arXiv:2606.16203v3 Announce Type: replace Abstract: Post-traumatic epilepsy (PTE) is a severe complication of traumatic brain injury (TBI). Yet, early identification remains challenging due to the complex structural and functional alterations it induces in the brain. To address this, we propose a dynamic multimodal Mixture-of-Experts (MoE) framework that integrates functional and structural connectivity through time-aware functional-structural encoding and class-conditioned expert routing. Within this framework, modality-specific and cross-modal experts learn complementary representations, while a Modality-Class MoE (MCoE) module dynamically adjusts expert weights according to each classification objective. Experimental results across three binary classification tasks demonstrate that the framework consistently outperforms static fusion baselines, and high-interpretability analyses further reveal meaningful regions of interest (ROIs) interactions. This dynamic multimodal expert framework effectively captures class-dependent brain interaction patterns and provides an interpretable approach for PTE diagnosis and risk stratification.
PAMAE: Phase-Aware-MoE Action Experts Towards Reliable Flow-Matching Vision-Language-Action Policies
arXiv:2606.27144v1 Announce Type: new Abstract: Reliable action generation for multi-stage robotic manipulation remains challenging for Vision-Language-Action (VLA) models. While existing flow-matching VLA policies offer strong multimodal grounding and generalization, they typically employ a single shared action expert, limiting their ability to capture phase-specific control patterns across distinct execution stages. We propose a plug-and-play Phase-Aware Mixture-of-Experts Action Module (PAMAE), as a step towards more reliable phase-consistent action generation. PAMAE replaces the original flow-matching action expert with a sparse expert mixture while preserving the pretrained VLA backbone. PAMAE introduces a phase-aware router that leverages execution-phase cues to allocate action generation across experts, supported by a lightweight phase prediction head and a routing alignment objective. To stabilize specialization, we adopt a two-stage training scheme that first warms up the expert module under the standard flow-matching loss and then optimizes phase-consistent routing under auxiliary supervision. On multi-stage manipulation simulation tasks, PAMAE improves task success by up to \textbf{9.2\%} over strong VLA baselines. Further ablations show that both phase-supervised routing and staged optimization are essential for the observed gains. Our results highlight phase-consistent expert allocation as an effective mechanism for improving the reliability and action quality of flow-matching VLA policies.
arXiv:2606.16325v2 Announce Type: replace Abstract: Few-shot medical image segmentation methods typically assume a single ground-truth annotation, overlooking systematic variability across expert raters commonly observed in clinical datasets. We propose an attention-based prototype calibration framework for few-shot multi-rater segmentation that models rater-specific deviations from a consensus representation in prototype space. A lightweight yet principled attention operator directly refines rater prototypes without modifying the backbone feature extractor, making the approach fully compatible with existing prototype-based few-shot segmentation methods. This design preserves semantic consistency while enabling personalized segmentation outputs with minimal computational overhead. Experiments on multi-rater medical imaging datasets demonstrate consistent improvements over baseline prototype approaches, highlighting the effectiveness of structured prototype calibration for modeling annotation variability.
arXiv:2606.17936v2 Announce Type: replace Abstract: In autonomous racing, fast detection of other participants' movements is required to plan safe, collision-free trajectories with non-cooperative opponents. LiDAR detection is inherently slower and harder to deploy on edge devices than vision methods, causing delayed detections that limit object tracking performance during high-dynamic maneuvering. Utilizing monocular 3D detection enables an easy-to-deploy, low-latency detection of other participants on the racetrack. We present SPARK, a single-camera pose-estimation algorithm for autonomous racing using keypoint detection. It achieves long-range detection with high accuracy, exceeding the performance of state-of-the-art monocular camera detection algorithms while maintaining lower latency. By employing well-optimized YOLO models and leveraging the fixed geometry in the autonomous racing domain, the algorithm also exhibits low latency and resource usage. We evaluate the performance of our approach on real-world autonomous racing data and compare it to state-of-the-art LiDAR and camera detection algorithms. The source code is available at: https://github.com/TUMFTM/SPARK-camera-det
arXiv:2606.19852v2 Announce Type: replace Abstract: Information extraction from pathology reports is essential for cancer staging, tumor registry population. Yet key data remains embedded in narrative reports, making manual extraction labor-intensive and error-prone. Traditional supervised Natural Language Processing pipelines address this through fully supervised Named Entity Recognition and Relation Extraction, but require expensive manual annotation and suffer cascading failures when upstream entities are missed. In this study, we developed a zero-shot, agentic workflow, and evaluated five open-source generative Large Language Models (LLMs) to populate 13 College of American Pathologists synoptic fields from lung resection pathology reports. We compared them against a state-of-the-art supervised GatorTron NER-RE baseline using a novel, registry-aligned evaluation framework. The baseline achieved Micro-F1of 0.960, while the best zero-shot model (GPT-OSS-20B) achieved Micro-F1 of 0.893 (recall: 0.949), accurately extracting complex relations like Pathologic Stage without task-specific training. These results suggest that open-source, zero-shot agentic LLMs show great potential as a low-cost solution for extracting lung pathology information.
arXiv:2606.20390v2 Announce Type: replace Abstract: Automated skin cancer classification from dermoscopic images remains challenging due to heterogeneous lesion structure, strong intra-class variability, and subtle visual differences between benign and malignant cases. Existing CNN/ViT pipelines typically rely on global or patch-level features and often combine patient metadata via late fusion, which limits spatially grounded multimodal reasoning. We present a novel region-based graph learning framework that explicitly models lesions as graphs of spatially coherent superpixel regions represented as frozen CNN features. To capture fine-grained lesion arrangements, we encode inter-regional geometry as edge attributes and introduce a dedicated metadata context node connected to all regions, providing structured integration of demographic/clinical variables within the same relational space. Node representations are updated using our edge-aware graph transformer followed by attention-driven propagation, and a final graph-level embedding for benign-malignant classification. Experiments on four public benchmarks demonstrate that explicit region-level relational modeling and graph-native multimodal fusion yield consistent gains over the state-of-the-art. Consequently, we establish a new graph-centric perspective in which CNN features are modeled as relational nodes and improved through contextual integration, yielding more expressive and robust classifications.