# Research Papers — 2026-06 · bundle page 19 of 260

> Full-text markdown of papers #901–950 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-06/page/19
[← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/18.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-06/page/20.md)

### Tastes without distinction: silicon samples and the synthetic construction of tastes

Authors: Xiangyu Ma, Mengmi Zhang, Shannon Ang, Minne Chen
Published: 2026-06-29
Categories: cs.CL, econ.GN
arXiv: https://arxiv.org/abs/2606.30085v2

Large-language models have proven to be remarkable if inconsistent parrots of public attitudes and opinions. The extent to which LLMs are able to produce reasonable approximations of cultural taste remains an open empirical question that becomes more urgent by the day, with market research companies already offering provisional 'synthetic' survey panels and the contamination of standard survey data from LLM-generated responses. In this study, we build on past work on silicon sampling by extending considerations of their ecological, relational, and positional fidelity in the doomain of cultural tastes. We use large-language models from OpenAI, Anthropic, and DeepSeek to produce 554,940 silicon surrogates of survey respondents from the Survey of Public Participation in the Arts (SPPA). We find these silicon surrogates' tastes to be highly stylized facsimiles of human tastes. First, silicon samples are super-omnivorous with a systematic postive-bias for liking. These individual-level bias of silicon samples are not well-explained by the WEIRD-bias often discussed in the literature. Second, the complex relationality in real taste structures is completely distorted among silicon samples. Third, very little of the known cultural alignment between tastes and social space are preserved. Silicon samples juvenilize age-taste associations, resurrect anachronistic class-taste associations, and caricaturize gender- and race-taste associations. Key words: AI, taste, consumption, culture, silicon sampling, meta-analysis.

---

### Not-quite-human tastes: the stylized omnivorousness of LLM survey surrogates

Authors: Xiangyu Ma, Mengmi Zhang, Shannon Ang, Minne Chen
Published: 2026-06-29
Categories: cs.CL, econ.GN
arXiv: https://arxiv.org/abs/2606.30085v1

Large-language models have proven to be remarkable if inconsistent parrots of public attitudes and opinions. The extent to which LLMs are able to produce reasonable approximations of cultural taste remains an open empirical question that becomes more urgent by the day, with market research companies already offering provisional `synthetic' survey panels and the contamination of standard survey data from LLM-generated responses. In this study, we build on past work on silicon sampling by extending considerations of its algorithmic fidelity and alignment to the domain of cultural consumption. We use large-language models from OpenAI, Anthropic, and DeepSeek to each produce 277,470 (30x9249) silicon surrogates of survey respondents from the Survey of Public Participation in the Arts (SPPA). We find these silicon surrogates' tastes to be highly stylized facsimiles of human tastes. (1) Silicon samples have a systematic postive-bias for liking, resulting in inflated ecological estimates of tastes. The individual-level bias of silicon samples are not well-explained by the WEIRD-bias often discussed in the literature. (2) The complex relationality in real taste structures is completely lost among silicon samples. (3) Finally, very little of the known cultural alignment between tastes and social space are preserved. Silicon samples attenuate age-taste associations, resurrect anachronistic class-taste associations, caricaturize gender- and race-taste associations.

---

### One Forward Beats Two: InnerZoom for Accurate and Efficient GUI Grounding

Authors: Chen Liu, Ling Chen, Hanzhang Zhou, Liangyu Chen, Chenglin Cai, Xin Yu, Steven Hoi, Yue Wang
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30084v1

MLLM-based GUI grounding methods commonly formulate target localization as autoregressive coordinate generation, enabling models to leverage the strong instruction-following and semantic understanding capabilities of MLLMs. However, this formulation requires the model to retain region-level target evidence while decoding coordinate tokens with the spatial precision demanded by GUI clicking. Our diagnostic analysis reveals that target-region awareness emerges in intermediate decoder layers but is neither retained nor translated into the final coordinate prediction. Existing ZoomIn-style methods address this issue through an external crop-and-rerun pass, which improves localization but increases end-to-end latency and computational cost. To retain the accuracy benefits of two-pass zooming without this extra cost, we propose InnerZoom, a single-forward framework for cross-layer evidence bridging. InnerZoom transforms target-related cues from the original forward pass into a compact cross-layer evidence state, then preserves, refines, and reinjects this state throughout later decoding layers to guide coordinate prediction. Extensive experimental results suggest that InnerZoom-4B achieves state-of-the-art performance on all six GUI grounding benchmarks, obtaining 64.7 on OSWorld-G, 40.2 on UI-Vision, 73.1 on OSWorld-GR, and 87.6 on MMBench-GUI, surpassing the previous best results by 4.1, 3.2, 2.9, and 2.3 points, respectively. Under a controlled 4B setting, InnerZoom improves the same SFT+RL baseline by 5.3 points on average and outperforms two-pass ZoomIn by 1.3 points on average, while reducing end-to-end latency by up to 31.8% and TFLOPs by about 29%. Code and models will be publicly available.

---

### Clinical Risk-Aware Multi-Level Grading for Coronary Artery Stenosis through Curved Feature Reconstruction

Authors: Shishuang Zhao, Hongtai Li, Junjie Hou, Yuhang Liu
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30082v1

Developing a multi-level grading model for coronary artery stenosis holds great clinical significance for the diagnosis of coronary artery disease. However, designing an effective multi-level deep learning algorithm faces significant challenges. Specifically, utilizing CCTA or 3D SCPR images alone presents inherent shortcomings: CCTA images are difficult to analyze due to the tortuous paths of blood vessels, while 3D SCPR images are prone to abnormal distortions that hinder accurate grading. Furthermore, different stenosis grades are associated with varying clinical risks, and incorporating this association into the algorithm is non-trivial. To address the former problems, we propose the Curved Feature Reconstruction (CFR) module, which uses vessel curves as prior and employs a point-by-point correspondence strategy to precisely align and fuse features from both 3D SCPR and CCTA images. Meanwhile, a Clinical Risk-Aware (CR) Loss is employed to introduce clinical risk relevance into the network training so that the algorithm can better align with the clinical diagnosis. The experimental results on a in-house dataset reveal that our approach significantly outperforms other methods, and several ablation studies also demonstrate the effectiveness of our proposed designs.

---

### Online Data Selection for Instruction Tuning via Gaussian Processes

Authors: Jun Wang, Quoc Phong Nguyen, Julien Monteil, Vu Nguyen
Published: 2026-06-29
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2606.30077v1

With Large Language Model (LLM) pre-training and fine-tuning shifting its focus from data volume to data quality, quality data selection has emerged as a critical research topic. Existing online data selection methods for LLM training are typically "batch-constrained", limiting optimization to local utility within random batches. To overcome this, we propose GAIA (Global Adaptive Instruction tuning via GAussian processes), a framework that formulates data valuation as a global estimation process. GAIA employs Gaussian Process regression to model continuous utility manifolds across the semantic space, utilizing an adaptive strategy fusion mechanism to dynamically prioritize high-utility samples. By casting the strategy-posterior update as an instance of the classical fixed-share Hedge framework for tracking the best expert, we inherit a dynamic-regret guarantee that characterizes GAIA's robustness under non-stationary quality scores during training. Empirical evaluations on three datasets demonstrate that GAIA significantly outperforms state-of-the-art baselines like \greats, establishing our method as a scalable and robust solution for efficient instruction tuning.

---

### ACPO: Agent-Chained Policy Optimization for Multi-Agent Reinforcement Learning

Authors: Daiki E. Matsunaga, Junho Na, Tri Wahyu Guntara, Scott Sanner, Pascal Poupart, Jongmin Lee, Kee-Eung Kim
Published: 2026-06-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2606.30072v2

Cooperative tasks in Multi-Agent Reinforcement Learning (MARL) require agents to collectively maximize a shared return. Under the Centralized Training with Decentralized Execution (CTDE) paradigm, policy gradients have remained difficult to compute directly. Prior methods largely follow two approaches: independent factorized updates with centralized critics, which lack general joint-improvement guarantees without value decomposition assumptions, or alternating best-response updates, which can converge to suboptimal Nash Equilibria. In this paper, we show the joint policy gradient admits an exact decentralized decomposition of per-agent terms, each formed from per-agent score functions and decentralized critics. Based on this decomposition, we develop Agent-Chained Policy Optimization (ACPO), where actors are trained independently, with their updates together constituting a single step on the joint policy gradient. Central to this result is a serialized view of the simultaneous joint decision in which agents commit actions one at a time, each conditioning on a belief over preceding actions that ties the independent per-agent updates into a single joint step. We evaluate on-policy and off-policy instantiations of ACPO on Multi-Robot Warehouse, SMACv2, and MA-MuJoCo, where it outperforms strong baselines, with the gap widening as the number of agents grows.

---

### Predictive Objectives Discard Exogenous Control-Relevant Features: A Controlled Mechanistic Study

Authors: Ayan Pendharkar
Published: 2026-06-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.30068v1

Joint-embedding predictive (JEPA-style) objectives learn representations by predicting future latents. In doing so they can discard features that are exogenous (uncontrollable by the agent) yet control-relevant, even when those features are trivially encodable. This occurs because the objective optimizes temporal predictability rather than control-relevance. We isolate this failure mode in a controlled 2x2 experimental design that varies feature controllability and relevance independently, using a predictability knob that decouples a feature's temporal predictability from its control-relevance. Comparing six objectives: reconstruction, JEPA, action-conditioned JEPA, controllability-based JEPA, inverse dynamics under a random policy, and reward-grounded JEPA, we observe that all evaluated reward-free predictive objectives leave the exogenous control-relevant feature near chance accuracy, while a reward-grounded variant retains it selectively. The remedy is label-efficient and robust: as little as 2% of reward-labeled transitions recovers the feature, the effect holds across two environments with different surface forms, and it persists across latent dimensions from 16 to 1024. Comparing the learned latent geometry against bisimulation theory's prediction, the JEPA latent realizes only a small fraction of the class separation a supervised reference attains.

---

### Neural Subspace Reallocation: Continual Learning as Retrieval-Based Subspace Memory Management

Authors: Byeong Hoon Yoon
Published: 2026-06-29
Categories: cs.LG, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2606.30067v1

We introduce Neural Subspace Reallocation (NSR), which reframes continual learning as memory management over parameter subspaces. Instead of treating Low-Rank Adaptation (LoRA) modules as disposable per-task adapters, NSR manages them as compressible, retrievable memory units on a frozen backbone through a recurring cycle: (1) compress learned LoRAs via SVD, (2) reserve them in a TaskKnowledgeBank, (3) recall related past LoRAs by embedding similarity to warm-start new or returning tasks, and (4) reallocate the active subspace accordingly, with distillation protecting prior tasks. We prove that in cyclic environments any memoryless allocation policy incurs cumulative regret Omega(T(M-1)Delta_switch) relative to a history-aware policy backed by the Bank (Theorem 1). Empirically, on Split-CIFAR-100 the Bank reduces cyclic recovery time by 10x, exactly as predicted, and on the heterogeneous 5-Datasets benchmark NSR achieves the highest accuracy and the least forgetting, about 9x closer to zero backward transfer than the memoryless heuristics. Crucially, we run a controlled study that isolates which component matters: holding the Bank fixed and varying only the allocation rule, we find that a simple similarity-based retrieval rule matches or beats a learned reinforcement-learning controller (recovering recurring tasks in 0 vs 1.8 steps and reaching equal accuracy). Our central, honest finding is therefore that the memory mechanism -- compression and similarity retrieval -- rather than a learned allocation policy, drives continual-learning performance under fixed capacity. A memory-budget analysis confirms the compressed Bank stays small -- 0.29 MB of parameter memory per task -- so a top-K retention cap bounds the total footprint while preserving fast recovery for retained tasks.

---

### Data-Driven Energy-Based Learning via Gibbs Measures on Hierarchical Structures

Authors: L. U. Abdullaev, F. Herrera, U. A. Rozikov, M. V. Velasco
Published: 2026-06-29
Categories: cs.LG, math.PR
arXiv: https://arxiv.org/abs/2606.30064v1

We introduce a data-driven probabilistic framework for learning systems based on Gibbs measures on hierarchical structures. Unlike standard empirical risk minimization, where a dataset is used to identify a single optimal parameter, our approach transforms the empirical loss function into an interaction potential defining an energy-based model. The resulting Gibbs distribution describes a family of equilibrium learning states generated by the data. We formulate the consistency conditions of the associated finite-volume distributions and derive nonlinear integral fixed-point equations whose solutions characterize the admissible learning states. These equations provide a rigorous connection between empirical loss landscapes and probabilistic inference on trees. For translation-invariant solutions, the problem reduces to the analysis of positive compact operators induced by data-dependent kernels, allowing us to establish existence and uniqueness conditions in the one-dimensional setting. Furthermore, we show that hierarchical learning systems may exhibit phase-transition phenomena: for certain empirical kernels on Cayley trees, multiple Gibbs measures emerge beyond a critical inverse temperature, corresponding to distinct equilibrium prediction regimes. Numerical experiments with non-separable kernels illustrate the appearance of multiple solution branches and demonstrate the coexistence of several data-induced learning states. Our results provide a new perspective on energy-based learning, where data do not merely determine an optimal model through minimization but define an entire probabilistic landscape of possible inference states.

---

### Little Brains, Big Feats: Exploring Compact Language Models

Authors: Dari Baturova, Elena Bruches, Ivan Chernov, Roman Derunets, Arsenii Fomin, Andrey Kostin
Published: 2026-06-29
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2606.30062v1

While large language models have been dominating the research landscape recently, small language models remain highly relevant across various domains; yet, they receive far less attention. In this study, we investigate how smaller language models perform during the generation stage within a Retrieval-Augmented Generation (RAG) system. To benchmark these models effectively, we utilised both open-source and proprietary datasets covering diverse subject areas and question types. Our findings demonstrate that a RAG system with small language models can be executed directly on-device without requiring any GPU hardware within a reasonable time. The experimental code and links to the supplementary materials can be accessed through the GitHub repository: https://github.com/SibNN/SLM-RAG-EVAL.

---

### From Failure Taxonomy to Intervention: A Diagnostic Methodology for Industry-Scale AVLM in Video and Live-Streaming Platform Moderation

Authors: Shuchang Ye, Jinqiang Yu, Zhujun Xiao, Yajing Kong, Yist Y. Lin, Yang Ma, Jiaxi Liu, Xiaolei Xu, Zheng Yu
Published: 2026-06-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.30059v1

Industry-scale video and live-streaming moderation imposes requirements that are difficult to satisfy with generic pretrained public models or external APIs, including adaptation to platform-specific data distributions, policy-specific objectives, and product-level safety constraints. As a result, platforms must undertake internal model development, naturally turning to shared public research for guidance. However, existing multimodal foundation-model studies primarily report architectures, training recipes, data scaling strategies, and benchmark results, but provide less systematic guidance on how failures should be localized and translated into targeted model-development interventions. Interventions are essential because deployment failures are rarely self-explanatory. Similar failures can originate from different causes. Without targeted interventions, improvement reduces to heuristic trial-and-error, where benchmark improvements are weakly attributable, and failures are difficult to trace to their underlying causes. To address this gap, we present a diagnostic methodology for industry-scale Audio-Visual-Language Models AVLM development. The methodology maps model failures into a taxonomy of observable failure signatures and links each class of failure to an intervention space. We instantiate this methodology across the development and alignment lifecycle of an AVLM foundation model for a large-scale video and live-streaming platform. The resulting system supports over 100 regions and is designed for noisy, ambiguous, and highly diverse content drawn from global platform traffic.

---

### Emergence of a Shared Canonical Object Frame from In-the-Wild Videos

Authors: Tom Fischer, Martin Sundermeyer, Adam Kortylewski, Eddy Ilg
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30058v1

Comparing object orientations and positions across different instances requires their poses to be expressed in a shared canonical frame. Establishing such frames has traditionally required manual annotation, creating a scaling bottleneck that limits category and instance diversity. We show that a shared canonical frame can instead emerge from self-supervised training on object-centric videos captured in the wild, using only noisy camera poses from Structure-from-Motion. Our key idea is to route all training sequences through a shared geometric bottleneck: a coarse canonical mesh that carries no category-specific detail. By learning dense correspondences from image pixels to this mesh, and estimating per-sequence alignments from noisy SfM geometry, a common canonical frame emerges from multi-view consistency and the semantic priors of the feature extractor, without any canonical pose labels or category conditioning. Trained in a self-supervised manner on 160,000 in-the-wild object videos, our method achieves competitive accuracy on category-level pose estimation benchmarks compared to methods that rely on canonical pose supervision. The code and checkpoint is available on https://github.com/Fischer-Tom/Emergent-Canonical-Frame/.

---

### Illuminating Unified Multimodal Model for Free-form Interleaved Text-Image Generation

Authors: Chonghuinan Wang, Zhikai Chen, Chunwei Wang, Yecong Wan, Junwei Yang, Zhixin Wang, Wei Zhang, Jiaqi Xu, Renjing Pei, Xiaohe Wu, Fan Li, Wangmeng Zuo
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30054v1

The advancement of generative AI models capable of producing text and image marks a critical step forward in the realm of multimodal intelligence, particularly for tasks involving the interleaving of both modalities. To advance this intelligence to the next stage, it is crucial for models to autonomously generate free-form interleaved text-image sequences. In this paper, we introduce ILLUME-X, an advanced unified multimodal paradigm that enables high-quality, free-form interleaved text-image generation by improving multimodal data efficiency and stabilizing the multimodal training process. ILLUME-X comprises three key components: (i) an expanded training data pipeline optimized for interleaved text-image generation, (ii) a progressive training strategy with self-adaptive objectives for free-length multimodal token sequences, and (iii) an objective and comprehensive evaluation method ILScore for interleaved text-image sequences. Notably, our ILLUME-X outperforms previous unified models across multiple interleaved text-image generation tasks like style transfer, image decomposition and storytelling.

---

### Notes on generative modeling: flow matching, diffusion, optimal transport and Schr{ö}dinger bridge

Authors: Titouan Vayer
Published: 2026-06-29
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2606.30053v1

These notes recapitulate the high level mathematical principles behind different techniques for generative modeling. I show the connections between optimal transport and standard techniques such as Schr{ö}dinger bridge and flow matching.

---

### Bridging the Gap Between Image Restoration and Navigational Safety in Hazy Conditions: A New Visibility Estimation Metric for Maritime Surveillance

Authors: Wentao Feng, Guobei Peng, Wengang Mao, Ryan Wen Liu
Published: 2026-06-29
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2606.30049v1

Visibility distance is critical to maritime navigational safety because it determines the effective observation range of shipborne and shore-based monitoring systems. Under hazy conditions, degraded visual information shortens observable distance and increases navigational risks and economic losses. Although numerous image dehazing methods have been developed, conventional image quality assessment metrics, such as PSNR, SSIM, FSIM, FADE, and NIQE, cannot establish a physically interpretable relationship between restoration quality and practical visibility thresholds. To address this limitation, this work proposes a visibility-oriented evaluation framework that links dehazing performance with visible-distance estimation. First, a Maritime Simulated Visibility Dataset (MSVD) is constructed using Unity3D to simulate maritime traffic scenes under graded visibility conditions. The dataset provides paired hazy and clear images with precise visibility annotations, enabling quantitative analysis of visibility restoration. Second, a dehazing visibility evaluation metric is developed by using object detection accuracy as an intermediate indicator. By establishing a mapping between visibility distance and detection performance, the proposed metric converts image restoration improvements into measurable visibility gains. Six representative dehazing methods are evaluated using both conventional image quality metrics and the proposed visibility metric. Experimental results under different imaging conditions demonstrate that MSVD provides a reliable benchmark for evaluating dehazing performance across graded visibility levels, while the proposed metric enables interpretable and reliable visible-distance estimation, thereby supporting the assessment of navigational safety and operational efficiency.

---

### Argus: Metric Panoramic 3D Reconstruction for Indoor Scenes

Authors: Xi Li, Linyuan Li, Yan Wu, Tong Rao, Kai Zhang, Xinchen Hui, Cihui Pan
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30047v3

Metric feed-forward 3D reconstruction for panoramic data remains under-explored due to the lack of large-scale panoramic RGB-D training data. We present Realsee3D, a hybrid dataset of 10K indoor scenes (1K real, 9K synthetic) with 299K panoramic viewpoints and precise metric annotations, and Argus, a feed-forward network trained on it for metric panoramic 3D reconstruction. In the sparse unordered capture setting of Realsee3D, a poorly chosen coordinate anchor can cause global pose drift. Argus addresses this with a learned covisibility module that selects the geometrically optimal reference view to anchor the metric world frame. To further improve multi-task learning, we decompose the bidirectional pixel-to-world mapping into interpretable sub-steps with per-step supervision and cross-coordinate joint constraints, reinforcing geometric consistency across prediction branches. On the Realsee3D benchmark, Argus achieves state-of-the-art metric performance in camera pose estimation, depth estimation, and point cloud reconstruction. Project page: https://argus-paper.realsee.ai.

---

### Walking in the Implicit: Interactive World Exploration via Neural Scene Representation

Authors: Zhiqi Li, Chengrui Dong, Zhenhua Du, Hangning Zhou, Cong Qiu, Hailong Qin, Mu Yang, Dongxu Wei, Peidong Liu
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30045v1

Interactive video generation systems for camera-controlled world exploration roll out growing sequences of latent video frames, entangling state transition with high-frequency observation synthesis. We propose Walking in the Implicit, a scene-centric paradigm that changes the rollout variable from frame latents to a fixed-length, renderable implicit state, termed Neural Implicit Scene (NIS). This factorizes interactive generation into stochastic transition of a compact scene state and deterministic pose-conditioned rendering given the sampled state. We instantiate this paradigm as NeuWorld: a transformer VAE learns locally anchored NIS from sparse posed frames, and a diffusion transformer evolves NIS conditioned on future camera trajectories and geometry-aware retrieved history. By reusing the VAE encoder as a unified conditioner, NeuWorld maps camera, reference-image, and history cues into the same NIS modality, avoiding external heterogeneous encoders. Trained from scratch on public posed-view data without pretrained video backbones or auxiliary 3D reconstructors, NeuWorld achieves strong long-horizon consistency with favorable inference efficiency.

---

### Building Multi-Task Agentic LLMs via Two-Phase Distillation

Authors: Huaijie Wang, Shusheng Xu, Yi Wu, Kaifeng Lyu
Published: 2026-06-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.30044v1

A key step toward artificial general intelligence is to train models that can perform multiple tasks. In this paper, we study how to build such models by first training separate RL experts for individual tasks and then consolidating them via distillation, as an alternative to directly training a single model on mixed tasks. We show that off-policy distillation degrades in multi-task settings due to the mode-covering nature of forward KL: aggregating data from multiple tasks introduces a large number of behavioral modes that can exceed the student's capacity, forcing it to average across behaviors and leading to degraded performance. In contrast, on-policy distillation is mode-seeking but requires strong initialization. Inspired by these observations, we propose a two-phase approach: off-policy distillation followed by on-policy refinement. Evaluation across conversational agents and text-based games confirms that this two-phase approach matches single-task RL expert performance for each individual task, whereas off-policy or on-policy distillation alone fails to match this performance.

---

### Heads, Not Backbones: Output Heads Dominate Architectures on Fat-Tailed Returns

Authors: Sichao He, Yansong Zhang
Published: 2026-06-29
Categories: cs.LG, q-fin.RM, q-fin.ST
arXiv: https://arxiv.org/abs/2606.30037v1

In a deep forecasting pipeline for fat-tailed financial returns at short horizons, which matters more - the backbone architecture or the output head? We compare four modern backbones (TimesNet, DLinear, N-BEATS, iTransformer) under three output heads: a point head, a single-Gaussian density head, and a Gaussian mixture density head with K=4 components. On S and P 500 monthly log-returns (1871-2023) under anchored walk-forward validation, the three heads form a strict gradient: switching from point to Gaussian improves CRPS by about 1.3 percent; switching from Gaussian to mixture adds a further about 2.4 percent. Switching between backbones, in contrast, changes CRPS by less than 1.5 percent on the point-head row and on the backbone-mean axis; density-head backbone spread is larger (up to 5.1 percent on the h=1 Gaussian row, driven by N-BEATS) but the head gradient (3.7 percentage points) still dominates. The Model Confidence Set on squared errors does not exclude any of the 12 variants at the 5 percent level: the head separates them only on distributional metrics (CRPS, pinball, coverage), not on squared error. The mixture head incremental value over a single Gaussian is largest in the highest-volatility regimes (13.9 percent in 1970s stagflation at h=12), confirming the mixture captures tail risk beyond what a unimodal Gaussian can express. The picture is horizon-dependent: the head dominates at short horizons, but at long horizons (h >= 6) the backbone re-takes the lead - an h-split we document against classical baselines (section 5.1). We conclude that on fat-tailed returns at short horizons, the head dominates the backbone, and the mixture distribution adds genuine value over a single Gaussian during crisis periods when risk-management decisions actually matter.

---

### Consensus Clustering of Free-Viewing Gaze Data: New Insights into Human-Information Interaction

Authors: Beryl Gnanaraj, Jaya Sreevalsan-Nair, Saqib Alam Ansari, Maanasa Rajaraman
Published: 2026-06-29
Categories: cs.CV, cs.HC, cs.LG
arXiv: https://arxiv.org/abs/2606.30035v2

Free-viewing gaze data provides a rich, task-free window into human visual attention. Conventional exploratory data analysis of the data provides user attention patterns through fixations and areas of interest. However, despite the richness of this gaze data, its human-information interaction (HII) patterns are understudied. We address this gap using consensus clustering of gaze data with respect to users and stimulus characteristics. We present a novel end-to-end unsupervised ensemble learning system for consensus clustering of free-viewing gaze datasets, EnsembleGaze. With a goal of characterizing the user behavior and stimulus type, we propose a feature engineering step based on statistical descriptors of fixation-based distributions. EnsembleGaze involves consensus voting of selected clustering methods implemented on the feature vector to compute the co-association matrix. Using the separate consensus clustering of users and stimuli as a baseline, we further propose two high-dimensional clustering strategies for determining gaze clusters based on joint user and image characterization. They are consensus subspace clustering and spectral biclustering. Clustering performance is evaluated using selected standard metrics and is further interpreted through image-level properties. Our system provides a replicable method for the unsupervised analysis of fixation behavior in scene perception research. Our results show that image stimuli groupings are highly consistent across methods, reflecting a robust ambient-versus-focal viewing mode distinction, whereas user groupings are image-context-dependent, a structure that only biclustering and the two-step conditional approaches are architecturally capable of recovering. Testing on the publicly available datasets revealed dataset-specific patterns, with each offering complementary insights through distinct clustering strategies.

---

### CogSENet: Blind Image Deblurring with Blur-Conditioned Semantic Routing and Explicit Frequency Fusion

Authors: Pan Wang, Yihao Hu, Xiujin Liu
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30030v1

Blind image deblurring demands the recovery of high-fidelity details and coherent structures from complex, unknown degradations. Current blind image deblurring methods struggle with real-world, spatially varying degradations, and lack the semantic awareness necessary to reliably differentiate valid textures from artifacts. To bridge this gap, we propose CogSENet, a dynamic, semantic-aligned reconstruction framework inspired by the eagle's visual system. By mimicking the eagle's active saccadic scanning, we devise a Semantic-Driven State Space Module (SDSSM) with semantic-aware token regrouping via differentiable routing, enabling prompt-conditioned long-range dependency modeling. To ensure physically interpretable recovery of textures and structures, a BiFreqFusionBlock (BFFB) mirrors functional differentiation of the eagle's retina by decomposing features into high and low frequencies using wavelet transforms. Finally, we estimate a continuous Blur Field (CBF) from blur image and fuse it with CLIP semantic priors to modulate the deepest latent features, emulating focal adaptation and enabling adaptive restoration under spatially non-uniform blur. Extensive experiments demonstrate that CogSENetoutperforms state-of-the-art deblurring methods in both visual quality and structural fidelity with fewer parameters, while also performing favorably on dehazing, deraining, and denoising tasks.

---

### Cross-Modal Iteration Distillation for Robust IHD Screening: The IDNet Framework and A New Benchmark

Authors: Yongchang Gao, Junjie Pang, Shuaiyu Yang, Yusheng Yang, Xichao Jia, Shaojie Li, Hongfei Zhang, Jia Mu
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30027v1

Color Fundus Photography (CFP) offers a low-cost and non-invasive route for ischemic heart disease (IHD) screening, but current studies are limited by scarce public benchmarks and ineffective fusion of retinal images with sparse clinical variables. We propose IDNet, a multimodal framework with a Cross-Modal Distillation Aggregator (CDA) that uses learnable queries to sequentially integrate left-eye, right-eye, and clinical features, mitigating the imbalance between high-dimensional visual features and low-dimensional tabular inputs. We also construct a reproducible UK Biobank benchmark with open-source curation and quality-control pipelines, yielding 50,410 images from 25,205 subjects. On this benchmark, IDNet outperforms image-only, clinical-only, and several multimodal baselines, and CDA consistently improves multiple visual encoders as a plug-in fusion module.

---

### MuseBench: Benchmarking Intent-Level Audiovisual Arts Understanding in MLLMs

Authors: Yuxuan Fan, Gyusik Seo, Jing Hao, Jaemin Cho, Mohit Bansal, Jaehong Yoon
Published: 2026-06-29
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2606.30026v1

Audiovisual arts encompass diverse creative disciplines, including cinema, visual arts, stage performance, and game design, where artistic meaning arises from deliberate combinations of visual, auditory, and narrative elements (e.g., fear amplified through claustrophobic framing, or grief conveyed through silence and lingering close-ups). True artistic understanding extends beyond recognizing what is depicted to reasoning about why it is expressed through particular creative choices. Despite the strong progress of multimodal large language models (MLLMs), this critical aspect of artistic understanding remains underexplored, as existing benchmarks largely measure perceptual recognition while overlooking reasoning about creative intent. To address this gap, we introduce Musebench, a comprehensive benchmark designed to evaluate MLLMs on nuanced artistic understanding. It comprises 4,016 questions spanning cinematic arts, static visual arts, stage performing arts, and game arts, distilled from over 10K candidate video essays that pair professional commentary with visual demonstration. To capture the open-ended nature of artistic analysis at scale, the benchmark combines single-select and variable-option multi-select questions. All questions are generated and refined through a four-phase iterative pipeline combining shortcut filtering, adversarial distractors, and expert validation. Comprehensive zero-shot evaluation of 28 state-of-the-art MLLMs reveals that even the best-performing model achieves only 48.29% accuracy, substantially below human expert performance of 87.18%, exposing a significant gap in current models' creative domain expertise.

---

### IBRSteG: Learning a Generalizable Steganography Framework for 3D Gaussian Splatting

Authors: Fanye Kong, Hongyu Xia, Yu Zheng, Boyang Gong, Jie Zhou, Jiwen Lu
Published: 2026-06-29
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2606.30024v1

Recent advances in deep learning have notably improved steganographic message hiding. However, designing a generalizable steganographic approach for 3D Gaussian Splatting (3DGS) that can embed meaningful 3D scene content remains challenging. In this paper, we propose IBRSteG, a generalizable framework for 3DGS steganography that enables undetectable concealment of secret scenes within a steganographic scene. Unlike existing approaches whose parameter generation is rigidly coupled with the specific scene, we formulate 3D steganography as a feed-forward 3D Gaussian embedding process that generalizes across different 3DGS scenes. To realize this, we introduce GAS (Gaussian Attributes Steganographer), a network that learns a scene-independent embedding function by injecting the attributes of secret 3D Gaussian points into a cover scene, thereby directly reconstructing the steganographic scenes without per-scene finetuning or optimization. By transforming 3D Gaussian into these structured attributes, these attributes are compatible with 2D learning paradigms and benefit from their structured nature, thereby enhancing generalization to unseen 3DGS scenes. Extensive experiments on established datasets demonstrate that IBRSteG can effectively conceal different scenes with high visual quality, and achieves superior capacity and security. Code is available at https://github.com/LingXiang2023/IBRSteG.

---

### Uncertainty Estimation in Pathology Foundation Models via Deep Mutual Learning

Authors: Gbègninougbo Aurel Davy Tchokponhoue, Sevda Öğüt, Ali Idri, Dorina Thanou, Pascal Frossard
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30020v1

Pathology foundation models (PFMs) offer generalizable representations for whole-slide image (WSI) analysis, yet their clinical adoption remains limited. Specifically, their predictions lack reliable confidence estimates, and no single PFM is universally best across tasks, which severely undermines trust in medical settings. To overcome this, we propose $\mathtt{DICE}$, a plug-and-play framework that ensembles $K$ frozen PFMs and models their disagreement as a proxy for uncertainty estimation. To ensure this proxy yields meaningful estimates, we align the ensemble members via deep mutual learning, and theoretically show that this objective upper-bounds the model uncertainty. Additionally, we demonstrate that the ensemble's consensus localizes abnormalities at the patch level without any explicit supervision. We evaluate $\mathtt{DICE}$ on three challenging WSI benchmarks. Notably, our framework provides reliable uncertainty estimates that accurately flag failure-prone cases under in- and out-of-distribution settings, while matching or outperforming SOTA baselines in classification, calibration, and localization. Overall, $\mathtt{DICE}$ takes a crucial step toward translating PFMs into uncertainty-aware decision-support systems.

---

### OmniDance: Multimodal Driven Dance Video Generation with Large-scale Internet Data

Authors: Kaixing Yang, Jiashu Zhu, Xulong Tang, Ziqiao Peng, Xiangyue Zhang, Chubin Chen, Puwei Wang, Jiahong Wu, Xiangxiang Chu, Hongyan Liu, Jun He
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30019v1

Music-driven dance video generation aims to synthesize expressive human motion that is temporally aligned with music while maintaining high visual fidelity. Despite recent progress, existing methods still face two key limitations: the lack of large-scale, high-quality dance video datasets, and the absence of principled frameworks for integrating music as a complementary conditioning signal into Video Generation Foundation Models. To address these limitations, we introduce CIPE-Dance, a large-scale Internet-sourced dance video dataset with choreography-informed text annotations, constructed via a progressive expert pipeline. To the best of our knowledge, CIPE-Dance is the largest dataset for dance video generation to date, comprising 300k high-quality clips over 400 hours and covering diverse dancers, environments, and dance genres. We further propose OmniDance, a framework-level recipe for integrating music into a TI2V foundation model without sacrificing its original controllability or visual fidelity. Motivated by the complementary roles of text as low-frequency semantics and music as high-frequency temporal dynamics, OmniDance co-designs a depth-aware specialization architecture, an anchored easy-to-hard curriculum learning strategy, and a modality-specialized time-dependent CFG strategy, enabling unified TI2V, MI2V, and MTI2V generation. Extensive experiments on CIPE-Dance demonstrate that OmniDance achieves state-of-the-art performance across all three tasks and exhibits robust multimodal integration capability. Project is available at https://github.com/AMAP-ML/OmniDance.

---

### Monte Carlo Energy Aggregation for Mobile 3D Gaussian Splatting

Authors: Xiaobiao Du, YuAn Wang, Hao Li, Bosheng Wang, Xun Sun, Xin Yu
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30017v1

Recent advances in 3D Gaussian Splatting have demonstrated unprecedented success in novel view synthesis. However, the substantial inference and storage overhead driven by high-order Spherical Harmonics (SH) are primary bottlenecks for mobile platforms. In this paper, we present Flux-GS, a real-time Gaussian Splatting method designed to achieve high-fidelity rendering with significantly reduced overhead for resource-constrained mobile platforms. We first propose a Monte Carlo Specular Energy Aggregator, sampling third-order radiance residuals and aggregating specular energy into a compact latent space. In this way, our method effectively preserves visually salient lighting features in lower-order bands without expensive distillation or pre-training. To mitigate the high-frequency details lost during compression, we introduce an Attribute-Conditioned SH Enhancement module. This module predicts Gaussian-aware offsets based on intrinsic Gaussian attributes, which enhance the first-order SH representation prior to inference, without extra inference costs. Furthermore, the original single-view gradient-based densification is prone to producing excessive Gaussians and overfitting to a certain view. We address these limitations by proposing a Multi-view Alpha-based Densification and Pruning strategy. By leveraging multi-view guidance, we ensure multi-view structure consistency and the precise removal of redundant primitives. Extensive experiments demonstrate that Flux-GS achieves substantial parameter reduction while maintaining competitive visual quality, offering a robust and scalable solution for real-time mobile rendering. Code: \textcolor{magenta}{\href{https://xiaobiaodu.github.io/flux-gs-project/}{https://xiaobiaodu.github.io/flux-gs-project/}}.

---

### Parametric Skills

Authors: Xuan Zhao, Haonan He, Qingyu Yang, Minglei Li, Jingqi Ye, Zelin Tan, Bo Wan, Peng Ye
Published: 2026-06-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2606.30015v1

Since intelligence fundamentally relies on efficient skill acquisition (Chollet, 2019), the ability to leverage skills is critical. For LLMs, skills, manually authored or extracted from task trajectories, are textual recipes encoding mature problem-solving experience and are critical to agentic capabilities. Despite widespread deployment, their utility is limited by the model's ability to comprehend and follow skill instructions, especially under complex and long-context scenarios, where key instructions are difficult to locate and adhere to. To address this limitation, we propose ParametricSkills, a framework that can convert free-form textual skills into parameters at test time, enabling context-free skill exploitation. Specifically, we first construct a large-scale, high-quality skill library, and synthesize single-turn and multi-turn skill exploitation trajectories built around these skills with OpenCode. Using these data, we then train a hypernetwork that parameterizes both the skill content and the test-time exploitation methodology by receiving textual skills and converting them into LoRA adapters. Experimental results on six complex software engineering (SWE) subtasks demonstrate that, the proposed ParametricSkills averagely outperforms in-context learning by 6.44 points as judged by DeepSeek-V4-Flash, while also achieving significantly higher BERT Score and F1 score, confirming its effectiveness. Beyond performance, we further find that parametric skills, being inherently accumulative, offer a preliminary yet promising avenue toward test-time continual learning.

---

### Shell-Supervised Gaussian Splatting for Urban Real-to-Sim Reconstruction

Authors: Yuan Yang, Peijun Lu, Fangzhou Lu, Sai Fan, Siqi Yan, Chenyuan Zhang, Haobo Liang, Yichen Wang
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30014v1

Real-to-sim reconstruction for embodied AI requires geometry that is useful for collision reasoning, navigation, and agent-environment interaction, not only photorealistic novel-view synthesis. However, close-range urban facades are difficult for video-to-3D reconstruction: glass, reflections, repeated windows, and weak texture can produce visually plausible renderings with unstable surface geometry. We introduce shell-supervised Gaussian Splatting, a reconstruction-stage framework that uses an external facade structural shell as lightweight geometric supervision for video-driven Gaussian reconstruction. The method aligns an exterior shell to the video reconstruction frame, renders per-view depth, camera-space normal, and valid-mask maps, and applies these cues through mask-gated losses during Gaussian optimization. This design preserves RGB-driven appearance while regularizing only visible shell-supported facade regions. Experiments on anonymized close-range urban facade scenes show improved facade orientation and visible-surface point-cloud consistency over photo-only, monocular-cue, and surface-oriented Gaussian baselines, while maintaining comparable held-out rendering quality.

---

### SkelEM: Training-Signal Decoupling of Skeleton and Diffusion for Self-supervised Axial Super-Resolution in Volume Microscopy

Authors: Bohao Chen, Yanchao Zhang, Yanan Lv, Chenxun Deng, Hua Han, Xi Chen
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30012v1

Volume microscopy, including electron and light microscopy, suffers from severe anisotropic resolution due to physical axial sectioning. Existing self-supervised axial super-resolution (ASR) methods face a trilemma bounded by overly smoothed regression textures, structural hallucinations of pure diffusion models, and prohibitive inference latency. In this paper, we propose Skeleton-refinE Microscopy (SkelEM), a self-supervised framework that decouples ASR at the training-signal level: a frozen topological network and a diffusion refiner are optimized by disjoint objectives, separating low-frequency topology formulation from high-frequency detail enhancement. Building on this deterministic skeleton, we exploit a unified cycle-consistent mechanism on input sparse slices to simultaneously extract a real-domain residual prior and bidirectionally align the diffusion refiner, washing away cross-plane artifacts without synthetic bias. By truncating the reverse diffusion process with this physical prior, SkelEM achieves high-fidelity detail restoration in merely $\le 5$ steps. To rigorously assess cross-instrument generalization, we further introduce BRAVE-ASR, a new benchmark of co-aligned anisotropic and isotropic volumes acquired on a Plasma-FIB instrument. Across public benchmarks, SkelEM achieves the most favorable balance across the fidelity-perception trade-off among self-supervised methods, with state-of-the-art downstream membrane segmentation performance and robust zero-shot generalization across distinct modalities.

---

### T3R: Deeper Test-Time Adaptation for Graph Neural Networks via Gradient Rotation

Authors: Huy Truong, Alexander Lazovik, Victoria Degeler
Published: 2026-06-29
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2606.30011v1

Graph Neural Networks (GNNs) deployed in real-world systems typically have fixed weights, often leading to degraded performance under distribution shifts. This issue can be mitigated by conventional fine-tuning, but in many real-world cases, collecting labeled data is expensive or infeasible. A potential approach is Test-Time Training (TTT), which adapts models' weights using unlabeled test data, yet it is typically limited to shallow updates that affect only a subset of model parameters. We propose T3R, leveraging multiple Rotograd matrices to improve task affinity between the target and auxiliary tasks, essential for effective test-time training. T3R further introduces a rotation technique that reorients self-supervised signals using these matrices to create surrogate gradients for the target task, allowing deeper adaptation across nearly the entire architecture. Empirically, T3R reduces MAE by 0.172 points over standard inference in regression datasets and achieves at least 9.37% relative improvement on cross-domain OGB classification benchmarks compared to models without adaptation. These results highlight the potential to develop an adaptation pipeline for graph-based systems, particularly in settings where conventional fine-tuning or retraining is infeasible.

---

### Node-to-Neighborhood Semantic Consistency: Text-Topology Alignment for TAGs Anomaly Detection

Authors: Bochen Lin, Jianxiang Yu, Jiayi Wu, Lin Qi, Huang Lu, Xiang Li
Published: 2026-06-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2606.30009v1

Graph anomaly detection (GAD) on text-attributed graphs (TAGs) is vital for applications such as fraud detection and academic integrity verification. Existing approaches generally fall into two paradigms. GNN-based methods effectively capture structural patterns but struggle to capture fine-grained textual semantics. Methods integrating LLMs with graphs improve semantic understanding yet fail to fully comprehend topological relationships among neighboring nodes. Moreover, both paradigms overlook the correspondence between textual semantics and graph topological relationships, limiting their ability to identify nodes whose semantics are inconsistent with their neighborhoods. In this paper, we formalize TAG anomaly detection as a node-to-neighborhood semantic consistency problem, where anomalies may arise from either textual semantic mismatch or topological deviation between a node and its neighbors. We propose N2NSC (Node-to-Neighborhood Semantic Consistency), a framework that captures the correspondence between graph topology and textual semantics through two complementary fusion paths. The two pathways work synergistically, enabling the LLM to fully leverage both textual and structural neighborhood information for anomaly detection. Extensive experiments across eight datasets demonstrate that N2NSC consistently outperforms current state-of-the-art methods.

---

### LLM Agents Are Latent Context Managers: Eliciting Self-Managed Context via State Proprioception

Authors: Binyan Xu, Haitao Li, Kehuan Zhang
Published: 2026-06-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2606.30005v5

Long-horizon tool agents are bottlenecked by how their context grows toward the limits of the context window. Recent systems make context management agent- or system-controlled, but they either learn compression policies that discard evidence or manage context in a layer the agent never sees. We argue that both miss a more basic gap: frontier language models are proprioceptively blind to their own context. From the prompt alone they cannot reliably infer block size, recency, or the remaining budget, all of which are needed for keep-or-archive decisions. We introduce VISTA (Visible Internal State for Tool Agents), a training-free, model-agnostic layer that represents working memory as typed addressable blocks, surfaces a runtime dashboard of token usage, recency, archive status, and remaining budget, and archives blocks as recoverable full-fidelity payloads. On LOCA-Bench, BrowseComp-Plus, and GAIA, the same untrained interface transfers across 1M-, 100K-, and 10K-scale trajectories. On LOCA-Bench it lifts Gemini-3-Flash from 22.7 to 50.7%, reaches 58.0% on BrowseComp-Plus, and remains competitive on GAIA. Gains grow with context pressure and transfer across backbones, while ablations confirm that the dashboard matters beyond archive and recovery tools.

---

### GeoEdit: Geometry-Aware Object Editing via Dual-Branch Denoising

Authors: Yi He, Jiangming Wang, Xinyu Wang, Mark Fong, Songchun Zhang, Yuxuan Xue, Hai-Tao Zheng, Yue Ma
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.30003v1

Precisely manipulating objects in a single photograph (translation, rotation, scaling) while obeying 3D physical constraints remains unsolved for diffusion-based editors. Current 2D methods lack spatial awareness and produce perspective violations. Forcing structural proxies into the latent space also disrupts variance homogeneity, and the resulting self-attention leakage leads to ghosting and background blur. The core difficulty is asymmetric: the relocated object must follow a rigid geometry, yet the uncovered background needs freedom to synthesize plausible content. We present GeoEdit, a training-free Lift-Manipulate-Render-Denoise pipeline that satisfies both constraints. We decouple scene and object in 3D, align them through point correspondence, and render a geometry-aligned proxy with a structural depth map. A Dual-Branch Denoising stage then refines this proxy: a video diffusion backbone preserves object identity, while 3D constraints are injected into the foreground within a narrow denoising window at matching noise variance (variance-homogeneous injection). The background denoises freely. Because the injected signal matches the native latent statistics, self-attention stays undisturbed. We also introduce GeoEditBench, a pose-aware benchmark covering object translation, object rotation, and camera movement with pose-aware evaluation metrics. Experiments confirm consistent gains in geometric accuracy, identity fidelity, and background quality. Our codes are available at https://github.com/Heey731/GeoEdit.

---

### SICAGE: Speaker-Independent Culture-Aware Gesture Generation using TED4C-L Dataset

Authors: Ariel Gjaci, Antonio Sgorbissa, Vittorio Murino
Published: 2026-06-29
Categories: cs.CV, cs.GR, cs.HC, cs.SD
arXiv: https://arxiv.org/abs/2606.30001v1

Recent co-speech gesture generation methods often overlook cultural differences, limiting their effectiveness in human-agent interaction. Moreover, culture-conditioned models are rarely evaluated under speaker-disjoint splits, so apparent "cultural" behavior may be confounded with speaker-specific gesturing style. We introduce SICAGE, a modular framework for culture-aware co-speech gesture generation that conditions motion synthesis models on speaker-independent cultural representations. SICAGE learns these representations from audio and text by treating each speaker as a separate domain while imposing invariance across speakers. This encourages representations to remain culture-discriminative while reducing dependence on speaker identity. The resulting cultural embeddings condition a multimodal generator to produce culturally appropriate gestures. We instantiate this idea with two domain generalization approaches: adversarial learning and Fishr regularization. We further introduce ALaDiT, a real-time diffusion-based gesture generator designed to efficiently incorporate the learned cultural embeddings. To validate our method, we built TED4C-L, a 106-hour multimodal dataset of 764 TED speakers from four cultural groups. Experiments show that SICAGE improves motion realism, diversity, beat synchronization, semantic relevance, and cultural consistency.

---

### AlgoSkill: Learning to Design Algorithms by Scheduling Human-Like Skills

Authors: Xinyuan Song, Zekun Cai, Liang Zhao
Published: 2026-06-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2606.29999v1

Designing an algorithm from a natural-language problem statement requires identifying the problem structure, reading constraints, choosing a suitable paradigm, checking correctness, and refining complexity. Existing large language model (LLM) methods often rely on direct generation or generic self-refinement, leaving these steps implicit. We propose AlgoSkill, which models algorithm design as sequential decision-making over a typed library of algorithmic skills, including abstraction, constraint analysis, state design, data-structure selection, proof checking, counterexample construction, and complexity refinement. A learned scheduler proposes skills from the current design state, while a Monte Carlo Tree Search (MCTS) controller explores skill sequences using verification feedback from compilation, testing, stress testing, and complexity analysis. Experiments on competitive programming and combinatorial optimization benchmarks show that AlgoSkill improves over direct LLM generation, chain-of-thought prompting, self-refinement, and MCTS without typed skills. Ablations show that typed skills, verification-based repair, and search-based scheduling each contribute to performance. These results support treating automatic algorithm design as verification-guided skill scheduling rather than one-shot code generation.

---

### Rigel: Self-Distilled Score Adaptation for Image and Video Captioning Evaluation

Authors: Shuitsu Koyama, Kazuki Matsuda, Yuiga Wada, Shinnosuke Hirano, Daichi Yashima, Komei Sugiura
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.29997v2

Automatic evaluation of image and video captioning is essential for benchmarking multimodal systems, although standard evaluation metrics show limited alignment with human judgments. Recent approaches using large language models (LLMs), commonly referred to as LLM-as-a-Judge, have improved alignment with human judgments but still suffer from a mismatch between large-vocabulary language modeling and evaluation over a small label set. To address this, we propose Rigel, an automatic evaluation metric for image and video captioning, based on self-distilled score adaptation. The metric employs an evaluation-specific scoring head distilled from a frozen LLM, which captures judgment signals in a task-aligned space without relying on large-vocabulary token sets. We then refine the LLM backbone with human judgment data. To train Rigel, we constructed the Vid-Lepus dataset, which contains 3,338 video clips, 33,380 reference captions, and 5,637 candidate captions. Experiments on multiple benchmarks show that Rigel outperforms state-of-the-art metrics, achieving over 10-point improvements on ActivityNet-Fact in the reference-free setting.

---

### Rigel: Self-Distilled Score Adaptation for Image and Video Captioning Evaluation

Authors: Shuitsu Koyama, Kazuki Matsuda, Yuiga Wada, Shinnosuke Hirano, Daichi Yashima, Komei Sugiura
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.29997v1

Automatic evaluation of image and video captioning is essential for benchmarking multimodal systems, although standard evaluation metrics show limited alignment with human judgments. Recent approaches using large language models (LLMs), commonly referred to as LLM-as-a-Judge, have improved alignment with human judgments but still suffer from a mismatch between large-vocabulary language modeling and evaluation over a small label set. To address this, we propose Rigel, an automatic evaluation metric for image and video captioning, based on self-distilled score adaptation. The metric employs an evaluation-specific scoring head distilled from a frozen LLM, which captures judgment signals in a task-aligned space without relying on large-vocabulary token sets. We then refine the LLM backbone with human judgment data. To train Rigel, we constructed the Vid-Lepus dataset, which contains 3,338 video clips, 33,380 reference captions, and 5,637 candidate captions. Experiments on multiple benchmarks show that Rigel outperforms state-of-the-art metrics, achieving over 10-point improvements on ActivityNet-Fact in the reference-free setting.

---

### Are We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning

Authors: Sangmook Lee, Minbeom Kim, Jeonghye Kim, Dohyung Kim, Sojeong Rhee, Kyomin Jung
Published: 2026-06-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2606.29985v2

Diversity in LLM mathematical reasoning is critical for exploration, but common diversity metrics mostly capture surface-level variation rather than differences in how a problem is solved. We address this gap by introducing approach-level diversity: variation in strategies across correct solutions to the same problem. Using a human-calibrated LLM judge framework, we show that prior diversity measures are unreliable proxies for approach-level diversity, and this mismatch carries over to diversity-aware RLVR, where target metrics are preserved while approach-level diversity declines. Investigating when approach-level diversity helps and whether it can be directly induced, we find that approach-diverse candidate sets improve test-time scaling. However, optimizing an LLM judge diversity reward during training causes the policy to exploit judge-specific preferences rather than broaden its approaches, leaving direct optimization of approach-level diversity as an open problem. Together, our work introduces the notion of approach-level diversity and uncovers a systematic divergence between surface- and approach-level signals, marking a step toward LLMs that reason in genuinely diverse, human-like ways.

---

### Are We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning

Authors: Sangmook Lee, Minbeom Kim, Jeonghye Kim, Dohyung Kim, Sojeong Rhee, Kyomin Jung
Published: 2026-06-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2606.29985v1

Diversity in LLM mathematical reasoning is critical for exploration, but common diversity metrics mostly capture surface-level variation rather than differences in how a problem is solved. We address this gap by introducing approach-level diversity: variation in strategies across correct solutions to the same problem. Using a human-calibrated LLM judge framework, we show that prior diversity measures are unreliable proxies for approach-level diversity, and this mismatch carries over to diversity-aware RLVR, where target metrics are preserved while approach-level diversity declines. Investigating when approach-level diversity helps and whether it can be directly induced, we find that approach-diverse candidate sets improve test-time scaling. However, optimizing an LLM judge diversity reward during training causes the policy to exploit judge-specific preferences rather than broaden its approaches, leaving direct optimization of approach-level diversity as an open problem. Together, our work introduces the notion of approach-level diversity and uncovers a systematic divergence between surface- and approach-level signals, marking a step toward LLMs that reason in genuinely diverse, human-like ways.

---

### Be Faithful When Response: Returning Fluent and Grounded Answers for Vision-Language Models Reinforcement Learning

Authors: Peng, Lee, Yin Zhang, Yanglin Zhang, Haonan Wu, Zishan Liu, Ruoxi Zang, Xin Zhu, Jiayin Zheng, Jian Yao, Zefeng Ji, Fei Ma
Published: 2026-06-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2606.29984v1

Reinforcement Learning (RL) is an important paradigm for improving the reasoning capabilities of Vision-Language Models (VLMs). However, directly applying RL to rollout multimodal reasoning can lead to instability, due to the exploitation of language priors, the neglect of visual evidence, and the generation of reasoning traces that are fluent yet not visually grounded. The question arises: Can initially steer the policy toward visually faithful reasoning regime before applying reinforcement learning? To this end, we propose a Faithful Warm-Start (FWS) strategy that first curates samples with explicit vision-language causal relationships from six general VQA benchmarks to construct the FaithfulQA dataset, where each of the image-question pairs gains a certain degree of visual observations, question requirements, commonsense knowledge, domain knowledge, and the final answer. Subsequently, a VLM-based judge is employed to further purify the dataset, ensuring strong causal consistency and visual faithfulness. This warm-start stage equips the model with the capability to understand causally grounded vision-language patterns before subsequent RL optimization under sparse answer-level rewards. Experimental results show that such faithful supervision improves answer accuracy, stabilizes RL training, and reduces visually unsupported reasoning.

---

### Stabilizing Extrapolation in Looped Transformers via Learned Stochastic Stopping

Authors: Hsun-Yu Kuo, El Mahdi Chayti, Patrik Reizinger, Wieland Brendel, Martin Jaggi
Published: 2026-06-29
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2606.29983v1

Looped Transformers, which repeatedly apply a shared transformer block, are an architecturally natural fit for variable-length algorithmic tasks. Although they can exhibit strong length generalization beyond the length of training sequences, this behavior is brittle, yielding high out-of-distribution (OOD) variance, even across well-performing in-distribution solutions. We trace this variance to the spurious correlation in simple algorithmic tasks between sequence length and number of loops. Introducing stochasticity into the number of loops during training sharply reduces OOD variance and stabilizes predictions across inference-time loop counts. To improve upon heuristic randomization schemes, we further analyze RL-Halting as a learned stochastic schedule and find that it generally improves the accuracy-stability trade-off. Across binary addition, Dyck-1, Unique Set, and Copy, learned stochastic stopping often improves this trade-off but can also stabilize a suboptimal computation. Our work suggests that "when to stop" should be treated as a training-time design choice, not merely an inference-time computation-allocation rule.

---

### Exploration and Online Transfer with Behavioral Foundation Models

Authors: Louis Bagot, Mathieu Lefort, Laëtitia Matignon
Published: 2026-06-29
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2606.29980v2

Zero-shot Transfer in Reinforcement Learning (RL) aims to train an agent that can generate optimal policies for any reward function, without additional learning at transfer time, while training only on reward-free trajectories. For their generality over tasks, such models are sometimes called ``Behavioral Foundation Models'' (BFMs). While they have shown strong performances and improvements in recent years, the current framework and algorithms still assume that, during the transfer phase, the agent is informed offline about the reward (the task to solve) through a dataset of state-reward pairs, which it uses to pick the best policy to deploy. However, in practice if the reward is a black-box (e.g. direct user feedback), it is not possible to generate such a dataset: it is necessary to observe the reward through interactions with the environment. In other words, the current framework of offline transfer is not aligned with the traditional RL setting of online learning through trial-and-error, which requires exploration in order to find rewards. This paper proposes to tackle this new online transfer in zero-shot RL, with the key insight that the BFM itself can be used to generate exploration policies. We show that it is possible to frame this online learning problem in terms of a bandit-like exploration-exploitation problem. More precisely, at each step the bandit algorithm recommends a policy, the BFM executes it in the environment, which yields a reward and a new state; we repeat the process until we converge to the optimal policy. In the popular context of linear reward approximation, we derive a formulation inspired by Upper Confidence Bound and show that exploration can be achieved through the minimization of the eigenvalues of an uncertainty matrix. We evaluate qualitatively and quantitatively our framework on a simple environment to validate the concept of our method.

---

### A multi-architecture study of specificity refinement and false-positive mechanism analysis in prostate MRI

Authors: Yongbo Shu, Kewen Chen, Yifeng Yuan, Zirui Xin, Luo Lei, Yang Yang, Xi Chen, Aijing Luo
Published: 2026-06-29
Categories: eess.IV, cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2606.29977v1

Objectives: To characterize residual false positives in prostate MRI detection, and to evaluate a lightweight post-hoc refinement head for case-level specificity. Materials and Methods: This retrospective study used PI-CAI (5-fold cross-validation) and Prostate158 (n=158; external). A context-aware evidence head and an 89,216-parameter refinement head were trained on a frozen detection backbone; the evidence head was also trained on four further backbones (bare nnU-Net, bare U-Net, bare Mamba, MIGF-Mamba). For each false-positive region, T2-weighted, apparent-diffusion-coefficient, and high-b-value contrast ratios versus peri-lesional rings were compared against ground-truth lesions and contralateral benign regions. Results: False positives were closer to true cancers than to benign tissue in evidence and raw T2-weighted and apparent-diffusion-coefficient contrast, reproducing 35/35 across five architectures (Cohen's d 1.10; FP/benign evidence ratio 2.38x) and 105/105 across modality-perturbation scenarios. On PI-CAI fold-0, refinement raised case-level specificity from 0.469 to 0.549 (+17.2%) at preserved sensitivity (0.943); 5-fold cross-validation showed fold-conditional behavior (9/15 observations positive; range -22% to +28%). On Prostate158, both models saturated (McNemar pooled p=0.69), while the false-positive contrast-matching finding replicated. Conclusion: Residual false positives are contrast-matched to cancer (sharing raw imaging features rather than histologically confirmed mimicry), reproducing across five architectures -- a data-level imaging property, not model-specific artifacts; post-hoc refinement adds practical specificity in-domain but is fold-conditional.

---

### Learning Efficient 4D Gaussian Representations from Monocular Videos with Flow Splatting

Authors: Shengjun Zhang, Jinzhao Li, Xin Fei, Yueqi Duan
Published: 2026-06-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.29976v1

Reconstructing dynamic 3D scenes from monocular videos is challenging due to scene complexity and temporal dynamics. With the advancement of 3D Gaussian Splatting in novel view synthesis, existing methods extend 3D Gaussians to 4D domain with deformation fields, trajectories or spatiotemporal 4D volumes to model scene element deformation. However, these methods suffer from long training time, low rendering speed or high memory consumption for per-frame reconstruction of 4D volumes, without fully exploiting dense dynamic information. To address this issue, we propose Flow Splatting, which constructs the velocity field and enables the conventional splatting technique to render optical flow from the velocity field to supervise dynamics learning process from monocular videos. Specifically, we extend 4D volumes with time varying means and covariance to represent complex dynamics. Then, we construct and approximate the velocity field naturally based on this representations. While conventional volume rendering techniques support to render color fields, we extend the volume rendering strategy to splat the velocity field by considering the influence of camera motions. We conduct experiments on various benchmarks to demonstrate the efficiency and effectiveness of our method. Compared to the state-of-the-art methods, our model achieves better image quality with less time consumption and higher rendering speed.

---

### Atompack: A Storage and Distribution Layer for Read-Heavy Atomistic ML Training Datasets

Authors: Ali Ramlaoui, Daniel T. Speckhard, Sagar Pal, Fragkiskos D. Malliaros, Alexandre Duval, Victor Schmidt
Published: 2026-06-29
Categories: cs.LG, cond-mat.mtrl-sci
arXiv: https://arxiv.org/abs/2606.29975v1

Atomistic machine learning datasets are increasingly used for training: large immutable snapshots are read repeatedly, shuffled across epochs, staged across clusters' storage systems, and republished as reusable scientific artifacts. This workload differs from interactive scientific curation, where mutable records and ad hoc inspection are often more important than random indexed throughput. We present Atompack, an append-oriented storage format and distribution layer designed around a simple workload: training pipelines usually consume complete molecular records, while the order of records is randomized by the learning algorithm. Atompack appends records efficiently during dataset construction, then commits an immutable index and serves records through a memory-mapped read path optimized for training. We compare Atompack with HDF5, LMDB, and ASE baselines representing array stores, key-value records, serialized records, and object-oriented databases. The benchmarks measure sequential reads, shuffled reads, shared-filesystem behavior, write throughput, and artifact size. On a representative 64-atom workload, Atompack is 96x faster than ASE LMDB on shuffled training-style reads while producing artifacts about 79\% smaller. The results indicate that serving complete molecule records, rather than field chunks or reconstructed objects, improves shuffled training throughput while keeping artifacts compact enough for public distribution.

---

### First-Order Temporal Logic Tensor Networks

Authors: Luca Boscarato, Ivan Donadello, Alessandro Artale, Marco Montali, Fabrizio Maria Maggi
Published: 2026-06-29
Categories: cs.AI, cs.LG, cs.LO
arXiv: https://arxiv.org/abs/2606.29972v1

Most of the existing neuro-symbolic AI methods focus on the scenario of static knowledge where objects do not change according to a temporal dimension. Temporal neuro-symbolic works are still under explored and are mainly developed for time-interval logic or propositional linear temporal logic. There is a lack of models studying linear temporal logics with predicates that deal with objects whose properties and relations change through the time. We present First-Order Temporal Logic Tensor Networks (FOT-LTN) that is an extension of Logic Tensor Networks (LTN) that fills this gap by considering a linear-temporal dimension. In particular, FOT-LTN joins the syntax of First-Order Linear Temporal Logic with the fuzzy (and real-valued) semantics of LTN obtaining a framework that supports both temporal operators and quantifiers and is totally differentiable. A first evaluation regards a temporal knowledge graph completion task on two synthetic datasets showing better performance of FOT-LTN with respect to dedicated (purely neural) methods.

---

### NeuReasoner: Theory-grounded Mapping of Reasoning Elicitation Boundaries

Authors: Aydin Javadov, Shyngys Aitkazinov, Tobias Hoesli, Florian von Wangenheim, Bjoern Schuller, Joseph Ollier
Published: 2026-06-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.29971v2

A growing body of work suggests that the reasoning capabilities of large language models are largely latent in their base form, with post-training primarily amplifying rather than introducing them. However, this evidence comes mainly from mathematical and coding benchmarks, leaving the boundary conditions of that claim largely unexplored, namely which cognitive tasks can be recovered through elicitation and where that recovery fails. To investigate this, we introduce NeuReasoner, a theory-grounded elicitation instrument. At each step, an orchestrator pairs a Neuro Lens, inspired by functional specificity, with a Cognitive Lens, drawn from the Erotetic Theory of Reasoning, and integrates their outputs through internal modularization of a single model, without external tools. We evaluate NeuReasoner on CogBench, a suite of behavioral tasks from cognitive psychology, alongside standard mathematical and coding benchmarks, measuring both its improvement over vanilla inference and its ability to match a model's post-trained thinking mode. At sufficient scale, NeuReasoner matches or exceeds thinking-mode baselines on arithmetic reasoning, code generation, Bayesian reasoning, and reward learning; these gains persist against self-consistency and iterative-refinement baselines matched to NeuReasoner's per-decision call budget. Using NeuReasoner allows us to find clear boundaries: risk-taking and decision making under uncertainty remains hard to recover through elicitation alone, and model scale interacts with elicitation in both directions: widening its advantage on some cognitive signatures while erasing it on others. Overall, through NeuReasoner as a modular, interpretable, theory-grounded elicitation instrument, we empirically map where reasoning elicitation succeeds and fails, beyond the mathematical and coding benchmarks where prior claims have rested.

---

### NeuReasoner: Theory-grounded Mapping of Reasoning Elicitation Boundaries

Authors: Aydin Javadov, Shyngys Aitkazinov, Tobias Hoesli, Florian von Wangenheim, Bjoern Schuller, Joseph Ollier
Published: 2026-06-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.29971v1

A growing body of work suggests that the reasoning capabilities of large language models are largely latent in their base form, with post-training primarily amplifying rather than introducing them. However, this evidence comes mainly from mathematical and coding benchmarks, leaving the boundary conditions of that claim largely unexplored, namely which cognitive tasks can be recovered through elicitation and where that recovery fails. To investigate this, we introduce NeuReasoner, a theory-grounded elicitation instrument. At each step, an orchestrator pairs a Neuro Lens, inspired by functional specificity, with a Cognitive Lens, drawn from the Erotetic Theory of Reasoning, and integrates their outputs through internal modularization of a single model, without external tools. We evaluate NeuReasoner on CogBench, a suite of behavioral tasks from cognitive psychology, alongside standard mathematical and coding benchmarks, measuring both its improvement over vanilla inference and its ability to match a model's post-trained thinking mode. At sufficient scale, NeuReasoner matches or exceeds thinking-mode baselines on arithmetic reasoning, code generation, Bayesian reasoning, and reward learning; these gains persist against self-consistency and iterative-refinement baselines matched to NeuReasoner's per-decision call budget. Using NeuReasoner allows us to find clear boundaries: risk-taking and decision making under uncertainty remains hard to recover through elicitation alone, and model scale interacts with elicitation in both directions: widening its advantage on some cognitive signatures while erasing it on others. Overall, through NeuReasoner as a modular, interpretable, theory-grounded elicitation instrument, we empirically map where reasoning elicitation succeeds and fails, beyond the mathematical and coding benchmarks where prior claims have rested.

---

### RiverONE: Generating Knowledge-Intensive VLM by Simulated Quantum Machines

Authors: Xindian Ma, Xinyu Long, Yefei Zhang, Yanchen Liu, Xianghao Li, Yufu Wen, Yike Hu, Yuedong Zhu, Zeyang Ma, Wen Qin, Yikun Wang, Peng Yang et al.
Published: 2026-06-29
Categories: quant-ph, cs.AI
arXiv: https://arxiv.org/abs/2606.29966v1

Quantum computing provides a powerful paradigm for representing and transforming high-dimensional information through superposition, entanglement, and measurement-induced nonlinear features. While current quantum hardware is not yet practical for direct large-scale vision-language model (VLM) inference, simulated quantum computation can be used during model construction to generate structured parameters for compact classical AI systems. We build RiverONE, a lightweight vision-language model for quantum calibration plot understanding, using simulated quantum computation. It employs a specialized visual encoder and an InternVL-based language backbone. To compensate for compression-induced information loss, we introduce quantum-generated parameters, which are materialized as classical tensors after training. This allows RiverONE to run entirely on classical GPUs at inference time, with no quantum hardware or runtime quantum simulation. With approximately 1.9 billion parameters, RiverONE achieves at least 95\% of the performance of NVIDIA Ising Calibration 1 on quantum calibration plot understanding tasks while using less than 10\% of its parameter count. These results suggest that simulated quantum computation can serve as a practical construction-stage mechanism for building lightweight, knowledge-intensive scientific VLMs. Our code is available at https://github.com/THeWakeSystems/RiverOne.

---