# Research Papers — 2026-04 · bundle page 24 of 233

> Full-text markdown of papers #1151–1200 of 11612 published in 2026-04, newest first. AI-readable mirrors of the full 11612-paper 2026-04 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-04/page/24
[← Previous bundle](https://aipod.dev/papers/archive/2026-04/page/23.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-04/page/25.md)

### AHASD: Asynchronous Heterogeneous Architecture for LLM Adaptive Drafting Speculative Decoding on Mobile Devices

Authors: Ma Zirui, Fan Zhihua, Li Wenxing, Wu Haibin, Zhang Fulin, Ye Xiaochun, Li Wenming
Published: 2026-04-28
Categories: cs.AR, cs.AI
arXiv: https://arxiv.org/abs/2604.25326v3

Speculative decoding enhances the inference efficiency of large language models (LLMs) by generating drafts using a small draft language model (DLM) and verifying them in batches with a large target language model (TLM). However, adaptive drafting inference on a mobile single-NPU-PIM system faces idle overhead in traditional operator-level synchronous execution and wasted computation in asynchronous execution due to fluctuations in draft length. This paper introduces AHASD, a task-level asynchronous mobile NPU-PIM heterogeneous architecture for speculative decoding. Notably, AHASD achieves parallel drafting on the PIM and verification on a single NPU through task-level DLM-TLM decoupling and specifically, it incorporates Entropy-History-Aware Drafting Control and Time-Aware Pre-Verification Control to dynamically manage adaptive drafting algorithm execution and pre-verification timing, suppressing invalid drafting based on low-confidence drafts. Additionally, AHASD integrates Attention Algorithm Units and Gated Task Scheduling Units within LPDDR5-PIM to enable attention link localization and sub-microsecond task switching on the PIM side. Experimental results for different LLMs and adaptive drafting algorithms show that AHASD achieves up to 4.2$\times$ in throughput and 5.6$\times$ in energy efficiency improvements over a GPU-only baseline, and 1.5$\times$ in throughput and 1.24$\times$ in energy efficiency gains over the state-of-the-art GPU+PIM baseline, with hardware overhead below 3% of the DRAM area.

---

### R$^3$-SQL: Ranking Reward and Resampling for Text-to-SQL

Authors: Hojae Han, Yeonseok Jeong, Seung-won Hwang, Zhewei Yao, Yuxiong He
Published: 2026-04-28
Categories: cs.SE, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2604.25325v1

Modern Text-to-SQL systems generate multiple candidate SQL queries and rank them to judge a final prediction. However, existing methods face two limitations. First, they often score functionally equivalent SQL queries inconsistently despite identical execution results. Second, ranking cannot recover when the correct SQL is absent from the candidate pool. We propose R$^3$-SQL, a Text-to-SQL framework that addresses both issues through unified reward for ranking and resampling. R$^3$-SQL first groups candidates by execution result and ranks groups for consistency. To score each group, it combines a pairwise preference across groups with a pointwise utility from the best group rank and size, capturing relative preference, consistency, and candidate quality. To improve candidate recall, R$^3$-SQL introduces agentic resampling, which judges the generated candidate pool and selectively resamples when the correct SQL is likely absent. R$^3$-SQL achieves 75.03 execution accuracy on BIRD-dev, a new state of the art among methods using models with disclosed sizes, with consistent gains across five benchmarks.

---

### Quantifying mandibular positioning error and simulated temporomandibular joint-space changes in patient-specific occlusal splints

Authors: Agnieszka Anna Tomaka, Krzysztof Domino, Michał Tarnawski, Dariusz Pojda
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25322v2

Patient-specific occlusal positioning splints can be regarded as physical realisations of planned mandibular transformations. However, the achieved mandibular pose may differ from the planned one because of acquisition, registration, fabrication, and positioning errors. This study presents a transformation-based biomedical engineering framework for quantifying mandibular positioning accuracy and propagating the resulting error to a simulated temporomandibular joint configuration. Multimodal 3D data, including CBCT, facial motion acquisition, and dental scans, were integrated in a common coordinate system. Positioning splints corresponding to selected mandibular poses were designed and fabricated, and their realised positions were evaluated using repeated scans of plaster models. Discrepancies between planned and achieved positions were represented as rigid-body error transformations and analysed in SE(3), together with surface-distance metrics. The estimated transformations were propagated to CBCT-derived TMJ structures to quantify changes in condyle-fossa distance maps. The results demonstrate a systematic translational component and anisotropic variability of mandibular positioning error, with measurable propagation to simulated TMJ-space changes. The proposed framework provides an objective method for documenting planned and achieved mandibular configurations and for analysing positioning uncertainty in patient-specific splint workflows.

---

### Edge-Cloud Collaborative Reconstruction via Structure-Aware Latent Diffusion for Downstream Remote Sensing Perception

Authors: Yun Li, Xianju Li
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25319v1

The exponential surge in high-resolution remote sensing data faces a severe bottleneck in satellite-to-ground transmission. Limited downlink bandwidth forces the use of extreme high-ratio compression, which irreversibly destroys high-frequency structural details essential for downstream machine perception tasks like object detection. While current super-resolution techniques attempt to recover these details, regression-based methods often yield over-smoothed textures, and generative diffusion models frequently introduce structural hallucinations that mislead detection systems. To address this trade-off, we propose the Structure-Aware Latent Diffusion (SALD) framework, an asymmetric edge-cloud collaborative SR system. At the resource-constrained edge, the system decouples imagery into a highly compressed low-frequency payload and a lightweight soft structural prior. Transmitting this decoupled representation minimizes bandwidth consumption. On the powerful cloud side, we introduce a Structure-Gated Large Kernel (SGLK) module and a Semantic-Guidance Engine (SGE) within the diffusion backbone. These modules leverage the transmitted structural priors to gate large-kernel convolutions, effectively capturing long-range dependencies inherent in aerial scenes while actively suppressing generative hallucinations. Extensive experiments on both the MSCM and UCMerced datasets demonstrate that, even under extreme bandwidth constraints, SALD achieves superior perceptual quality (LPIPS) and significantly enhances downstream performance in both scene classification and small-target detection.

---

### Cutscene Agent: An LLM Agent Framework for Automated 3D Cutscene Generation

Authors: Lanshan He, Haozhou Pang, Qi Gan, Xin Shen, Ziwei Zhang, Yibo Liu, Gang Fang, Bo Liu, Kai Sheng, Shengfeng Zeng, Chaofan Li, Zhen Hui et al.
Published: 2026-04-28
Categories: cs.GR, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2604.25318v1

Cutscenes are carefully choreographed cinematic sequences embedded in video games and interactive media, serving as the primary vehicle for narrative delivery, character development, and emotional engagement. Producing cutscenes is inherently complex: it demands seamless coordination across screenwriting, cinematography, character animation, voice acting, and technical direction, often requiring days to weeks of collaborative effort from multidisciplinary teams to produce minutes of polished content. In this work, we present Cutscene Agent, an LLM agent framework for automated end-to-end cutscene generation. The framework makes three contributions: (1)~a Cutscene Toolkit built on the Model Context Protocol (MCP) that establishes \emph{bidirectional} integration between LLM agents and the game engine -- agents not only invoke engine operations but continuously observe real-time scene state, enabling closed-loop generation of editable engine-native cinematic assets; (2)~a multi-agent system where a director agent orchestrates specialist subagents for animation, cinematography, and sound design, augmented by a visual reasoning feedback loop for perception-driven refinement; and (3)~CutsceneBench, a hierarchical evaluation benchmark for cutscene generation. Unlike typical tool-use benchmarks that evaluate short, isolated function calls, cutscene generation requires long-horizon, multi-step orchestration of dozens of interdependent tool invocations with strict ordering constraints -- a capability dimension that existing benchmarks do not cover. We evaluate a range of LLMs on CutsceneBench and analyze their performance across this challenging task.

---

### Towards Robust Deep Learning-based Rumex Obtusifolius Detection from Drone Images

Authors: Fabian Dionys Schrag, Mehmet Ozgur Turkoglu, Konrad Schindler, Ralph Lukas Stoop
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25316v1

Domain adaptation (DA) addresses the challenge of transferring a machine learning model trained on a source domain to a target domain with a different data distribution. In this work, we study DA for the task of Rumex obtusifolius (Rumex) image classification. We train models on a published, ground vehicle-based dataset (source) and evaluate their performance on a custom target dataset acquired by unmanned aerial vehicles (UAVs). We find that Convolutional Neural Network (CNN) models, specifically ResNets, generalize poorly to the target domain, even after fine-tuning on the source data. Applying moment-matching and maximum classifier discrepancy, two established DA techniques, substantially improves target-domain performance. However, Vision Transformer (ViT) models pretrained with self-supervised objectives (DINOv2, DINOv3) handle domain shifts intrinsically well, surpassing even moment-matching-trained ResNets, likely due to the rich, general-purpose representations acquired during large-scale pretraining. Using ViTs fine-tuned on the source dataset, we demonstrate high classification performances in the range of F1=0.8 on our target dataset. To support further research on DA for weed detection in grassland systems, we publicly release our UAV-based target dataset AGSMultiRumex, comprising data from 15 flights over Swiss meadows.

---

### SaliencyDecor: Enhancing Neural Network Interpretability through Feature Decorrelation

Authors: Ali Karkehabadi, Jamshid Hassanpour, Houman Homayoun, Avesta Sasan
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25315v1

Gradient-based saliency methods are widely used to interpret deep neural networks, yet they often produce noisy and unstable explanations that poorly align with semantically meaningful input features. We argue that a fundamental cause of this behavior lies in the geometry of learned representations: correlated feature dimensions diffuse attribution gradients across redundant directions, resulting in blurred and unreliable saliency maps. To address this issue, we identify feature correlation as a structural limitation of gradient-based interpretability and propose SaliencyDecor, a training framework that enforces feature decorrelation to improve attribution fidelity without modifying saliency methods or model architectures by reshaping the feature space toward orthogonality, our approach promotes more concentrated gradient flow and improves the fidelity of saliency-based explanations. SaliencyDecor jointly optimizes classification, prediction consistency under feature masking, and a decorrelation regularizer, requiring no architectural changes or inference-time overhead. Extensive experiments across multiple benchmarks and architectures demonstrate that our method produces substantially sharper and more object-focused saliency maps while simultaneously improving predictive performance, achieving accuracy gains across the datasets. These results establish our method as a principled mechanism for enhancing both interpretability and accuracy, challenging the conventional trade-off between explanation quality and model performance.

---

### Golden RPG: Confidence-Adaptive Region-Aware Noise for Compositional Text-to-Image Generation

Authors: Hao Li
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25314v1

Compositional text-to-image (T2I) generation requires a model to honour multiple sub-prompts that describe distinct image regions. Recent work shows that the \emph{starting noise} of a diffusion model carries significant semantic information: ``golden'' noise predicted from text can substantially raise prompt fidelity. We observe that this noise prediction is, however, fundamentally global: the same network is asked to summarise a long, multi-region prompt with a single text embedding, which becomes the bottleneck whenever the prompt describes scenes with spatially-separated entities. We introduce \textbf{Golden RPG}, a region-aware noise predictor that extends a frozen NPNet with two trainable additions: (i) a per-region \textbf{FiLM adapter} that reshapes the predicted noise according to each sub-prompt; and (ii) a \textbf{Region Cross-Attention} layer injected between two stages of the Swin backbone, allowing different spatial locations to attend to different sub-prompt tokens. To prevent the regional conditioning from degrading samples whose prompts are already easy, we further propose a \textbf{Confidence-Adaptive Blending} head that dynamically predicts, per sample, how strongly the regional signal should override the global signal. We evaluate on the original RPG benchmark (20 prompts, 100 samples) and on four multi-region categories of T2I-CompBench (1{,}200 images, six competing methods). Golden RPG achieves the highest Cross-Region-Coherence score on every category, while matching the strongest baselines on absolute CLIP-Score and CLIP-IQA. A paired user study further shows a $\boldsymbol{\sim}$67\% preference over the strongest baseline. The adapter contains $\sim$2M trainable parameters and adds only $0.6$\,s of inference overhead on top of SDXL.

---

### Faithfulness-QA: A Counterfactual Entity Substitution Dataset for Training Context-Faithful RAG Models

Authors: Li Ju, Junzhe Wang, Qi Zhang
Published: 2026-04-28
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2604.25313v2

Retrieval-Augmented Generation (RAG) models frequently produce answers grounded in parametric memory rather than the retrieved context, undermining the core promise of retrieval augmentation. A fundamental obstacle to fixing this unfaithfulness is the lack of training data that explicitly requires models to prefer context over internal knowledge. We introduce Faithfulness-QA, a large-scale dataset of 99,094 samples constructed through counterfactual entity substitution. Starting from two established extractive QA benchmarks--SQuAD and TriviaQA--we automatically identify answer-bearing named entities in each context, replace them with type-consistent alternatives drawn from a curated bank of 76,953 entities, and thereby manufacture controlled knowledge conflicts between context and parametric memory. Rigorous quality filtering ensures 100% pass rates across four automated checks on random 200-sample audits. We release the full dataset, the construction pipeline, and a typed entity bank covering eight named entity categories. Faithfulness-QA is designed as a training resource for attention-based faithfulness objectives and as an evaluation benchmark for measuring context-grounding behavior in RAG systems. Data and code are available at https://github.com/qzhangFDU/faithfulness-qa-dataset.

---

### Rapid tracking through strongly scattering media with physics-informed neuromorphic speckle analysis

Authors: Yuqing Cao, Shuo Zhu, Rongzhou Chen, Jingyan Chen, Ni Chen, Edmund Y. Lam
Published: 2026-04-28
Categories: cs.CV, eess.IV
arXiv: https://arxiv.org/abs/2604.25310v1

This work addresses the critical problem of tracking fast-moving objects through strongly scattering media in a low-light environment. Different from existing approaches that use frame-based cameras with fixed exposure times, which trade off signal-to-noise ratio for temporal resolution, we introduce computational neuromorphic tracking (CNT), a physics-informed framework that combines asynchronous event sensing with task-driven speckle analysis for robust motion estimation. We formulate the neuromorphic speckle aggregation as a spatiotemporal speckle representation, jointly optimizing the temporal and spatial parameters to maximize tracking stability under extreme conditions. Extensive experiments demonstrate that our method enables robust motion tracking of 10x faster motion and under 10x dimmer illumination compared to conventional systems. These improvements significantly broaden the operational regime for tracking through scattering media, providing an efficient and scalable solution for demanding scenarios involving rapid motion and low-light conditions.

---

### QFlash: Bridging Quantization and Memory Efficiency in Vision Transformer Attention

Authors: Sehyeon Oh, Yongin Kwon, Jemin Lee
Published: 2026-04-28
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2604.25306v1

FlashAttention improves efficiency through tiling, but its online softmax still relies on floating-point arithmetic for numerical stability, making full quantization difficult. We identify three main obstacles to integer-only FlashAttention: (1) scale explosion during tile-wise accumulation, (2) inefficient shift-based exponential operations on GPUs, and (3) quantization granularity constraints requiring uniform scales for integer comparison. To address these challenges, we propose \textit{QFlash}, an end-to-end integer FlashAttention design that performs softmax entirely in the integer domain and runs as a single Triton kernel. On seven attention workloads from ViT, DeiT, and Swin models, QFlash achieves up to 6.73$\times$ speedup over I-ViT and up to 8.69$\times$ speedup on Swin, while reducing energy consumption by 18.8\% compared to FP16 FlashAttention, without sacrificing Top-1 accuracy on ViT/DeiT and remaining competitive on Swin under per-tensor quantization. Our code is publicly available at https://github.com/EfficientCompLab/qflash.

---

### RCProb: Probabilistic rule extraction from classification tree ensembles

Authors: Josue Obregon
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.25304v3

Tree ensembles provide strong classification performance but usually behave as black-box models. Post-hoc interpretability techniques such as RuleCOSI+ extract a small ruleset that approximates the ensemble, but this simplification can leave the probabilities attached to the extracted rules unreliable. In particular, RuleCOSI+ assigns empirical class probabilities to the extracted rules and repeatedly uses those rule statistics during its greedy combination and simplification procedure. We present RCProb, a probabilistic extension that uses smoothed atomic class-conditional evidence for the expensive search stages and a support-adaptive mixture with an ensemble-informed m-estimate for the final rule probabilities. The method is evaluated on 18 binary and 5 multiclass datasets using random forest (RF) and gradient boosting machine (GBM) ensembles. Relative to RuleCOSI+, the median paired log-loss reduction is 71.9% for RF and 62.5% for GBM, with both differences remaining significant after Holm correction. The number of rules decreases by 38.7% for RF and 38.5% for GBM, while the primary tests do not detect a macro-F1 difference. Confidence-ECE also decreases for both ensembles, with statistical support for RF after correction. A separate controlled experiment with dedicated calibration data shows that native RCProb probabilities are competitive with RuleCOSI+ followed by temperature scaling, although additional post-hoc calibration can still improve RCProb. The results show that probability estimation is an important part of rule extraction and not only a post-processing step.

---

### RCProb: Probabilistic Rule Extraction for Efficient Simplification of Tree Ensembles

Authors: Josue Obregon
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.25304v1

Tree ensembles are widely used in industrial machine learning due to their strong predictive performance and efficient training procedures. However, as the number of trees in an ensemble grows, the resulting models become increasingly difficult for humans to interpret. To address this limitation, explainable artificial intelligence (XAI) studies methods that generate interpretable models capable of explaining complex predictors. One approach consists of extracting decision rules from tree ensembles while attempting to preserve the predictive performance of the original model. In previous work, we introduced RuleCOSI+, a greedy heuristic algorithm for extracting compact rule-based models from tree ensembles. Although RuleCOSI+ produces accurate and interpretable rule sets, it relies on repeated empirical frequency counting over the training data to estimate rule confidence, which becomes computationally expensive for large datasets. In this paper, we propose RCProb, a probabilistic reformulation of RuleCOSI+ designed to reduce the computational cost of rule extraction. RCProb estimates rule statistics using Dirichlet-smoothed class priors and Beta-smoothed condition likelihoods combined through a Naive Bayes formulation, avoiding repeated dataset scans. Experiments on 33 benchmark datasets show that RCProb maintains competitive predictive performance while reducing runtime by approximately $22\times$ compared with RuleCOSI+, while producing more compact rule sets on average.

---

### DenseScout: Algorithm-System Co-design for Budgeted Tiny Object Selection on Edge Platforms

Authors: Zhouzhi Xiong, Zimo Zeng, Yi Chen, Shuqi Xu, Yunfeng Yan, Donglian Qi
Published: 2026-04-28
Categories: cs.CV, eess.IV
arXiv: https://arxiv.org/abs/2604.25300v2

Deploying high-resolution tiny-object perception on edge platforms requires not only accurate localization, but also selecting a small set of informative patches under compute, transport, and latency constraints. We study budgeted tiny-object selection, where a frontend ranks patch centers from a lightweight proxy and a downstream detector processes only the selected regions. DenseScout is a 1.01M-parameter deployment-oriented dense-response selector that removes detector-style box regression and directly optimizes ranked patch-center prioritization. Its contribution lies in the task-specific selector formulation, the alignment among output representation, supervision, and decoding, and its joint design with transport-aware execution and QoS-oriented evaluation. Under unified protocols on VisDrone and DOTA, DenseScout provides stronger low-budget recall than the evaluated detector-derived selectors; controlled fixed-K inspection experiments further demonstrate advantages over selection-style proxy baselines. Cross-platform profiling on Jetson Orin NX and RK3588 shows that deployable utility depends jointly on selector quality, memory movement, and heterogeneous runtime realization. These results support treating edge tiny-object perception as a selection-and-deployment co-design problem rather than evaluating model accuracy and runtime in isolation.

---

### DenseScout: Algorithm-System Co-design for Budgeted Tiny Object Selection on Edge Platforms

Authors: Xiong Zhouzhi, Zimo Zeng, Yi Chen, Shuqi Xu, Yunfeng Yan, Donglian Qi
Published: 2026-04-28
Categories: cs.CV, eess.IV
arXiv: https://arxiv.org/abs/2604.25300v1

Deploying tiny object perception on edge platforms is challenging because practical systems must satisfy both strict compute budgets and end-to-end latency constraints. A common strategy is to first select a small number of candidate patches from a high-resolution image and then apply downstream processing only to the selected regions. However, existing detector-based frontends are not well aligned with this setting: strong offline detection accuracy does not necessarily yield effective low-budget patch prioritization, nor does it guarantee usable performance once transport and inference delays are considered. In this work, we study budgeted tiny object selection on edge platforms from a joint algorithm--system perspective. We present DenseScout, a lightweight dense-response selector with only 1.01M parameters, which directly ranks candidate patch locations from a high-resolution scene via a lightweight proxy input and is better aligned with low-budget tiny-object prioritization than detector-style frontends. To bridge offline selector quality and deployable utility, we further develop a transport-aware runtime realization on heterogeneous edge devices and adopt QoS-constrained recall, which counts a target as successfully perceived only if it is covered by the selected regions and the end-to-end processing finishes before the deadline. Experiments show that DenseScout consistently outperforms detector-based baselines in offline budgeted patch-selection evaluation, especially in low-budget regimes, while cross-platform results on RK3588 and Jetson Orin NX show that deployable performance depends jointly on selector quality and runtime realization efficiency. These results suggest that edge tiny object perception should be optimized as an algorithm--system co-design problem rather than as isolated model selection.

---

### The Thinking Pixel: Recursive Sparse Reasoning in Multimodal Diffusion Latents

Authors: Yuwei Sun, Yuxuan Yao, Hui Li, Siyu Zhu
Published: 2026-04-28
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2604.25299v1

Diffusion models have achieved success in high-fidelity data synthesis, yet their capacity for more complex, structured reasoning like text following tasks remains constrained. While advances in language models have leveraged strategies such as latent reasoning and recursion to enhance text understanding capabilities, extending these to multimodal text-to-image generation tasks is challenging due to the continuous and non-discrete nature of visual tokens. To tackle this problem, we draw inspiration from modular human cognition and propose a recursive, sparse mixture-of-experts framework integrated into conventional diffusion models. Our approach introduces a recursive component within joint attention layers that iteratively refines visual tokens over multiple latent steps while efficiently sharing parameters via sparse selection of neural modules. At each step, a gating network is devised to dynamically select specialized neural modules, conditioned on the current visual tokens, the diffusion timestep, and the conditioning information. Comprehensive evaluation on class-conditioned ImageNet image generation tasks and additional studies on the GenEval and DPG benchmark demonstrate the superiority of the proposed method in enhancing model image generation performance.

---

### LegalMidm: Use-Case-Driven Legal Domain Specialization for Korean Large Language Model

Authors: Youngjoon Jang, Chanhee Park, Hyeonseok Moon, Young-kyoung Ham, Jiwon Moon, Jinhyeon Kim, JuKyung Jung, Heuiseok Lim
Published: 2026-04-28
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2604.25297v1

In recent years, the rapid proliferation of open-source large language models (LLMs) has spurred efforts to turn general-purpose models into domain specialists. However, many domain-specialized LLMs are developed using datasets and training protocols that are not aligned with the nuanced requirements of real-world applications. In the legal domain, where precision and reliability are essential, this lack of consideration limits practical utility. In this study, we propose a systematic training framework grounded in the practical needs of the legal domain, with a focus on Korean law. We introduce LegalMidm, a Korean legal-domain LLM, and present a methodology for constructing high-quality, use-case-driven legal datasets and optimized training pipelines. Our approach emphasizes collaboration with legal professionals and rigorous data curation to ensure relevance and factual accuracy, and demonstrates effectiveness in key legal tasks.

---

### Learning from Medical Entity Trees: An Entity-Centric Medical Data Engineering Framework for MLLMs

Authors: Jianghang Lin, Haihua Yang, Deli Yu, Kai Wu, Kai Ye, Jinghao Lin, Zihan Wang, Yuhang Wu, Liujuan Cao
Published: 2026-04-28
Categories: cs.CL
arXiv: https://arxiv.org/abs/2604.25296v1

Multimodal Large Language Models (MLLMs) have shown transformative potential in medical applications, yet their performance is hindered by conventional data curation strategies that rely on coarse-grained partitioning by modality or department. Such fragmented approaches fail to capture the hierarchical and interconnected nature of clinical medical knowledge, limiting the models' ability to perform fine-grained recognition and complex reasoning. In this paper, we propose a novel Entity-Centric Medical Data Engineering framework. We automatically extract entities from authoritative medical literature to construct a Medical Entity Tree (MET), a hierarchical structure that systematically encodes diseases, anatomical structures, modalities, and symptoms into a unified knowledge repository. Building upon the MET, we propose an advanced data engine that includes: (1) node-guided retrieval to anchor raw data to specific medical concepts, (2) a two-stage hybrid filtering and alignment pipeline to ensure precise visual-semantic correspondence, and (3) knowledge-aware data synthesis to generate enriched captions and targeted reasoning VQA pairs, leveraging structural constraints. Extensive evaluations across six medical benchmarks demonstrate that our approach significantly enhances the medical capabilities of general-purpose MLLMs, improving their ability to handle complex clinical queries and achieve state-of-the-art performance in diverse medical contexts.

---

### Optimization-Free Topological Sort for Causal Discovery via the Schur Complement of Score Jacobians

Authors: Rui Wu, Hong Xie
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.25295v1

Continuous causal discovery typically couples representation learning with structural optimization via non-convex acyclicity penalties, which subjects solvers to local optima and restricts scalability in high-dimensional regimes. We propose a decoupled paradigm that shifts the causal discovery bottleneck from non-convex optimization to statistical score estimation. We introduce the Score-Schur Topological Sort (SSTS), an algorithm that extracts topological order directly from unconstrained generative models, bypassing constrained structure optimization. We establish that the causal hierarchy leaves a geometric signature within the score function: iterative graph marginalization is mathematically equivalent to computing the Schur complement of the Score-Jacobian Information Matrix (SJIM) under linear conditions. This translates the acyclicity constraint into an algebraic procedure with a dominant cost of O(d^3) operations. For non-linear systems, we formulate the expectation gap of Schur marginalization and introduce Block-SSTS to compress extraction depth, bounding structural error. Empirically, SSTS allows causal structural analysis on non-linear graphs up to d=1000. At this scale, our framework indicates that once the non-convex optimization bottleneck is mathematically bypassed, the structural fidelity of continuous causal discovery is bounded by the finite-sample estimation variance of the global score geometry. By reducing graph extraction to matrix operations, this work reframes scalable causal discovery from a constrained optimization problem to a statistical estimation challenge.

---

### Exploring Time Conditioning in Diffusion Generative Models from Disjoint Noisy Data Manifolds

Authors: Liuzhuozheng Li, Zhiyuan Zhan, Shuhong Liu, Dengyang Jiang, Zanyi Wang, Guang Dai, Jingdong Wang, Mengmeng Wang
Published: 2026-04-28
Categories: cs.LG, cs.CV
arXiv: https://arxiv.org/abs/2604.25289v1

Practically, training diffusion models typically requires explicit time conditioning to guide the network through the denoising sampling process. Especially in deterministic methods like DDIM, the absence of time conditioning leads to significant performance degradation. However, other deterministic sampling approaches, such as flow matching, can generate high-quality content without this conditioning, raising the question of its necessity. In this work, we revisit the role of time conditioning from a geometric perspective. We analyze the evolution of noisy data distributions under the forward diffusion process and demonstrate that, in high-dimensional spaces, these distributions concentrate on low-dimensional hyper-cylinder-like manifolds embedded within the input space. Successful generation, we argue, stems from the disentanglement of these manifolds in high-dimensional space. Based on this insight, we modify the forward process of DDIM to align the noisy data manifold with the flow-matching approach, proving that DDIM can generate high-quality content without time conditioning, provided the noisy manifold evolves according to the flow-matching method. Additionally, we extend our framework to class-conditioned generation by decoupling classes into distinct time spaces, enabling class-conditioned synthesis with a class-unconditional denoising model. Extensive experiments validate our theoretical analysis and show that high-quality generation is achievable without explicit conditional embeddings.

---

### The optimal betting wealth growth rate

Authors: Ashwin Ram, Aaditya Ramdas
Published: 2026-04-28
Categories: math.ST, math.PR, stat.ML
arXiv: https://arxiv.org/abs/2604.25280v1

This paper characterizes the best possible rate of growth of wealth in a Kelly betting game when repeatedly betting against a general i.i.d. null hypothesis $\mathscr{P}$, but the data are drawn i.i.d from an arbitrary alternative $Q$. We prove that it equals $\lim_{n \to \infty}n^{-1}\inf_{P \in (\mathscr P)^n)^{\circ\circ}} \mathrm{KL}(Q^n,P)$, where ${\mathscr P}^n = \{P^n: P \in \mathscr{P}\}$ and $(\mathscr {P}^n)^{\circ\circ}$ is its bipolar, i.e., this rate is achievable and one cannot do better. This quantity is in general smaller than a more popular quantity in the literature, $\mathrm{KL}_{\inf}(Q,\mathscr{P}) := \inf_{P \in \mathscr P}\mathrm{KL}(Q,P)$. If $\mathrm{KL}_{\mathrm{inf}}(\cdot,\mathscr P)$ is weakly lowersemicontinuous (w.l.s.c.) at $Q$, we show that the two quantities are equal; in particular, this happens when $\mathscr P$ is weakly compact. For simple alternatives, we provide the first matching necessary and sufficient condition for when power-one sequential tests exist (without assumptions on $\mathscr P, Q$). We also derive the optimal worst-case growth rate against composite $\mathscr Q$. We emphasize that test supermartingales on reduced filtrations suffice for all i.i.d. testing problems, and more general e-processes are not required. We thus completely generalize the recent results of Larsson et al.~\cite{larsson2025numeraire} to the sequential setting.

---

### OmniVTG: A Large-Scale Dataset and Training Paradigm for Open-World Video Temporal Grounding

Authors: Minghang Zheng, Zihao Yin, Yi Yang, Yuxin Peng, Yang Liu
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25276v1

Video Temporal Grounding (VTG), the task of localizing video segments from text queries, struggles in open-world settings due to limited dataset scale and semantic diversity, causing performance gaps between common and rare concepts. To overcome these limitations, we introduce OmniVTG, a new large-scale dataset for open-world VTG, coupled with a Self-Correction Chain-of-Thought (CoT) training paradigm designed to enhance the grounding capabilities of Multimodal Large Language Models (MLLMs). Our OmniVTG is constructed via a novel Semantic Coverage Iterative Expansion pipeline, which first identifies gaps in the vocabulary of existing datasets and collects videos that are highly likely to contain these target concepts. For high-quality annotation, we leverage the insight that modern MLLMs excel at dense captioning more than direct grounding and design a caption-centric data engine to prompt MLLMs to generate dense, timestamped descriptions. Beyond the dataset, we observe that simple supervised finetuning (SFT) is insufficient, as a performance gap between rare and common concepts still persists. We find that MLLMs' video understanding ability significantly surpasses their direct grounding ability. Based on this, we propose a Self-Correction Chain-of-Thought (CoT) training paradigm. We train the MLLM to first predict, then use its understanding capabilities to reflect on and refine its own predictions. This capability is instilled via a three-stage pipeline of SFT, CoT finetuning, and reinforcement learning. Extensive experiments show our approach not only excels at open-world grounding in our OmniVTG dataset but also achieves state-of-the-art zero-shot performance on four existing VTG benchmarks. Code is available at https://github.com/oceanflowlab/OmniVTG.

---

### Combating Visual Neglect and Semantic Drift in Large Multimodal Models for Enhanced Cross-Modal Retrieval

Authors: Guosheng Zhang, Linkai Liu, Keyao Wang, Haixiao Yue, Zhiwen Tan, Xiao Tan
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25273v1

Despite significant progress in Unified Multimodal Retrieval (UMR) powered by Large Multimodal Models (LMMs), existing embedding methods primarily focus on sample-level objectives via contrastive learning while overlooking the crucial subject-level semantics. This limitation hinders the model's ability to group semantically coherent subjects in complex multimodal queries, manifesting as semantic alignment deviation--where models fail to accurately localize salient text-referred regions in visual content. Moreover, without explicit guidance to model salient visual subjects, LMMs tend to over-rely on textual cues, resulting in visual modality neglect and suboptimal utilization of visual knowledge. To this end, we propose Salient Subject-Aware Multimodal Embedding (SSA-ME), a novel framework designed to enhance fine-grained representation learning through saliency-aware modeling. SSA-ME leverages LMMs and visual experts to identify and emphasize salient visual concepts in image-text pairs, and introduces a saliency-guided objective to better align cross-modal attention with semantically meaningful regions. Additionally, a feature regeneration module recalibrates visual features based on the derived saliency maps, ensuring a balanced and semantically coherent integration across modalities. Extensive experiments show that our method achieves state-of-the-art performance on the MMEB benchmark, demonstrating that incorporating subject-level modeling substantially improves multimodal retrieval. Comprehensive qualitative analyses further illustrate the interpretability and effectiveness of our approach.

---

### Spectral bandits

Authors: Tomáš Kocák, Rémi Munos, Branislav Kveton, Shipra Agrawal, Michal Valko
Published: 2026-04-28
Categories: stat.ML, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2604.25272v1

Smooth functions on graphs have wide applications in manifold and semi-supervised learning. In this work, we study a bandit problem where the payoffs of arms are smooth on a graph. This framework is suitable for solving online learning problems that involve graphs, such as content-based recommendation. In this problem, each item we can recommend is a node of an undirected graph and its expected rating is similar to the one of its neighbors. The goal is to recommend items that have high expected ratings. We aim for the algorithms where the cumulative regret with respect to the optimal policy would not scale poorly with the number of nodes. In particular, we introduce the notion of an effective dimension, which is small in real-world graphs, and propose three algorithms for solving our problem that scale linearly and sublinearly in this dimension. Our experiments on content recommendation problem show that a good estimator of user preferences for thousands of items can be learned from just tens of node evaluations.

---

### Online learning with Erdős-Rényi side-observation graphs

Authors: Tomáš Kocák, Gergely Neu, Michal Valko
Published: 2026-04-28
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2604.25271v1

We consider adversarial multi-armed bandit problems where the learner is allowed to observe losses of a number of arms beside the arm that it actually chose. We study the case where all non-chosen arms reveal their loss with a fixed but unknown probability $r$, independently of each other and the action of the learner. We propose two algorithms that work for different ranges of $r$. We show that after $T$ rounds in a bandit problem with $N$ arms, the expected regret of our first algorithm is $O(\sqrt{(T /r) \log N })$ whenever $r\ge(\log T)/(2N)$, while our second algorithm achieves a regret of $O(\sqrt{(T/r) \log (N+T)})$ for smaller values of $r$. We also give a quick estimation procedure that decides the range of~$r$. All our bounds are within logarithmic factors of the best achievable performance of any algorithm that is even allowed to know~$r$.

---

### Online combinatorial optimization with stochastic decision sets and adversarial losses

Authors: Gergely Neu, Michal Valko
Published: 2026-04-28
Categories: cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2604.25269v1

Most work on sequential learning assumes a fixed set of actions that are available all the time. However, in practice, actions can consist of picking subsets of readings from sensors that may break from time to time, road segments that can be blocked or goods that are out of stock. In this paper we study learning algorithms that are able to deal with stochastic availability of such unreliable composite actions. We propose and analyze algorithms based on the Follow-The-Perturbed-Leader prediction method for several learning settings differing in the feedback provided to the learner. Our algorithms rely on a novel loss estimation technique that we call Counting Asleep Times. We deliver regret bounds for our algorithms for the previously studied full information and (semi-)bandit settings, as well as a natural middle point between the two that we call the restricted information setting. A special consequence of our results is a significant improvement of the best known performance guarantees achieved by an efficient algorithm for the sleeping bandit problem with stochastic availability. Finally, we evaluate our algorithms empirically and show their improvement over the known approaches.

---

### Dynamic UGV-UAV Cooperative Path Planning in Uncertain Environments

Authors: Ninh Nguyen, Srinivas Akella
Published: 2026-04-28
Categories: cs.RO, cs.AI
arXiv: https://arxiv.org/abs/2604.25267v1

This paper addresses the Dynamic UGV-UAV Cooperative Path Planning (DUCPP) problem involving one unmanned ground vehicle (UGV) assisted by one or more unmanned aerial vehicles (UAVs) operating on an uncertain road network with potentially impassable edges. DUCPP is particularly relevant for scenarios such as disaster response, emergency supply transport, and rescue operations, where a UGV must reach a specified destination in the presence of partially unknown road conditions. To enable the UGV to travel safely and efficiently to its destination, the UAV(s) dynamically inspect edges in the environment to identify and prune damaged or impassable edges from consideration. We present multiple strategies, including a bidirectional approach, to optimize UGV-UAV cooperation for finding a safe path in an uncertain road network. Furthermore, we explore the impact of using multiple UAVs on reducing the UGV's travel time, and evaluate the associated computation time. The proposed strategies are implemented and evaluated on 100 urban road networks. The results demonstrate that the bidirectional strategy achieves the best performance in most instances, and using multiple UAVs further reduces UGV travel time at the expense of increased computation time. This paper presents a robust framework for DUCPP to achieve efficient UGV-UAV cooperation for path planning and inspection, offering practical solutions for navigation in challenging and uncertain conditions.

---

### DGLight: DQN-Guided GRPO Fine-Tuning of Large Language Models for Traffic Signal Control

Authors: Chenbo Yu
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.25259v1

Traffic signal control (TSC) plays a central role in reducing congestion and maintaining urban mobility. This dissertation introduces DGLight, a critic-guided reinforcement-learning framework for adapting a pretrained large language model to TSC. DGLight first trains a CoLight-based Deep Q-Network critic to estimate traffic-aware action values from structured intersection states, then uses the frozen critic to score candidate language-model actions and optimize the policy with Group Relative Policy Optimization (GRPO). The resulting controller maps traffic states to interpretable reasoning traces and signal decisions while learning from dense per-state supervision rather than raw cumulative environment rewards. Experiments on TSC benchmarks covering Jinan and Hangzhou show that DGLight is the strongest overall method among the compared LLM-based controllers, remains competitive with strong RL baselines, and transfers well to city datasets not used to fit the critic. Qualitative examples further show that the model's generated reasoning is interpretable and aligned with the chosen signal phase. The project code is available $\href{https://github.com/yyccbb/FYP_LLMTSC}{here}$.

---

### AutoResearchBench: Benchmarking AI Agents on Complex Scientific Literature Discovery

Authors: Lei Xiong, Kun Luo, Ziyi Xia, Wenbo Zhang, Jin-Ge Yao, Zheng Liu, Jingying Shao, Jianlyu Chen, Hongjin Qian, Xi Yang, Qian Yu, Hao Li et al.
Published: 2026-04-28
Categories: cs.AI
arXiv: https://arxiv.org/abs/2604.25256v1

Autonomous scientific research is significantly advanced thanks to the development of AI agents. One key step in this process is finding the right scientific literature, whether to explore existing knowledge for a research problem, or to acquire evidence for verifying assumptions and supporting claims. To assess AI agents' capability in driving this process, we present AutoResearchBench, a dedicated benchmark for autonomous scientific literature discovery. AutoResearchBench consists of two complementary task types: (1) Deep Research, which requires tracking down a specific target paper through a progressive, multi-step probing process, and (2) Wide Research, which requires comprehensively collecting a set of papers satisfying given conditions. Compared to previous benchmarks on agentic web browsing, AutoResearchBench is distinguished along three dimensions: it is research-oriented, calling for in-depth comprehension of scientific concepts; literature-focused, demanding fine-grained utilization of detailed information; and open-ended, involving an unknown number of qualified papers and thus requiring deliberate reasoning and search throughout. These properties make AutoResearchBench uniquely suited for evaluating autonomous research capabilities, and extraordinarily challenging. Even the most powerful LLMs, despite having largely conquered general agentic web-browsing benchmarks such as BrowseComp, achieve only 9.39% accuracy on Deep Research and 9.31% IoU on Wide Research, while many other strong baselines fall below 5%. We publicly release the dataset and evaluation pipeline to facilitate future research in this direction. We publicly release the dataset, evaluation pipeline, and code at https://github.com/CherYou/AutoResearchBench.

---

### Personalized Cross-Modal Emotional Correlation Learning for Speech-Preserving Facial Expression Manipulation

Authors: Tianshui Chen, Yujie Zhu, Jianman Lin, Zhijing Yang, Chunmei Qing, Feng Gao, Liang Lin
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25255v1

Speech-preserving facial expression manipulation (SPFEM) aims to enhance human expressiveness without altering mouth movements tied to the original speech. A primary challenge in this domain is the scarcity of paired data, namely aligned frames of the same individual with identical speech but different expressions, which impedes direct supervision for emotional manipulation. While current Visual-Language Models (VLMs) can extract aligned visual and semantic features, making them a promising source of supervision, their direct application is limited. To this end, we propose a Personalized Cross-Modal Emotional Correlation Learning (PCMECL) algorithm that refines VLM-based supervision through two major improvements. First, standard VLMs rely on a single generic prompt for each emotion, failing to capture expressive variations among individuals. PCMECL addresses this limitation by conditioning on individual visual information to learn personalized prompts, thereby establishing more fine-grained visual-semantic correlations. Second, even with personalization, inherent discrepancies persist between the visual and semantic feature distributions. To bridge this modality gap, PCMECL employs feature differencing to correlate the modalities, providing more precisely aligned supervision by matching the change in visual features to the change in semantic features. As a plug-and-play module, PCMECL can be seamlessly integrated into existing SPFEM models. Extensive experiments across various datasets demonstrate the superior efficacy of our algorithm.

---

### Below-Chance Blindness: Prompted Underperformance in Small LLMs Produces Positional Bias Rather than Answer Avoidance

Authors: Jon-Paul Cacioli
Published: 2026-04-28
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2604.25249v1

Detecting sandbagging--the deliberate underperformance on capability evaluations--is an open problem in AI safety. We tested whether symptom validity testing (SVT) logic from clinical malingering detection could identify sandbagging through below-chance performance (BCB) on forced-choice items. In a pre-registered pilot at the 7-9 billion parameter instruction-tuned scale (3 models, 4 MMLU-Pro domains, 4 conditions, 500 items per cell, 24,000 total trials), the plausibility gate failed. Zero of 12 model-domain cells showed significant below-chance performance under sandbagging instruction. Exploratory analyses revealed three qualitatively distinct failure modes. Qwen-2.5-7B and Phi-3.5-mini largely ignored the sandbagging instruction, with 62-88% response identity with the honest baseline. Llama-3-8B complied substantially but implemented underperformance as a positional heuristic, collapsing its response distribution onto middle-alphabet options (E at 31.8%, F at 26.1%) regardless of where the correct answer fell. This produced accuracy boosts of up to 33 percentage points when the correct answer coincidentally occupied the model's preferred position. An explicit anti-task instruction ("pick the least likely answer") drove two of three models below chance, with accuracy as low as 0.024. The capability for answer-aware avoidance therefore exists but is not activated by "deliberately underperform." BCB did not fail as a logical marker of answer-aware avoidance. It was not observed in this regime because the model showing the largest behavioural shift exhibited behaviour consistent with a position-dominant response policy rather than content-aware answer avoidance. We propose that positional-distribution shift may be a more effective behavioural signature than below-chance accuracy for detecting prompted underperformance at this model scale.

---

### Learning Structure, Energy, and Dynamics: A Survey of Artificial Intelligence for Protein Dynamics

Authors: Haocheng Tang, Liang Shi, Ya-Shi Zhang, Xixian Liu, Jian Tang, Jiarui Lu
Published: 2026-04-28
Categories: q-bio.BM, cs.LG
arXiv: https://arxiv.org/abs/2604.25244v1

Protein dynamics underlie many biological functions, yet remain difficult to characterize due to the high computational cost of molecular dynamics simulations and the scarcity of dynamic structural data. This survey reviews recent advances in artificial intelligence for protein dynamics from three perspectives: learning from structural ensembles and trajectories, learning from physical energy signals, and learning to accelerate molecular simulations. We summarize representative methods for conformation ensemble generation, trajectory generation, Boltzmann generators, physics-aware adaptation, machine learning potentials, coarse-grained modeling, and collective variable discovery. We further discuss available datasets and key open challenges, such as scalability, thermodynamic consistency, kinetic fidelity, and integration with experimental constraints.

---

### Categorical Optimization with Bayesian Anchored Latent Trust Regions for Structural Design under High-Dimensional Uncertainty

Authors: Zhangyong Liang, Jie Hou, Huanhuan Gao, Manyu Xiao
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.25241v2

Categorical structural optimization under aleatoric uncertainty is challenging because each design variable must be selected from a finite catalog of admissible instances, while each candidate design may require expensive stochastic finite-element evaluations. Existing latent-space optimization strategies can reduce the dimensionality of catalog attributes, but they often treat the reduced space as a continuous search domain. The resulting continuous optimum must then be rounded off to a nearby catalog instance, which may alter the objective value, constraint status, or physical interpretation of the design. To address this issue, this paper proposes the \textbf{C}ategorical \textbf{O}ptimization with \textbf{B}ayesian \textbf{A}nchored \textbf{L}atent \textbf{T}rust Regions (\textbf{COBALT}) framework for high-dimensional categorical Optimization Under Uncertainty. COBALT first embeds the physical catalog into a low-dimensional latent representation and locks the mapped instances as a discrete anchored graph. A data-independent random tree decomposition is then used to provide bounded-complexity additive modeling over high-dimensional categorical variables. On this anchored domain, an additive SAAS-GP surrogate is fitted to heteroscedastic MC-FEA observations, and a trust-region discrete graph acquisition search selects the next admissible catalog configuration without continuous relaxation or rounding-off. The proposed strategy is applied to robust design optimization of complex bar structures, considering structural weight, strain energy, and local buckling performance. By evaluating only valid catalog designs through the MC-FEA oracle, COBALT preserves physical admissibility throughout the active learning loop and improves the efficiency of robust categorical structural optimization.

---

### VLM Judges Can Rank but Cannot Score: Task-Dependent Uncertainty in Multimodal Evaluation

Authors: Divake Kumar, Sina Tayebati, Devashri Naik, Ranganath Krishnan, Amit Ranjan Trivedi
Published: 2026-04-28
Categories: cs.LG, cs.CL, cs.CV, stat.ML
arXiv: https://arxiv.org/abs/2604.25235v2

Vision-language models (VLMs) are increasingly used as automated judges for multimodal systems, yet their scores provide no indication of reliability. We study this problem through conformal prediction, a distribution-free framework that converts a judge's point score into a calibrated prediction interval using only score-token log-probabilities, with no retraining. We present the first systematic analysis of conformal prediction for VLM-as-a-Judge across 3 judges and 14 visual task categories. Our results show that evaluation uncertainty is strongly task-dependent: intervals cover ~40% of the score range for aesthetics and natural images but expand to ~70% for chart and mathematical reasoning, yielding a quantitative reliability map for multimodal evaluation. We further identify a failure mode not captured by standard evaluation metrics, ranking-scoring decoupling, where judges achieve high ranking correlation while producing wide, uninformative intervals, correctly ordering responses but failing to assign reliable absolute scores. Finally, we show that interval width is driven primarily by task difficulty and annotation quality, i.e., the same judge and method yield 4.5x narrower intervals on a clean, multi-annotator captioning benchmark. Code: https://github.com/divake/VLM-Judge-Uncertainty

---

### DRAGON: A Benchmark for Evidence-Grounded Visual Reasoning over Diagrams

Authors: Anirudh Iyengar Kaniyar Narayana Iyengar, Tampu Ravi Kumar, Gaurav Najpande, Manan Suri, Dinesh Manocha, Puneet Mathur, Vivek Gupta
Published: 2026-04-28
Categories: cs.CV, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2604.25231v2

Diagram question answering (DQA) requires models to interpret structured visual representations such as charts, maps, infographics, circuit schematics, and scientific diagrams. Recent vision-language models (VLMs) often achieve high answer accuracy on these tasks, yet correct answers do not guarantee that models ground their reasoning in the diagram regions that support the prediction. Models may instead rely on textual correlations or dataset artifacts without identifying the visual evidence required to verify the answer. This limitation prevents reliable evaluation of diagram reasoning and reduces interpretability. We introduce DRAGON, a benchmark for evaluating evidence-grounded visual reasoning in diagrams. Given a diagram, a question, and the correct answer, a model must predict bounding boxes that correspond to the visual elements required to justify the answer. These evidence regions may include answer-bearing components, textual labels, legends, axes, connectors, and other supporting structures involved in the reasoning process. The DRAGON dataset contains 11,664 annotated question instances from six diagram QA datasets: ChartQA, Circuit-VQA, InfographicsVQA, MapIQ, MapWise, and AI2D, with a 2,445-instance test set carrying human-verified evidence annotations and a standardized evaluation framework. Our evaluation on recent VLMs reveals that even Claude Opus 4.6 achieves a Grounding IoU of only 23.8% on AI2D, showing that faithful visual grounding remains an open challenge. DRAGON supports future research on models that ground their predictions in visual evidence.

---

### DRAGON: A Benchmark for Evidence-Grounded Visual Reasoning over Diagrams

Authors: Anirudh Iyengar Kaniyar Narayana Iyengar, Tampu Ravi Kumar, Gaurav Najpande, Manan Suri, Dinesh Manocha, Puneet Mathur, Vivek Gupta
Published: 2026-04-28
Categories: cs.CV, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2604.25231v1

Diagram question answering (DQA) requires models to interpret structured visual representations such as charts, maps, infographics, circuit schematics, and scientific diagrams. Recent vision-language models (VLMs) often achieve high answer accuracy on these tasks, yet correct answers do not guarantee that models ground their reasoning in the diagram regions that support the prediction. Models may instead rely on textual correlations or dataset artifacts without identifying the visual evidence required to verify the answer. This limitation prevents reliable evaluation of diagram reasoning and reduces interpretability. We introduce DRAGON, a benchmark for evaluating evidence-grounded visual reasoning in diagrams. Given a diagram, a question, and the correct answer, a model must predict bounding boxes that correspond to the visual elements required to justify the answer. These evidence regions may include answer-bearing components, textual labels, legends, axes, connectors, and other supporting structures involved in the reasoning process. The DRAGON dataset contains 11,664 annotated question instances collected from six diagram QA datasets: ChartQA, Circuit-VQA, InfographicsVQA, MapIQ, MapWise, and AI2D. We release a 2,445-instance benchmark test set with human-verified reasoning evidence annotations and a standardized evaluation framework. We evaluate eight recent VLMs and analyze their ability to localize reasoning evidence across diverse diagram domains. DRAGON enables systematic evaluation of diagram reasoning and supports future research on models that ground their predictions in visual evidence.

---

### Value-Sensitive AI for Prayer: Balancing the Agencies Between Human and AI Agents in Spiritual Context

Authors: Soonho Kwon, Dong Whi Yoo, Shaowen Bardzell, Younah Kang
Published: 2026-04-28
Categories: cs.HC, cs.AI
arXiv: https://arxiv.org/abs/2604.25230v1

We present four conceptual value-sensitive AI systems to examine how the presence of AI could influence praying experiences. Drawing on key values and practices associated with praying identified through a diary study, we designed AI systems intended to "assist" prayer practices. These designs were presented to participants through speculative design workbooks, serving as provocations to co-reflect on how the intervention of AI systems might shape their praying experiences. Our findings suggest that a sense of authenticity (or feeling a genuine connection to the divine) is a crucial value, while the presence of AI was often perceived as diminishing this authenticity, particularly when AI assumed too much agency in guiding praying practices. Based on our findings, we argue that AI system designs for deeply value-laden experiences should preserve users' agency in shaping their own experiences by maintaining interpretive openness, perhaps by leveraging AI's inexplicability as a resource for personal meaning-making or by recognizing non-use of AI as a legitimate design choice.

---

### ValueBlindBench: Agreement-Gated Stress Testing of LLM-Judged Investment Rationales Before Returns Are Observable

Authors: Sidi Chang, Peiying Zhu, Yuxiao Chen
Published: 2026-04-28
Categories: cs.AI, q-fin.CP
arXiv: https://arxiv.org/abs/2604.25224v2

LLM-based financial agents increasingly produce investment rationales before the outcomes needed to evaluate them are observable. This creates a delayed-ground-truth evaluation problem: realized returns remain the eventual arbiter of investment quality, but they arrive too late and are too noisy to guide many model-development and governance decisions. LLM judges offer a tempting shortcut for pre-deployment evaluation of AI-finance systems, but unvalidated judges may reward verbosity, confidence, or rubric mimicry rather than financial judgment. This paper introduces ValueBlindBench, a preregistered agreement-gated stress-test protocol for deciding when LLM-judged investment-rationale claims are publishable, qualified, or invalid. In a controlled market-state capital-allocation prototype with 1,000 honest decision cycles and 100 preregistered adversarial controls (1,100 trajectories, 5,500 judge calls), ValueBlindBench clears the aggregate agreement gate at \(\barκ_w = 0.7168\) but prevents several overclaims. Lower-rank systems collapse into a tie-class, one rubric dimension fails the per-dimension gate (\texttt{constraint\_awareness}, \(\barκ_w = 0.2022\)), single-judge rankings are family-dependent, and terse-correct rationales receive a \(Δ= -2.81\) rubric-point penalty relative to honest rationales. A targeted anchor-specificity probe further shows that financial constructs such as constraint awareness are operationally load-bearing. The scientific object is therefore not a leaderboard and not a claim to measure true investment skill. ValueBlindBench is a pre-calibration metrology layer for AI-finance evaluation: it governs whether a proposed LLM-judge-based investment-rationale claim is stable enough, agreed enough, and uncontaminated enough to be reported at all.

---

### DATAREEL: Automated Data-Driven Video Story Generation with Animations

Authors: Ridwan Mahbub, Syem Aziz, Mizanur Rahman, Mahir Ahmed, Shadikur Rahman, Shafiq Joty, Enamul Hoque
Published: 2026-04-28
Categories: cs.AI
arXiv: https://arxiv.org/abs/2604.25220v2

Data videos combine animated visualizations with synchronized narration to communicate quantitative information and are widely used in journalism, education, and public communication. Automatically generating them requires deciding what story to tell, designing effective visualizations, and producing executable animations synchronized with narration. Despite rapid progress in vision-language models (VLMs), it remains unclear how well they can perform this task from a high-level communicative intent, largely because no standardized benchmark exists. We introduce DATAREEL, a benchmark for automated data-driven video story generation containing 328 real-world data reels. Given a data table, a communicative intent, a target duration, and a style reference image, a model must generate executable animation code with synchronized subtitles, which we render and evaluate. Evaluating eight proprietary and open-weight VLMs reveals a substantial capability gap: open-weight models exhibit execution failure rates of up to 39.8%, whereas proprietary models achieve stronger overall performance. Yet even the best-performing models frequently generate static charts, subtitle-animation desynchronization, unstable layouts, and poor adherence to the reference style. We further introduce a strong agentic baseline that decomposes generation into planning, coding, and verification, consistently outperforming direct prompting in both human and automatic evaluations. The task remains far from solved; we release DATAREEL at https://github.com/vis-nlp/DataReel to support future work.

---

### DATAREEL: Automated Data-Driven Video Story Generation with Animations

Authors: Ridwan Mahbub, Syem Aziz, Mahir Ahmed, Shadikur Rahman, Mizanur Rahman, Shafiq Joty, Enamul Hoque
Published: 2026-04-28
Categories: cs.AI
arXiv: https://arxiv.org/abs/2604.25220v1

Data videos are a powerful medium for visual data based storytelling, combining animated, chart-centric visualizations with synchronized narration. Widely used in journalism, education, and public communication, they help audiences understand complex data through clear and engaging visual explanations. Despite their growing impact, generating data-driven video stories remains challenging, as it requires careful coordination of visual encoding, temporal progression, and narration and substantial expertise in visualization design, animation, and video-editing tools. Recent advances in large language models offer new opportunities to automate this process; however, there is currently no benchmark for rigorously evaluating models on animated visualization-based video storytelling. To address this gap, we introduce DataReel, a benchmark for automated data-driven video story generation comprising 328 real-world stories. Each story pairs structured data, a chart visualization, and a narration transcript, enabling systematic evaluation of models' abilities to generate animated data video stories. We further propose a multi-agent framework that decomposes the task into planning, generation, and verification stages, mirroring key aspects of the human storytelling process. Experiments show that this multi-agent approach outperforms direct prompting baselines under both automatic and human evaluations, while revealing persistent challenges in coordinating animation, narration, and visual emphasis. We release DataReel at https://github.com/vis-nlp/DataReel.

---

### When the Forger Is the Judge: GPT-Image-2 Cannot Recognize Its Own Faked Documents

Authors: Jiaqi Wu, Yuchen Zhou, Dennis Tsang Ng, Xingyu Shen, Kidus Zewde, Ankit Raj, Tommy Duong, Simiao Ren
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25213v1

OpenAI's GPT-Image-2 has effectively erased the visual boundary between authentic and AI-edited document images: a single number on a receipt can be replaced in under a second for a few cents. We release AIForge-Doc v2, a paired dataset of 3,066 GPT-Image-2 document forgeries with pixel-precise masks in DocTamper-compatible format, and benchmark four lines of defence: human inspectors (N=120, n=365 pair-votes via the public 2AFC site CanUSpotAI.com), TruFor (generic forensic), DocTamper (qcf-568, document-specific), and the same GPT-Image-2 model as a zero-shot self-judge -- asked, to avoid the trivial "image is mostly real" reading, whether any region was generated or edited by an AI image model. Human 2AFC accuracy is 0.501, indistinguishable from chance: even side-by-side, inspectors cannot tell GPT-Image-2 receipt forgeries from authentic counterparts. The three computational judges sit only modestly above (TruFor 0.599, DocTamper 0.585, self-judge 0.532). The self-judge fails consistently, not by chance: across five prompt strategies and four policies for handling ambiguous responses, AUC never rises above 0.59. To rule out the possibility that the two forensic detectors are broken on our source domain rather than blind to AI inpainting, we calibrate each on a same-domain traditional-tampering set built for its training distribution: TruFor reaches AUC 0.962 on cross-camera splicing of our dataset, DocTamper reaches 0.852 on cross-document OCR-token splicing with two-pass JPEG re-encoding. Both retain near-published performance on traditional tampering; switching to GPT-Image-2 inpainting drops AUC by 0.27-0.36 (0.962->0.599 TruFor; 0.852->0.585 DocTamper), isolating a detection gap specific to GPT-Image-2 inpainting. We release the dataset, pipeline, four-judge protocol, and calibration sets.

---

### DiRe-RAPIDS: Topology-faithful dimensionality reduction at scale

Authors: Alexander Kolpakov, Igor Rivin
Published: 2026-04-28
Categories: cs.LG, cs.AI, cs.SE, cs.SI
arXiv: https://arxiv.org/abs/2604.25209v2

Dimensionality reduction methods such as UMAP and t-SNE are central tools for visualising high-dimensional data, but their local-neighborhood objectives can preserve sampling noise while distorting global topology. We show that standard local metrics reward this noise memorisation: top-performing embeddings invent cycles and disconnected islands absent from the data. We introduce a topology-faithfulness benchmark based on noisy manifolds with known homology, tune DiRe against it, and find Pareto-optimal configurations that match or beat GPU-accelerated UMAP on classification while recovering exact first Betti numbers on stress tests. On 723K arXiv paper embeddings, DiRe preserves 3-4 times more topological structure than UMAP at comparable wall-clock.

---

### Towards Seamless Lunar Mosaics: Deep Radiometric Normalization for Cross-Sensor Orbital Imagery Using Chandrayaan-2 TMC Data

Authors: Pratincha Singh, Jai Gopal Singla, Prashant Hemrajani, Nitant Dube, Amithabh, Hinal Patel
Published: 2026-04-28
Categories: cs.CV, astro-ph.IM
arXiv: https://arxiv.org/abs/2604.25208v1

Radiometric inconsistencies remain a major challenge in generating seamless lunar mosaics from multi-mission orbital imagery due to variability in illumination geometry, sensor characteristics, and acquisition conditions. This paper presents a deep learning-based radiometric normalization framework for multi-mission lunar mosaics constructed primarily from ISRO's Chandrayaan-2 Terrain Mapping Camera (TMC) data, supplemented with auxiliary imagery from the SELENE (Kaguya) mission. The proposed approach employs a conditional generative adversarial network (cGAN) comprising a U-Net-based generator and a PatchGAN discriminator to learn a nonlinear radiometric mapping from conventionally mosaicked lunar imagery to a photometrically consistent reference derived from LROC Wide Angle Camera (WAC) data. A patch-based training strategy with overlap-aware inference is adopted to enable scalable processing of large-area mosaics while preserving structural continuity across tile boundaries. Quantitative evaluation using Structural Similarity Index (SSIM), Peak Signal-to-Noise Ratio (PSNR), and Root Mean Square Error (RMSE) demonstrates consistent improvements over traditional histogram-based normalization techniques. The proposed framework achieves enhanced tonal uniformity, reduced seam artifacts, and improved structural coherence across multi-source lunar datasets. These results highlight the effectiveness of learning-based radiometric normalization for large-scale planetary mosaicking and demonstrate its potential for generating high-fidelity lunar surface maps from heterogeneous orbital imagery.

---

### BARRED: Synthetic Training of Custom Policy Guardrails via Asymmetric Debate

Authors: Arnon Mazza, Elad Levi
Published: 2026-04-28
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2604.25203v1

Deploying guardrails for custom policies remains challenging, as generic safety models fail to capture task-specific requirements, while prompting LLMs suffers from inconsistent boundary-case performance and high inference costs. Training custom classifiers achieves both accuracy and efficiency, yet demands substantial labeled data that is costly to obtain. We present BARRED (Boundary Alignment Refinement through REflection and Debate), a framework for generating faithful and diverse synthetic training data using only a task description and a small set of unlabeled examples. Our approach decomposes the domain space into dimensions to ensure comprehensive coverage, and employs multi-agent debate to verify label correctness, yielding a high-fidelity training corpus. Experiments across diverse custom policies demonstrate that small language models finetuned on our synthetic data consistently outperform state-of-the-art proprietary LLMs (including reasoning models) and dedicated guardrail models. Ablation studies confirm that both dimension decomposition and debate-based verification are critical for ensuring the diversity and label fidelity required for effective fine-tuning. The BARRED framework eliminates the reliance on extensive human annotation, offering a scalable solution for accurate custom guardrails.

---

### Geometry of tail allocation in conformal prediction intervals

Authors: Tianying Wang
Published: 2026-04-28
Categories: stat.ME, math.ST, stat.ML
arXiv: https://arxiv.org/abs/2604.25202v2

Lower and upper errors of a two-sided conformal prediction interval can have different scientific consequences. The division of target miscoverage between the two endpoints determines the corresponding tail-specific guarantees and can alter interval length at first order when tail scales differ. We characterize this allocation-length relation after separate one-sided split calibration, which preserves the tail-specific guarantees and marginal coverage whenever the allocation is selected independently of the calibration sample. Tail-quantile response to proportional rescaling determines the resulting length geometry. For regularly varying tails, normalized length converges to $g_γ(c)=c^{-ξ}+γ(1-c)^{-ξ}$, where $c$ is the upper-tail allocation fraction, $ξ$ is the tail index, and $γ$ is the lower-to-upper tail-scale ratio. A dominant tail produces a boundary optimum and makes the equal-tail interval asymptotically $2^ξ$ times as long as the optimum. Comparable tails produce an interior optimum, with equal-tail allocation optimal only at matching scales. An empirical allocation rule attains the corresponding optimum without estimating tail parameters. In the de Haan class the effect moves to an additive scale. Calibration resolution determines whether ordinary ranks can realize these allocations. When calibration tail counts remain bounded, two-sided rank feasibility also constrains the allocation. Tail homogeneity transfers the length relation over covariates, while opposite dominant tails preclude one globally efficient allocation.

---

### Making AI-Assisted Grant Evaluation Auditable without Exposing the Model

Authors: Kemal Bicakci
Published: 2026-04-28
Categories: cs.CR, cs.AI, cs.CY, cs.LG
arXiv: https://arxiv.org/abs/2604.25200v1

Public agencies are beginning to consider large language models (LLMs) as decision-support tools for grant evaluation. This creates a practical governance problem: the model and scoring rubric should not be exposed in a way that allows applicants to optimize against them, yet the evaluation process must remain auditable, contestable, and accountable. We propose a TEE-based architecture that helps reconcile these requirements through remote attestation. The architecture allows an external verifier to check which model, rubric, prompt template, and input representation were used, without exposing model weights, proprietary scoring logic, or intermediate reasoning to applicants or infrastructure operators. The main artifact is an attested evaluation bundle: a signed, timestamped record linking the original submission hash, the canonical input hash, the model-and-rubric measurement, and the evaluation output. The paper also considers a scenario-specific prompt injection risk: applicant-controlled documents may contain hidden or indirect instructions intended to influence the LLM evaluator. We therefore include a canonicalization and sanitization layer that normalizes document representations and records suspicious transformations before inference. We position the design relative to confidential AI inference, attestable AI audits, zero-knowledge machine learning, algorithmic accountability, and AI-assisted peer review. The resulting claim is deliberately narrow: remote attestation does not prove that an evaluation is fair or scientifically correct, but it can make part of the evaluation process externally verifiable.

---

### Kohn-Sham Hamiltonian from Effective Field Theory: Quasiparticle Band Narrowing from Frozen Core Dynamics

Authors: Xiansheng Cai, Han Wang, Kun Chen
Published: 2026-04-28
Categories: cond-mat.mtrl-sci, cond-mat.str-el, cs.AI, cs.LG, physics.comp-ph
arXiv: https://arxiv.org/abs/2604.25199v1

Kohn-Sham (KS) eigenvalues are routinely compared with angle-resolved photoemission (ARPES) and used as input for many-body methods, yet density functional theory (DFT) assigns them no physical meaning. For alkali and alkaline-earth metals, KS bandwidths overestimate ARPES measurements by 20-35%, a discrepancy that persists across all exchange-correlation functionals. We construct an effective field theory (EFT) of the inhomogeneous electron gas and show that two conditions imply KS bands are the quasiparticle bands, up to a frozen-core renormalization factor zcore: a scale separation between core excitation energies and the valence Fermi energy, and an approximate Galilean invariance of the uniform electron gas confirmed by diagrammatic Monte Carlo. This factor reflects dynamical core excitations that conventional pseudopotentials freeze out and no static potential can capture. The correction 1-zcore reaches 20-35% for alkali metals but falls below 5% for Al and Si, explaining both the failure and success of KS band theory. We derive a closed-form post-SCF formula and validate it for Li, Na, K, Ca, Mg, Al, and Si; the predicted quasiparticle bands resolve the long-standing ARPES bandwidth discrepancy, matching embedded dynamical mean-field theory at negligible cost. This work also exemplifies first-principles agentic science, a direction particularly suited to the AGI-for-Science paradigm: an LLM-co-developed derivation with controlled approximations, verified symbolically and against a few experiments, becomes a deterministic harness for agentic scale-out, resolving simultaneously the LLM audit bottleneck and the non-falsifiability of fit-based AI-for-science.

---

### Knowledge-Data Dually Driven Paradigm for Accurate Landslide Susceptibility Prediction under Data-Scarce Conditions Using Geomorphic Priors and Tabular Foundation Model

Authors: Yuting Yang, Gang Mei, Feng Chen, Yongshuang Zhang, Jianbing Peng
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.25196v1

Landslide susceptibility prediction is critical for geohazard risk assessment and mitigation. Conventional data-driven paradigm achieves high predictive accuracy but require sufficient conditioning factors and large-scale landslide inventories. However, in practical engineering applications across mountainous and plateau regions, data-scarce conditions are commonly observed, where such data requirements are rarely satisfied, rendering conventional data-driven paradigm inapplicable. To address this issue, we propose a knowledge-data dually driven paradigm for accurate landslide susceptibility prediction under data-scarce conditions. The essential idea behind the proposed novel paradigm is the integration of the geomorphic prior knowledge with scarce landslide data. To validate the proposed paradigm, we first applied it to a data-rich region in central Italy, where a conventional data-driven paradigm trained on the full dataset served as the baseline. By utilizing only 30% of the available landslide data, the proposed paradigm achieved comparable predictive accuracy to the baseline, demonstrating its effectiveness under data-scarce conditions. The paradigm was further evaluated in a genuinely data-scarce environment for application, the Qilian Permafrost Region of the Tibetan Plateau, where it also yielded reliable susceptibility predictions, confirming its applicability under data-scarce conditions.

---

### How Can Reinforcement Learning Achieve Expert-level Placement?

Authors: Ruo-Tong Chen, Ke Xue, Chengrui Gao, Yunqi Shi, Tian Xu, Peng Xie, Siyuan Xu, Mingxuan Yuan, Chao Qian, Zhi-Hua Zhou
Published: 2026-04-28
Categories: cs.AR, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2604.25191v2

Chip placement is a critical step in physical design. While reinforcement learning (RL)-based methods have recently emerged, their training primarily focuses on wirelength optimization, and therefore often fail to achieve expert-quality layouts. We identify the reward design as the primary cause for the performance gap with experts, and instead of formalizing intricate processes, we circumvent this by directly learning from expert layouts to derive a reward model. Our approach starts from the final expert layouts to infer step-by-step expert trajectories. Using these trajectories as demonstrations or preferences, we train a model that captures the latent implicit rewards in expert results. Experiments show that our framework can efficiently learn from even a single design and generalize well to unseen cases.

---

### Image Classification via Random Dilated Convolution with Multi-Branch Feature Extraction and Context Excitation

Authors: Wentao Jiang, Yuanchan Xu, Heng Yuan
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.25188v1

Image classification remains a fundamental yet challenging task in computer vision, particularly when fine-grained feature extraction and background noise suppression are required simultaneously. Conventional convolutional neural networks, despite their remarkable success in hierarchical feature learning, often struggle with capturing multi-scale contextual information and are susceptible to overfitting when confronted with noisy or irrelevant image regions. In this paper, we propose RDCNet (Image Classification Network with Random Dilated Convolution), a novel architecture built upon ResNet-34 that integrates three synergistic innovations to address these limitations: (1) a Multi-Branch Random Dilated Convolution (MRDC) module that employs parallel branches with varying dilation rates combined with a stochastic masking mechanism to capture fine-grained features across multiple scales while enhancing robustness against noise and overfitting; (2) a Fine-Grained Feature Enhancement (FGFE) module embedded within MRDC that bridges global contextual information with local feature representations through adaptive pooling and bilinear interpolation, thereby amplifying sensitivity to subtle visual patterns; and (3) a Context Excitation (CE) module that leverages softmax-based spatial attention and channel recalibration to dynamically emphasize task-relevant features while suppressing background interference. Extensive experiments conducted on five benchmark datasets -- CIFAR-10, CIFAR-100, SVHN, Imagenette, and Imagewoof -- demonstrate that RDCNet consistently achieves state-of-the-art classification accuracy, outperforming the second-best competing methods by margins of 0.02%, 1.12%, 0.18%, 4.73%, and 3.56%, respectively, thereby validating the effectiveness and generalizability of the proposed approach across diverse visual recognition scenarios.

---