arXiv:2607.11839v1 Announce Type: new
Abstract: This paper presents a cascaded Low-Rank Adaptation (LoRA)-based multimodal fusion framework for action and activity recognition in healthcare-oriented training environments. The proposed architecture combines parameter-efficient modality-specific adaptation with sequential fusion, enabling modalities to be integrated in stages without retraining previously learned components. Rather than assuming a fixed fusion structure, the framework first integrates more closely related modalities and then incorporates additional heterogeneous modalities, supporting scalable adaptation across datasets with different modality sets.We evaluate the framework on two healthcare-oriented training environment datasets: NurViD and the Nurse Training dataset. Across these datasets, preliminary results suggest that the proposed cascaded fusion strategy improves over individual modality models and provides competitive performance relative to previously reported dataset-specific baselines. Overall, these findings indicate that cascaded LoRA-based fusion is a promising parameter-efficient approach for integrating heterogeneous modalities in medical training action and activity recognition tasks. github: https://github.com/anonymous0-ai/LoRA-Based-Cascaded-Multimodal-Fusion-.git.
Science Journals
arXiv:2607.09740v1 Announce Type: new
Abstract: Safe motion planning in advanced driver-assistance systems and autonomous vehicles requires an accurate understanding of how the surrounding traffic scene is likely to evolve. However, many existing lane-change prediction methods remain centered on a single target vehicle, while multi-agent forecasting approaches often describe scene evolution only through future positions and provide limited explicit information about the maneuver associated with each vehicle. This study proposes a dynamic scene graph attention framework that predicts the lane-change intention and future trajectory of every relevant vehicle within a local traffic scene. The scene is represented as a time-varying interaction graph in which vehicles are modeled as nodes and their spatial and kinematic relationships are encoded through explicit edge features. Temporal graph-attention message passing captures evolving inter-vehicle dependencies and pre-maneuver cues, while an intention-guided decoder links each predicted maneuver to its corresponding future motion. A scene-level consistency objective further encourages compatible multi-vehicle futures. Experiments on the NGSIM I-80, NGSIM US-101, and highD datasets demonstrate consistent improvements over competing baselines. DSiGAT achieves intention prediction accuracies of 90.12% and 90.97% on NGSIM I-80 and US-101, respectively, and reduces trajectory RMSE by up to 52.94% relative to the strongest baseline. It also produces lower inter-agent collision rates and joint displacement errors, indicating more coherent scene-level predictions. Ablation, sensitivity, robustness, and qualitative analyses further validate the contribution of the proposed components and the effectiveness of the scene-focused formulation.
arXiv:2607.11422v1 Announce Type: new
Abstract: We study the truncated trapezoidal rule, a family of A-stable second order multistep methods parametrized by an integer $J \ge 2$, a compromise between BDF2 and the trapezoidal rule. We obtain a closed-form expression for the coefficients that minimize the principal error constant under the A-stability constraint and we derive an explicit formula for the corresponding principal error constant. The latter decreases to the optimal Dahlquist value $1/12$ as $J$ increases, and is strictly smaller than the BDF2 constant for every $J \ge 2$. We apply the truncated trapezoidal rule within the convolution quadrature framework. Its analyticity in a neighbourhood of the closed unit disk yields milder regularity and perturbation requirements than those of the trapezoidal rule, while the error constant can be made arbitrarily close to the optimal one by increasing $J$. Numerical experiments show that convolution quadrature based on the truncated trapezoidal rule remains stable under symbol perturbations, where the trapezoidal rule fails, while achieving a smaller error constant than BDF2.
arXiv:2607.11850v1 Announce Type: new
Abstract: We present Need for Speed Sort (NFS Sort), a recursive distribution-based sorting algorithm designed for numeric arrays. The algorithm partitions elements into equal-width value intervals, recursively refines dense buckets, and propagates analytical interval bounds between recursive calls, avoiding repeated scans for local minima and maxima. NFS Sort combines a fragment-based, cache-conscious scatter procedure for large subarrays with a lower-overhead auxiliary-array approach for smaller inputs. Small buckets are deferred to a final insertion-sort cleanup, while a comparison-based fallback is activated when recursive partitioning repeatedly fails to reduce the problem size. This mechanism guarantees a worst-case running time of O(n log n) and auxiliary space usage of O(log n). Experimental evaluation on synthetic inputs and real-world datasets from the SOSD benchmark suite compares NFS Sort with Balanced Learned Sort, IPS4o, Boost Spreadsort, PDQSort, and std::sort. The results show that NFS Sort is competitive or better than established state-of-the-art sorting methods across dataset sizes and distributions, outperforming the learned baseline particularly on smaller inputs while retaining strong performance at larger scales. Overall, NFS Sort combines efficient recursive distribution, practical memory management, and robust worst-case guarantees for high-performance numeric sorting.
arXiv:2607.10694v1 Announce Type: new
Abstract: We study the problem of optimal continual fine-tuning for a pre-trained Foundation Model deployed at a resource-limited device. At each time slot, a new batch of training data arrives, and the controller is faced with two options: either use the data to fine-tune the model and incur a compute cost, or do not fine-tune the model and discard the data. After the decision, the performance of the current model is measured in terms of an application-specific performance metric such as classification accuracy. Our objective is to learn an optimal policy that determines \emph{when to fine-tune the model} on a single task (e.g., sentiment analysis), under a finite compute budget. We formulate this online decision-making problem as a constrained Markov Decision Process, where the system state captures three essential aspects: (\textit{i}) model's performance, (\textit{ii}) computational budget, and (\textit{iii}) data distribution relevance to historic data encountered up to that point. The transition to the next state is stochastic and therefore, we propose a reinforcement learning-based method to solve this problem, namely the \emph{actor-critic} algorithm. We also consider the special case where the performance of fine-tuning for a given model can be predicted or estimated prior to decision; in this case the problem becomes a Dynamic Programming one. Experiments with a large pre-trained model on a widely-used text classification dataset demonstrate that our method consistently outperforms fine-tuning approaches with the same compute budget by more than $4\%$ in terms of accuracy and achieves $97\%$ of full-parameter fine-tuning accuracy while requiring only $25\%$ of the fine-tuning steps.
arXiv:2607.10565v1 Announce Type: new
Abstract: End-to-end motion planning has emerged as a promising paradigm in autonomous driving, directly mapping raw sensor data to control commands via deep neural networks. Despite its advantages, its large model size hinders deployment in resource-constrained platforms. In this paper, we present BucketKD, a bucket-based knowledge distillation framework that yields compact and safety-aware end-to-end planners. Compared to the state-of-the-art approach, which relies on simplified planning state representations, BucketKD discretizes critical environmental variables into adaptive buckets that capture richer scene semantics while preserving efficiency. In addition, we design a safety-aware waypoint attention mechanism that evaluates each waypoint's risk level by accounting for both obstacle proximity and relative motion through a time-to-collision (TTC) formulation widely used in transportation research. This enables the student model to better retain safety-critical behaviors during distillation. Extensive experiments in CARLA using the Bench2Drive dataset show that BucketKD significantly outperforms the state-of-the-art in both planning accuracy and safety while maintaining strong compression ratios.
arXiv:2607.11859v1 Announce Type: new
Abstract: Can large language models perform deep technical comprehension of computer architecture papers -- not summarization, but structured critique that names the core mechanism, surfaces buried assumptions, and connects a contribution beyond its own scope? We study Gauntlet, an open-source pipeline that analyzes a paper through five independent expert-persona reviewers and an adversarial synthesis stage. On 20 ISCA 2025 and HPCA 2026 papers, ten researchers each wrote their own analyses and then judged, for papers other than their own, the human analysis against Gauntlet's. Across the 20 comparisons evaluators preferred Gauntlet in 15 (human in 4, one tie); its advantage is significant on per-analyst totals (paired Wilcoxon, p < 0.01) and largest on Critical Rigor, vanishing only on Calibration. Where humans win, it is on trust and usefulness rather than depth: a confident wrong claim, a mechanism described but not taught, or unprioritized breadth. A 98-paper automated ablation shows the gain comes from the multi-agent structure -- the pipeline beats the same model run as a single rich-persona agent on 96% of papers -- and specifically from its synthesis pass. We release all analyses, scores, and the rubric as a community resource.
arXiv:2607.09772v1 Announce Type: new
Abstract: Autonomous driving systems require reliable safety validation before real-world deployment. However, large-scale road testing is costly, difffcult to reproduce, and inefffcient for exposing rare safety-critical scenarios. Conventional simulation improves repeatability, but an offfine simulator alone cannot continuously connect physical trafffc states, virtual reconstruction, algorithm evaluation, and scenario evolution. This paper proposes a risk-ffeld enhanced closed-loop digital twin framework for autonomous driving safety validation. The framework integrates physical data acquisition, data synchronization, virtual twin reconstruction, risk-aware scenario generation, autonomous driving algorithm evaluation, and safety analysis. A driving risk ffeld is introduced as a uniffed intermediate representation to describe obstacle, lane-departure, road-boundary, time-to-collision, and comfort-related risks around the ego vehicle. The risk ffeld ranks high-risk scenarios in the digital twin scenario library and provides dense safety guidance for reinforcement learning-based driving policies. A simulation-style evaluation protocol is designed to compare conventional reinforcement learning baselines, risk-penalty baselines, and the proposed risk-ffeld guided method. The study indicates that embedding explicit risk structure into digital twins can make autonomous driving validation more targeted, interpretable, and reusable, while its practical effectiveness remains bounded by model ffdelity, risk calibration, and sim-to-real transfer.
arXiv:2607.09782v1 Announce Type: new
Abstract: In evolutionary game dynamics, there exists a hypothesis, which states that, the dynamic structure of the game's steady -- state system is characterized by the linear superposition of eigenmanifolds, which depends specifically on the eigenvector structure at the Nash equilibrium and is ultimately governed by the game dynamics equations. This hypotheses has been supported widely in discrete strategy game. In continuous -- strategy game, using experimental data from human -- subject games, this paper finds that the hypothesis is supported in significant, too.
arXiv:2607.09826v1 Announce Type: new
Abstract: Dysgraphia is a specific learning disability that is prevalent among school-age children. It affects handwriting coherence, quality, fluency, and legibility, often hindering academic achievement and early learning development. This motor coordination disorder is typically diagnosed through subjective assessments based on clinician observation, which can be timeconsuming and prone to variability. In this paper, we introduce a deep learning-based framework for objective dysgraphia detection using online handwriting data captured via digitizing tablets. The proposed framework relies on two complementary branches: the first pipeline extracts both handcrafted and embedding-based kinematic features directly from raw temporal signals, while the second leverages image-based representations of the temporal signals generated using continuous wavelet transforms (CWT) and Gramian Angular Fields (GAF). The resulting features are then fused to leverage the complementary strengths of both representations. The four representations were evaluated separately and jointly using the publicly available DiaGraMo dataset, showing that the fusion of GAF, MOMENT, and hand-crafted kinematic features outperforms each individual representation, as well as other fusion schemes. These findings highlight the potential of the complementarity of image and signal based representations for more objective dysgraphia detection.
arXiv:2607.10749v1 Announce Type: new
Abstract: Reflections of water pose a significant challenge for computer vision systems, as standard deep learning models frequently confuse objects with their mirror images, producing spurious false positives and negatives in tasks such as object detection and semantic segmentation. As a result, detecting reflection axes in natural-water scenes is pivotal for reliable object detection and scene understanding. To mitigate this issue, we leverage the intrinsic imperfect reflective symmetry of water and introduce a Symmetry-Aware Water Reflection Detection Network, namely, SAWRD-Net, that couples dihedral group-equivariant convolutions with a matrix-decomposition decoder in an end-to-end framework. First, dihedral group convolutional layers extract geometry-consistent feature maps that explicitly encode both rotational and mirror symmetries. A Multi-scale Reflection Equivariant block then aggregates features across scales and employs a symmetric-attention mechanism to highlight reflection-relevant regions. The proposed matrix-decomposition decoder factorizes high-dimensional features into compact low-rank parameter and confidence spaces, after which the network directly regresses keypoints on the reflection axis. Then a robust principal component analysis fits the final axis. Evaluated on the largest available water reflection scene data set, SAWRD-Net achieves a true-positive rate of 0.890 against human annotations, outperforming all existing water reflection detectors.
arXiv:2607.09945v1 Announce Type: new
Abstract: Purpose: Motion compromises the utility of high-resolution 3D MRI, an established tool in quantitative neuroimaging research. Deep learning-based methods have shown promise for mitigating motion-induced artifacts, but their development typically requires simulated motion-corrupted data. Several open-source tools exist for this task, each implementing different algorithms. However, no scheme currently exists for evaluating the accuracy of these simulations, making it difficult for users to choose the most suitable tool. Developing such a scheme is the aim of this study. Methods: The essential ingredient of the desired scheme is a ground-truth reference simulation that does not suffer from sampling-induced error. To meet this requirement, the proposed scheme, APHABAMAS, leverages a digital phantom whose representations in both the image and Fourier domains can be expressed analytically under arbitrary rigid-body transformations. Results: APHABAMAS is used to quantify the sampling-induced errors of three existing simulation algorithms, establishing their first definitive accuracy-based ranking. Conclusions: APHABAMAS provides a rigorous tool for assessing the accuracy of high-resolution 3D MRI motion-artifact simulations. It allows the accuracy-based ranking of existing simulation algorithms to be established, thereby enabling informed selection of the most suitable algorithm for synthesizing motion-corrupted data.
arXiv:2607.09985v1 Announce Type: new
Abstract: Object pose estimation is a fundamental problem in 3D vision. Although recent state-of-the-art approaches achieve strong performance, they often overfit to existing benchmarks and exhibit limited generalization to novel categories and unseen scenes. We propose UniPose9D, a category-agnostic foundation model for 9D object pose estimation: given an instance mask/ROI and either an RGB-D observation or an RGB image with predicted depth, the model estimates rotation, translation, and metric size without category labels, CAD models, mean-shape priors, or reference views. Specifically, UniPose9D samples point pairs from the observed object geometry and uses DINOv2 and PointNet features to predict NOCS coordinates for each pair. To improve accuracy, we introduce a point-pair-based RANSAC N-hop Kabsch--Umeyama algorithm with an adaptive threshold. We further employ flow matching to address symmetric ambiguities and construct a large-scale training set by curating and aligning pose annotations from existing public datasets. Experiments across six datasets show that a single unified model can match or surpass specialist methods while generalizing to unseen objects and in-the-wild scenarios. Our code and model are available on https://github.com/qq456cvb/UniPose9D.
arXiv:2607.10085v1 Announce Type: new
Abstract: Modelling rarefied gas flow using the Boltzmann equation is vital in many areas. Due to the high dimensionality and coexistence of multiple characteristic scales, conventional solution strategies to this equation incur prohibitively high computational costs and are inadequate for rapid response in engineering design simulations. Based on proper generalised decomposition (PGD), we propose an \textit{a priori}, asymptotic-preserving reduced-order method to solve the high-dimensional, parametrised Shakhov kinetic model equation. The method reduces the original problem to a few low-dimensional problems by formulating separated representations for the low-rank solution, thereby mitigating the curse of dimensionality. To capture the hydrodynamic asymptotics, we incorporated solutions of some synthetic equations into the PGD algorithm. This treatment allows the PGD solver to automatically reduce to a macroscopic solver for the Navier-Stokes equations, whose solution naturally exhibits low-rank structure. By treating the rarefaction parameter as an additional coordinate, a parametrised solution can be computed once and for all over the entire range of rarefaction, enabling fast multiple queries to any points in the parameter space. Numerical examples are presented to demonstrate the capability of the method to simulate rarefied gas flow with certain accuracy and a significant reduction in computational costs.
arXiv:2607.10775v1 Announce Type: new
Abstract: We analyze chemical short-range order in equiatomic fcc NiCoV using molecular-dynamics snapshots generated with a machine-learned interatomic potential. Radial distribution functions identify stable coordination shells, while shell-resolved Warren-Cowley parameters and bond probabilities reveal continued chemical ordering after the radial structure has largely converged. The dominant signal is V-V avoidance in the first shell and V-V enrichment in the second shell, consistent with an L1$_2$-like local ordering tendency, while the third-shell response remains weak. Lagged Jensen-Shannon diagnostics show that bond statistics relax more slowly than the RDF. Principal component analysis of per-replica-centered bond probabilities resolves three collective modes: a V-sublattice ordering amplitude, a Ni-Co redistribution mode, and a Co-V exchange-like mode. These results show that scalar RDF convergence can miss slow chemical relaxation, and that shell-resolved bond statistics provide a compact route for tracking SRO development in multicomponent alloys.
arXiv:2607.09779v1 Announce Type: new
Abstract: The deep nonnegative matrix factorization (DNMF) technique is proposed to address the low interpretability of deep learning-based methods in extracting multilayer features from synthetic aperture radar (SAR) target samples. However, existing DNMF methods employ a layer-by-layer decomposition strategy, which is prone to causing error accumulation and local optimum, thereby hindering a consistent improvement in recognition accuracy as the number of layer increases. In this paper, a robust multilayer feature extraction method, termed generalized deep non-negative matrix factorization (G-DNMF), is proposed to address the above challenges in SAR automatic target recognition (ATR). The G-DNMF aims global optimality and derives the update rules for each parameter using lagrangian multiplier method. The new update formula indicates that both the DNMF method based on the encoding matrix and the mixing matrix are special cases of the proposed method, theoretically demonstrating the universality of proposed method. In general, the proposed method discards the layer-by-layer decomposition strategy, thereby effectively mitigating the risk of local optima and eliminating error accumulation, leading to a significant improvement in DNMF's multi-layer feature extraction capability. The experimental results, by presenting the feature images extracted from each layer by G-DNMF and the reconstructed original images, verified the proposed method's pure additive understanding of multi-layer features and demonstrated its interpretability. The experimental results based on MSTAR and OpenSARship datasets show that G-DNMF outperforms existing DNMF algorithms and their derivatives in terms of stability and recognition performance.
arXiv:2607.10689v1 Announce Type: new
Abstract: In the (fully) dynamic edge connectivity problem, the goal is to maintain the edge connectivity $\lambda_G$ of an $n$-vertex graph $G$ that undergoes edge insertions and deletions. Our main result is a randomized algorithm for maintaining edge connectivity in dynamic simple graphs using worst-case update and query time $\tilde{O}(n^{12/13})$, for all values of $\lambda_G$. This is the first algorithm that has $o(n)$ update and query time, as all existing algorithms achieve this only when $\lambda_G$ is below $n^{1/11}$ or above $n^{1/2}$ (up to polylogarithmic factors). We then use the tools developed for this purpose to design two additional algorithms. The first one is a deterministic algorithm for the exact same task, that uses $n^{1+o(1)}$ worst-case update and query time or $\tilde{O}(n)$ amortized update and query time; this gives a polynomial improvement over existing deterministic algorithms. The second one is a deterministic algorithm for the same task but in dynamic unweighted multigraphs, that uses $\tilde{O}(n^{3/2})$ worst-case update and query time.
arXiv:2607.11874v1 Announce Type: new
Abstract: Recent work in humanoid whole-body control has found success with a simple recipe: retarget human motion to robot kinematic references, then train policies via reinforcement learning (RL) to track them. But how does this recipe transfer to dexterous manipulation? The answer is not obvious, as manipulation involves complex, contact-rich dynamics and requires delicate regulation of contact modes and forces. We present REGRIND, a minimalist retargeting-guided RL pipeline that learns dexterous manipulation policies from a single human demonstration. REGRIND retargets human hand-object motion to a robot reference that preserves hand-object spatial and contact relationships, trains a residual RL policy in simulation to track object-centric keypoints along that reference, and transfers the resulting policy zero-shot to hardware with careful system identification. The resulting policies produce fluid, human-like behavior on two different multi-fingered hands across contact-rich tool-use tasks, including operating a pair of scissors and turning a screwdriver. Through systematic hardware experiments, we identify and analyze the key factors that govern sim-to-real transfer in dexterous manipulation, offering practical guidance for retargeting-based learning in contact-rich settings. Videos and code are available at https://yunhaifeng.com/REGRIND.
arXiv:2607.10238v1 Announce Type: new
Abstract: Video emotion analysis is typically framed as a static classification problem, treating each clip as an independent labeled unit. However, such a formulation overlooks a key psychological fact: emotions change as a result of cumulative reactions to consecutive causal events. To bridge this gap, we introduce Dynamic Affective Reasoning, the first large-scale benchmark for viewer-centric affect transitions and causal reasoning over consecutive video events. DAR contains 15,087 videos and 36,908 event-aligned affective segments annotated with 27 emotion categories. Unlike existing video-based emotion datasets, DAR presents a new viewer-centric perspective on fine-grained emotional expressions and transitions, and provides dense, temporally grounded, and causally explicit reasoning chains. Based on DAR, we formally define three challenging tasks: affective segmentation, fine-grained emotion classification, and affective reasoning. Complementing this benchmark, we propose DAR-R1, a two-stage framework that combines supervised fine-tuning with Group Relative Policy Optimization. Experiments across 10+ MLLMs show that DAR-R1 sets a new state-of-the-art for dynamic affective reasoning, in terms of both emotional localization and affective reasoning. Project page: https://github.com/Zhang-Zhiyan/DAR.
arXiv:2607.10240v1 Announce Type: new
Abstract: Short-answer VQA benchmarks conflate two distinct quantities: whether a model's answer is semantically correct, and whether that answer matches the surface form expected by the automatic evaluator. We study this conflation across six vision--language models and six benchmarks, using a human-validated semantic judge (97.6% precision) to audit over 37k official errors. A second text-only judge reproduces the same benchmark-level false-negative pattern, showing that the effect is not an artifact of a single audit model. On text-rich benchmarks, up to half of these errors are semantically acceptable answers penalized purely for surface-form mismatch. This instability is structured by answer type: extractive and multi-span answers are far more evaluator-sensitive than scalar answers. Benign prompt and context rewrites further destabilize official outcomes, flipping item-level correctness at substantial rates without changing the underlying task. A deterministic CPU-only contract repair confirms that the undercount is partially recoverable. These findings imply that official short-answer VQA scores should be accompanied by semantic audits and answer-type diagnostics to remain interpretable.
arXiv:2607.09721v1 Announce Type: cross
Abstract: To help evaluate the mathematical skills of current AI systems, we present a set of formulas for fundamental mathematical constants. These problems are attractive for AI evaluation because they are concrete and can be checked numerically to arbitrary precision, yet proving them may require non-obvious mathematics. Mathematical constants such as $\pi$, $e$, Catalan's constant, and special values of the Riemann zeta function have fascinated mathematicians for centuries. The search for formulas evaluating mathematical constants has produced some of the most beautiful mathematics in the field, especially in cases that yield irrationality proofs or fast convergence rates. Ramanujan's legacy is emblematic of this tradition. The list we provide contains two types of problems: formulas whose proofs are known to the authors but will remain encrypted for a short initial period; and formulas that are not yet proven. We are curious to see the achievements of AI in both cases.
arXiv:2607.10112v1 Announce Type: new
Abstract: Safety alignment in large language models remains brittle across languages: prompts reliably refused in English can elicit harmful compliance in non-English and low-resource settings. We introduce \textsc{Minionese}, a multilingual jailbreak benchmark spanning 18 languages, 4 resource tiers, and 4 perturbation types (standard translation, code-switching, transliteration, and translationese), paired with a geometric mechanistic analysis of refusal failure across language tiers. We show that each attack type produces a distinct vulnerability profile: transliteration vulnerability is mediated by script identity, code-switching maintains effectiveness through the lowest-resource tier, and a sharp safety regime transition between Tiers 2 and 3 is consistent across all models. Mechanistically, low-resource jailbreaks succeed by routing harmful content through a geometrically misaligned subspace that projects insufficiently onto the refusal directions, leaving the refusal mechanism intact but untriggered. These findings show that English-only safety evaluations are insufficient; they require accounting for script family, perturbation type, and per-language alignment coverage. The benchmark and analysis code is at https://github.com/Brentkong/Minionese-Comprehensive-Benchmark-and-Mechanistic-Study-of-Multilingual-LLM-Safety.git.
arXiv:2607.10263v1 Announce Type: new
Abstract: Bilevel optimization underpins many machine learning applications, including hyperparameter optimization, meta-learning, neural architecture search, and reinforcement learning. While hypergradient-based methods have advanced significantly, a gap persists between theoretical guarantees and practical single-loop implementations required for efficiency. We bridge this gap by establishing sharper convergence results for single-loop approximate implicit differentiation (AID) and iterative differentiation (ITD) methods, leveraging our proposed analytical framework, decoupled norm analysis (DNA). For AID, we improve the convergence rate from $\mathcal{O}(\kappa^6/K)$ to $\mathcal{O}(\kappa^5/K)$, where $\kappa$ is the condition number of the inner-level problem. For ITD, we prove that the asymptotic error is $\mathcal{O}(\kappa^2)$, exactly matching the known lower bound and improving upon the previous $\mathcal{O}(\kappa^3)$ guarantee. Numerical experiments on synthetic and real tasks corroborate our theoretical findings.
arXiv:2607.10748v1 Announce Type: new
Abstract: Computed Tomography (CT) diagnosis often relies on dynamic selection of imaging phases, such as non-contrast, arterial, or venous phases, based on preliminary findings, clinical suspicion, and diagnostic guidelines. This phase-wise decision process is critical for reducing unnecessary radiation exposure while supporting timely staging and treatment planning. However, phase-selection protocols can vary across hospitals, regions, and guidelines, while most existing CT-based AI methods assume that all phases are available and focus on static tasks under a fixed imaging phase, failing to model whether additional phases are required. This limitation stems from heterogeneous multi-phase representations, the need for knowledge-guided phase control beyond visual cues, and the lack of supervision for phase-sufficiency decisions in existing datasets. To address these challenges, we propose Policy-Driven CT-Agent (PD-CTAgent) for clinically consistent CT phase selection and diagnostic reasoning. PD-CTAgent introduces a Clinical Structure Abstraction Module (CSAM) to harmonize heterogeneous CT phases into a unified, phase-aware evidence representation. Based on this representation, a Knowledge-Guided Diagnostic Control Model (KDCM) evaluates phase sufficiency and iteratively requests additional phases when necessary. The policy-driven agent design further allows PD-CTAgent to flexibly follow different institutional, regional, or guideline-specific diagnostic protocols. Together, PD-CTAgent bridges static CT analysis and real-world clinical workflows. Experiments on two public datasets, LIDC and MCT-LTDiag, and one private dataset demonstrate its effectiveness and clinical consistency. Code will be made public upon acceptance.
arXiv:2607.10911v1 Announce Type: new
Abstract: In the age of large language models, Natural Language to SQL (NL2SQL) translation remains an open problem with many useful applications. We explore interactions between several NL2SQL pipeline extensions to inspire development of more lightweight models. Specifically, we integrate the NatSQL intermediate representation, include a preprocessing step and a fine-tuning step based on synthetic data, and develop a novel reranker model to improve SQL selection in the final beam. We perform an ablation study supplemented by a Shapley analysis of these different components integrated with two backbone architectures, SmBoP and RASAT. We find that simply combining all of them does not lead to best results, but that their impact depends on their interactions with the baseline system, as well as each other.