# Research Papers — 2026-09 · bundle page 150 of 266 > Full-text markdown of papers #7451–7500 of 13298 published in 2026-09, newest first. AI-readable mirrors of the full 13298-paper 2026-09 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-09/page/150 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/149.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/151.md) ### Learning Where to Focus: Self-Supervised Multi-Scale ViTs for Histopathology Authors: Anabel Stammer, Valay Bundele, Mehran Hosseinzadeh, Hendrik P. A. Lensch Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18578v1 Pathologists diagnose diseases by first locating suspicious tissue and then examining it at higher magnification, whereas self-supervised vision transformers (ViTs) allocate the same spatial resolution to every image region despite diagnostic evidence being sparse and spanning multiple biological scales. Recent pathology foundation models have substantially improved representation quality by scaling training data and model capacity, but largely retain uniform tokenization. We instead investigate whether pathology representations can be improved by learning where to allocate spatial resolution during self-supervised learning. To this end, we propose CRAFT (Coarse-to-fine Region-Adaptive Feature Tokenization), a DINO-based framework that learns image-dependent mixed-scale representations by using self-supervised attention to selectively refine informative regions while preserving coarse context, together with a symmetric cross-scale regularization objective that encourages complementary coarse and fine representations. Across CAMELYON16, TCGA-Lung subtype classification, and TCGA-LUAD survival prediction, CRAFT consistently outperforms comparable-scale self-supervised methods while requiring lower inference computation. Despite using only a compact 22M parameter backbone trained on comparatively small pathology datasets, CRAFT remains competitive with, and often surpasses, substantially larger pathology foundation models. --- ### Accurate Trace Estimation with Fewer Random Bits via Recursive TensorSketch Authors: Mohammad Azhar Khan, Rameshwar Pratap, Amit Sharma Published: 2026-09-16 Categories: cs.LG, cs.DS arXiv: https://arxiv.org/abs/2609.18577v1 We consider the problem of estimating the trace of an implicit matrix $\mathbf{A} \in \mathbb{R}^{d^p\times d^p}$ that can only be accessed through matrix-vector products queries. The \textit{Hutchinson trace estimator}% ~\cite{Girard1987algorithme, article-hutchinson} is a classical sketching method for this problem. Their estimator, $H_{m}(\mathbf{A}) = \frac{1}{m} \sum_{i=1}^{m} {\mathbf{z}^{(i)}}^T \mathbf{A} \mathbf{z}^{(i)}, \quad \text{where } \ {\mathbf{z}^{(i)}}\in \mathbb{R}^{d^p}$, and $z^{(i)}_j \in {N}(0, 1), j\in (d^p)$, satisfies the following guarantees: (i) $\mathbb{E}(H_{m}(\mathbf{A}))=\operatorname{tr}(\mathbf{A})$, and (ii) $\mathrm{Var}(H_{m}(\mathbf{A}))=\frac{2}{m}||\mathbf{A}||_F^2$. Generating one query vector $\mathbf{z}^{(i)}$ requires $O(d^p)$ random bits; thus, $m$ queries require $O(md^p)$ random bits, which can be prohibitive in large-scale applications. Recent work by Meyer et al.~\cite{meyer2025hutchinsonsestimatorbadkroneckertraceestimation} proposes a variant of the Hutchinson trace estimator in which each query vector in $\mathbb{R}^{d^p}$ is constructed as the Kronecker product of $p$ random vectors in $\mathbb{R}^d$, requiring $O(mpd)$ random bits for $m$ query vectors. The estimator of~\cite{meyer2025hutchinsonsestimatorbadkroneckertraceestimation} is unbiased; however, its variance grows exponentially with $p$. In this work, we address this limitation by proposing a sketching-based estimator that requires $O\!\big(p (d + m)\log m\big)$ random bits, yields an unbiased estimate of the trace, and simultaneously achieves a variance bound that grows polynomially with $p$. --- ### Deep learning emergent spacetime from fermionic spectral functions in holography Authors: Koji Hashimoto, Hyun-Sik Jeong, Keun-Young Kim, Daichi Takeda, Kwan Yun Published: 2026-09-16 Categories: hep-th, cond-mat.str-el, cs.LG arXiv: https://arxiv.org/abs/2609.18566v1 We present a physics-informed machine learning framework based on Neural Ordinary Differential Equations that solves the holographic inverse problem: reconstructing the bulk spacetime and gauge field of a charged AdS black hole directly from boundary fermionic spectral functions. Encoding the UV asymptotics, horizon regularity, and zero temperature extremality as hard constraints in the neural network architecture, our framework reliably reconstructs the extremal Reissner-Nordström AdS geometry across three quantum critical regimes set by the $U(1)$ probe charge---non-Fermi liquid, marginal Fermi liquid (strange metal), and Fermi-liquid-like states---and can jointly infer the probe charge itself to sub-percent accuracy. Relaxing the near-AdS boundary constraint uncovers a geometrical degeneracy: bulk profiles that differ throughout the radial direction but share the same near-horizon $AdS_2 \times \mathbb{R}^2$ data reproduce identical spectral functions near the Fermi surface. This isospectral non-uniqueness is precisely the bulk degeneracy expected on general holographic grounds at zero temperature, and its spontaneous emergence across independent training runs shows that the network isolates the IR CFT universality rather than overfitting a single UV completion. --- ### Variational Quantum Transformer Architecture for Synthetic Language Generation Authors: Julian Hager, Michael Kölle, Gerhard Stenzel, Tobias Rohe, Jonas Stein, Claudia Linnhoff-Popien Published: 2026-09-16 Categories: quant-ph, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.18565v1 We propose a compact NISQ-compatible quantum transformer architecture for synthetic QNLP sequence modelling. The model preserves the autoregressive next-token interface of a classical transformer, but replaces attention and feed-forward sublayers with variational quantum encoder blocks, connector circuits, decoder blocks and a direct two-qubit measurement readout. Token contexts are angle-encoded into small quantum registers, processed by parallel variational heads and encoder integration circuits and conditioned through decoder ancillae to produce a distribution over a four-token vocabulary. We evaluate several architecture variants on deterministic and lexicographic grammar-generation tasks against a compact classical transformer baseline. The quantum models are trainable end-to-end and learn nontrivial grammar structure, including perfect deterministic generation in individual runs and high lexicographic validity in the strongest variant. The classical baseline remains more accurate and stable and the quantum models are sensitive to initialization. The contribution is therefore not a claim of quantum advantage, but a concrete architecture and evaluation of transformer-inspired QNLP sequence modelling under near-term quantum constraints. --- ### Sim-to-Real Traffic Scene Understanding by Decoupling Semantics from Caption Generation with V-JEPA Authors: Nguyen Hoai Thuong Bui, Thanh Nguyen Vo, Trinh Tra Giang Nguyen, Ha Duc Bui Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18562v1 Track 2 of the AI City Challenge 2026 requires both visual question answering (VQA) and traffic event description generation under a challenging synthetic-to real domain shift. Existing vision-language approaches often entangle semantic understanding with language generation, making them susceptible to hallucination and inconsistent reasoning across event phases. In this work, we propose a decoupled semantic understanding framework that first resolves predefined traffic questions into structured semantic facts and subsequently uses these facts to guide caption generation. A frozen V-JEPA encoder extracts predictive scene representations, while a lightweight Llama-based predictor produces answers for VQA queries. To improve reliability, we introduce a training-free structured refinement mechanism that exploits statistical priors, inter-question relationships, and temporal event consistency to correct prediction errors. The refined semantic facts are then provided to Qwen3-VL-8B to generate pedestrian and vehicle descriptions for each traffic event. Experimental results on the official 2026 AI City Challenge Track 2 benchmark show that the proposed method achieves 87.09% VQA accuracy and an overall S2 score of 60.0853, ranking first among all participating teams. These results demonstrate that predictive world representations combined with structured semantic refinement enable more accurate and reliable traffic understanding, leading to higher-quality lan guage generation. --- ### The evolution of sex for artificial intelligence: a population-genetic framework for multigenerational model populations Authors: Giorgio F. Gilestro Published: 2026-09-16 Categories: cs.LG, cs.NE, q-bio.PE arXiv: https://arxiv.org/abs/2609.18560v1 Some aspects of AI development resemble a population process in which models are specialised, retrained on the output of peers, or combined by averaging weights. These practices lead to generations of models, in the biological sense studied by population genetics. Here, I develop this parallelism and interpret multigenerational model populations in terms of sexual and asexual reproduction, formally recombining the two fields. I test these analogies in an exact inheritance model, in trained networks (recurrent, feedforward and variational autoencoder generators) and in large language models, and show that they hold generally, with some measurable architecture-specific biases. Training recursively on model output is known to lead to model collapse, a process previously described as akin to genetic drift; I develop all that follows. A minimal model of a learner retrained on its parent's output reproduces the Wright-Fisher process exactly; verified real data added to each generation play the role of immigration, with the surprising finding that the absolute number of real data samples matters, not their share, exactly as in population genetics. Training a child on the average of its parents' outputs cancels the benefit of having several parents, matching blending inheritance (and reviving Jenkin's objection to Darwin), whereas combining parents so that each keeps its strongest contribution preserves it; merged language-model specialists exceeded every parent across seeds (the Fisher-Muller effect); and lineages become reproductively isolated, losing the ability to merge at all, when they have learned conflicting conventions and not when they have merely drifted apart. As AI societies become societies in time as well as in space, a mathematical framework for their inheritance acquires predictive power. Remarkably, that framework can be adapted almost wholesale from biology. --- ### Interpretable Patch-Based Deep Learning for Wildfire Spread Prediction from Ensemble Simulations Authors: Marcin Lawenda, Aleksandra Krasicka, David Caballero, Luis Torres, Łukasz Szustak Published: 2026-09-16 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.18555v1 Wildfire spread is traditionally predicted using physics-based simulators, which are physically interpretable but whose cost increases with each additional ensemble member. We ask how well deep learning surrogates can reproduce these simulations at a fraction of this cost, training them on 10,584 fire spread simulations at 2m resolution for the Rectoret region in Catalonia, Spain. Four architectures are compared: a patch-based U-Net, a transfer-learned ResNet-50, a physics-informed network constrained by the wind-driven advection equation and a Swin-Unet transformer. Among the terrain and vegetation variables, only surface fuel load predicts burn probability with any strength (r = 0.27) and including it lowers prediction error by 21%. The remaining variables correlate weakly and are highly duplicative. Next, an experiment with saliency, occlusion and rotation demonstrates the models' learning. Convolutional models rely primarily on distance from the current fire front, while Swin-Unet assigns more weight to fuel and terrain, a finding also noted in an unrelated wildfire dataset. When applied without retraining to the second region, Pedriza, all three convolutional models still predict fire spread, losing accuracy by a small but systematic margin. --- ### CARA: Collision-Aware Resolution Adaptation for Multiresolution Hash Encoding Based Image Fitting Authors: Linfeng Ye, Zhixiang Chi, Shayan Mohajer Hamidi, En-hui Yang, Konstantinos N. Plataniotis Published: 2026-09-16 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2609.18554v1 Multiresolution hash encodings have recently enabled fast and high-fidelity implicit neural representations by storing multi-scale features in fixed-size hash tables along a geometric resolution schedule. However, the standard design is data-agnostic: different resolution levels receive identical hash-table capacity despite large differences in image frequency content. As a result, some levels experience severe hash collisions while others underutilize parameters, leading to inefficient capacity allocation. To address this issue, we propose Collision-Aware Resolution Adaptation (CARA), a method that assigns per-level resolutions by balancing the effective information load across hash levels. This adaptive allocation reduces capacity bottlenecks and improves parameter efficiency. In addition, we introduce an invertible pixel-shuffle transform that reduces hash load factors by redistributing spatial information, thereby mitigating collision-induced information loss without enlarging the hash tables. To support evaluation on extremely high-resolution data, we also curate, to the best of our knowledge, the first uncompressed whole-slide image dataset for academic research. Experiments on Kodak images, gigapixel natural images, and raw whole-slide images demonstrate that CARA consistently improves the fidelity-parameter trade-off. Our method matches state-of-the-art performance while using only $27.76%$ of the parameters, and achieves up to $6.11$ dB PSNR improvement at comparable parameter counts. Code is provided in the supplementary. --- ### HAP: A Hand-Driven Active Perception Framework for Egocentric Head Motion Prediction Authors: Yunji Feng, Junyi Ma, Guanzhong Sun, Chenyang Xu, Hesheng Wang Published: 2026-09-16 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2609.18548v1 Egocentric motion forecasting has primarily focused on hands and manipulated objects, leaving future human head motion comparatively underexplored. During manipulation, the head both redirects perception toward the target to acquire task-relevant evidence and coordinates with body and hand motion. We therefore formulate future six Degree of Freedom (6-DoF) head-motion prediction conditioned on observed hand motion and inferred target context, and propose HAP, a Hand-Driven Active Perception framework. HAP infers confidence for each target object from observed hand motion and object geometry. Then constructs a dynamic Predictive Target-Centric Amodal Occlusion Graph (P-TAOG) representing current and potential occlusion among candidate objects. Directed graph and causal temporal reasoning encode the evolving target conditioned perceptual state, which is fused with hand and head motion history. A horizon-wise gate then blends the learned trajectory with a constant velocity prior. We further introduce Bottle, an egocentric RGB-D dataset of object manipulation toward specified targets, with coordinated head and hand motion under changing target visibility. Experiments on the public dataset and Bottle show that HAP achieves lower head motion prediction errors than representative baselines, supporting the value of hand driven intention and dynamic occlusion reasoning for anticipating human head motion. Code will be released at https://HAP-ego.github.io/HAP. --- ### STUNet-Fusion: Spatiotemporal Needle-Tip Localization in Ultrasound Video via Multi-Channel Motion Fusion Authors: Chia-Chi Hsu, Chia-Hsuan Hsu, Che-Chou Shen Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18546v1 Needle-tip localization in ultrasound remains challenging because the needle may appear weak, discontinuous, or partially invisible, while imaging artifacts and anatomical structures can produce similar responses. To address this problem, we propose STUNet-Fusion, a spatiotemporal framework for needle-tip localization in ultrasound videos. The proposed method formulates the input as a tri-channel spatio-temporal fusion tensor, comprising grayscale appearance, grid-based motion feature, and raw frame difference. A shared ResNet-34 encoder extracts spatial features, ConvLSTM integrates temporal dependencies, and a U-Net decoder reconstructs a dense probability heatmap. The final coordinates are extracted via a soft-argmax operation to achieve sub-pixel localization accuracy. Experimental results demonstrate that this spatiotemporal fusion strategy significantly improves localization robustness compared to conventional baselines. --- ### Revisiting the Objective of Echo Chamber Detection Authors: Abylaikhan Bexeit, Kushani Perera, Shanika Karunasekera, Jean Honorio Published: 2026-09-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.18545v1 In this paper, we study the detection of an echo chamber in a social network, i.e., the identification of a set of nodes that agree on a topic, while disagreeing with the rest of nodes. We argue that this problem is different from other social network analysis problems such as community detection, and from other graph problems such as maximum graph cut and maximum clique. To the best of our knowledge, we are the first to formalize the objective function of echo chamber detection, by using the theory of Fourier transforms of set functions (Stobbe and Krause, 2012). We propose scalable semidefinite relaxation, solved via an interior point method and sparse linear algebra. Experimentally, our algorithm recovers the ground truth echo chamber better than competing methods on small synthetic experiments. Our algorithm produces echo chambers with better network properties than competing methods on large real-world datasets. To independently validate our proposed objective function, we show that our algorithm finds echo chambers with more agreements with suspended users than competing methods on a small real-world dataset. --- ### Accuracy- and Real-Time-Aware 4D Radar Preprocessing for Autonomous Driving Perception Systems Authors: Woo-Jin Jung, Dong-Hee Paek, Jeong-Su Park, Seung-Hyun Kong Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18542v1 4D radar has emerged as a promising next-generation sensor for improving the robustness of autonomous driving perception systems because of its stable sensing capability under adverse weather conditions. However, deploying 4D radar in embedded environments with limited hardware resources requires radar-representation preprocessing that jointly considers perception accuracy, real-time performance, and computational complexity. This paper proposes a preprocessing framework for 4D-radar-based 3D object detection. First, Percentile-based 3D Shape Preservation (P3DP) extracts point clouds from radar tensors while preserving object-shape information and suppressing noise and false alarms. Second, Multi-frame-based Noise Point Discrimination using Kernel Density Estimation (MF-KDE) improves the density and reliability of sparse radar point clouds. Finally, Embedded \& NetScore (ENS) evaluates suitability for embedded deployment by jointly considering accuracy, real-time performance, adverse-weather robustness, and model complexity. --- ### SVMemAgent: A Streaming Video Memory Agent for Query-Agnostic Online Frame Selection Authors: Dohwan Ko, Ji Soo Lee, Pierce Chuang, Debojeet Chatterjee, Ashish Shenoy, Yichao Lu, Seungwhan Moon, Xin Luna Dong, Vikas Bhardwaj, Hyunwoo J. Kim Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18540v1 Most keyframe selection studies focus on offline settings, assuming access to the full video and query in advance. In contrast, real-world streaming scenarios require online frame selection under unknown video duration, without access to either the query or future frames during selection. To address this, we introduce Streaming Video Memory (SVMem), a compact and representative memory of previously observed content, updated continuously as the video stream unfolds. Building on this setting, we propose the Streaming Video Memory Agent (SVMemAgent), which dynamically maintains a memory by deciding at each timestep whether to replace an existing memory frame with the incoming frame or discard it. SVMemAgent is trained using Group Relative Policy Optimization (GRPO) with task-driven rewards derived from diverse question-answer pairs, implicitly exposing the policy to a distribution of queries during training so that SVMem retains generally informative frames at inference, when queries are unavailable. Experiments on both online and offline video benchmarks show that SVMemAgent consistently outperforms online frame selection baselines and achieves competitive performance with offline methods that assume access to the full video and query. Through task-driven rewards, SVMemAgent learns an emergent keyframe selection policy that prefers frames containing textual information, which may benefit downstream VideoQA tasks. --- ### Provable Guarantees and Efficient Learning of Structural Equation Models with Latent Confounders Authors: Weijian Yu, Jean Honorio Published: 2026-09-16 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.18535v1 Causal discovery aims to recover causal relationships from observed data. In various fields, exploring causal relationships among variables remains an important topic, but this task becomes challenging due to the existence of latent confounders. Ignoring such confounders can lead to false associations and incorrect edge directions. In this paper, we study the linear structural equation model with latent confounders. We propose an algorithm that iteratively identifies terminal (observed) nodes and reconstructs the directed acyclic graph of the observed variables. To do this, we recover the precision matrix of the observed variables as a sparse plus low-rank matrix: a sparse matrix captures the conditional dependencies among observed variables, while a low-rank matrix captures the combined influence of a few latent confounders. We establish that for $p$ observed variables, $r$ latent confounders and $s$ edges, our procedure correctly identifies the directed causal relationship among observed variables, for $n \gtrsim \max\{s\log p,\ r p\}$ samples. Experimental results validate our theoretical contributions. --- ### A Probe Shift Is Not a Fairness Fix: The Limits of Representation Steering in Speech Models Authors: Nicolas Bourrel, Abderrahmane Issam, Gerasimos Spanakis Published: 2026-09-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.18533v1 Automatic speech recognition (ASR) systems exhibit unequal error rates across speaker groups, motivating interventions on their internal representations. We ask whether speaker-linked attributes that are linearly readable from pretrained ASR encoders yield useful directions for reducing group word-error-rate (WER) gaps. Across Whisper-medium, HuBERT-large, and Wav2Vec2-large on Common Voice and the Speech Accent Archive, we probe every encoder layer for metadata-derived sex/gender, age, and native/accent labels; construct centroid and probe-derived directions; inject them at selected layers; and compare downstream probe trajectories with matched WER changes. Sex labels are highly decodable (best macro-F1 0.924--0.941), native/accent labels are also above chance (0.544--0.696), and age is weaker (0.354--0.397). Of 22 post-selected reruns, nine have 95% paired-bootstrap intervals entirely below zero, yet every absolute source-group WER reduction is below 0.7 percentage points. Conversely, a local target-class probe rate can rise from 8.09% to 99.87% while WER worsens. Linear readability is therefore neither evidence of causal use nor a reliable mitigation method. Our results motivate evaluating speech-bias interventions jointly at representation, propagation, and task levels. --- ### Machine Translation between English and Syriac (East Syriac Dialect) using Statistical Machine Learning Authors: Hadiana Sliwa, Hossein Hassani Published: 2026-09-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.18529v1 UNESCO considers the Assyrian (Syriac) language an endangered language. Although Assyrians speak the language worldwide, the speaking population is uncertain (ranging from 500,000 to 1,500,000). Syriac is also one of the least studied languages in Natural Language Processing (NLP). Despite advances in Machine Translation (MT) over the past decade, the lack of publicly available corpora and the orthographic complexity of the Syriac script, specifically the Madnkhaya script, have left this language entirely ignored in the computational linguistics literature. This study develops the first phrase-based Statistical MT (SMT) model for English-to-Assyrian MT using the Moses framework. We created a dataset of 38,847 sentence pairs from the complete English and Syriac Bible, merging a pre-existing New Testament dataset with an Old Testament built from scratch through PDF extraction, using custom segmentation scripts and manual alignment review by three bilingual annotators. The Syriac side of the corpus undergoes diacritic removal and Byte-Pair Encoding tokenization to reduce orthographic sparsity before training. We trained and evaluated six models using different configurations and splitting-scheme ratios, language model order, distortion limits, and the inclusion of an Operation Sequence Model. The best-performing configuration achieves a word-level BLEU score of 23.54. Human evaluation by 11 native Assyrian speakers resulted in mean adequacy and fluency scores of 3.42 and 3.34 out of 5, respectively. These results are consistent with comparable low-resource SMT models trained on Biblical corpora for morphologically rich Semitic languages. The corpora, scripts, and trained model are publicly available, providing the research community with the first systematically curated English-Syriac dataset and a reproducible baseline for future MT and broader NLP work on this endangered language. --- ### Provable Guarantees for Spectral Structured Prediction Authors: Violet Zheng, Jean Honorio Published: 2026-09-16 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.18527v1 Structured prediction is the simultaneous prediction of multiple labels, and is widely used in various fields, such as natural language processing and computer vision. In this paper, we study binary node label recovery on signed graphs with edge-flip noise, a model introduced by (Globerson et al., 2015), via a simple spectral method that decodes node labels from the signs of the principal eigenvector of the noisy signed adjacency matrix. We develop graph structure-agnostic theoretical guarantees for approximate inference of node labels as well as guarantees for maximum angle deviation with respect to the ground truth node labels. By leveraging tools from matrix concentration theory and eigenvector perturbation analysis, we derive new concentration inequalities that explicitly quantify the effect of the spectral gap of the adjacency matrix, number of nodes, degree distribution, and noise level. As a corollary, we relate our general results to the Cheeger constant and provide results for different classes of graphs. We perform several synthetic experiments to validate our theory. To the best of our knowledge, we are the first to provide theoretical guarantees for the spectral-based approach. As a byproduct of our analysis, we derive technical results that might be of independent interest and useful for other machine learning problems. --- ### TRIPROBE: Probing Task Separability Beyond Classification for XAI Authors: Amirhossein Sadough, Freek Hens, Aleksa Bokšan, Mohammad Mahdi Dehshibi, Mahyar Shahsavari Published: 2026-09-16 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.18525v1 Modern evaluation of learning pipelines often reduces to downstream accuracy, leaving open the question of why tasks succeed or fail. TriProbe addresses this gap with a multi-level probing framework for explainable diagnosis of task separability. Rather than treating models as black boxes, TriProbe traces how separability evolves across inputs, learned features, and final classifiers. It decomposes multi-task problems into binary subtasks and applies three complementary probes: a Foundational Probe on input spaces, a Latent Probe on feature representations, and a Final Probe on classifier outputs. Using Maximum Fisher's Discriminant Ratio as a principled separability metric, TriProbe identifies bottlenecks and affected task pairs. Experiments on the Roshambo sEMG benchmark show how TriProbe reveals hidden breakdowns, guiding data collection, validation, and architecture design. --- ### VoiceTrace: A Benchmark and Retrieval Framework for Who-Said-What Speech Retrieval Authors: Aaron Yee, Fengjie Lu, Jiarui Hai, Chenang Jiang, Helin Wang, Siwei Tu, Weitao You, Lingyun Sun Published: 2026-09-16 Categories: cs.SD, cs.AI, eess.AS arXiv: https://arxiv.org/abs/2609.18521v1 Speech retrieval has become increasingly important as spoken content continues to grow across meetings, lectures, podcasts, and videos. Existing benchmarks and models have advanced semantic search over spoken content, but largely focus on \emph{what} is said while overlooking \emph{who} says it. In many real-world scenarios, however, users need to retrieve speech based jointly on semantic content and a target speaker, where the speaker may be specified naturally through a reference speech utterance rather than a predefined identity. To address this gap, we introduce \textbf{VoiceTrace-Bench}, a benchmark for hybrid speech retrieval in which each query combines text specifying \emph{what} to retrieve with reference speech specifying \emph{who} to retrieve. This setting requires models to integrate complementary semantic and speaker information directly from heterogeneous query inputs. Motivated by the joint audio-text modeling capabilities of audio-language models (ALMs), we develop \textbf{VoiceTrace}, a two-stage retrieval framework consisting of \textbf{VoiceTrace-Emb}, an embedding model that learns unified representations for efficient large-scale retrieval, and \textbf{VoiceTrace-Reranker}, a reranking model that jointly examines each query--candidate pair for fine-grained relevance estimation. Experiments show that VoiceTrace achieves state-of-the-art performance on established semantic speech retrieval benchmarks, while substantially outperforming cascade-based approaches on VoiceTrace-Bench, demonstrating its effectiveness for both conventional semantic retrieval and the new hybrid retrieval setting. --- ### AeroWeaver: An Embodied-Agent Harness for Weaving Aerial Skills into Distributed, Adaptive Swarm Execution Authors: Jiabin Lou, Yirong Yang, Haopeng Wang, Xuxin Lv, Xinyu Liu, Diyuan Hou, Xuehong Liu, Rongye Shi, Wenjun Wu Published: 2026-09-16 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.18520v1 Collective intelligence is a collaborative autonomy paradigm in which multiple agents pursue shared objectives through local perception, information exchange, and coordinated action. UAV swarms embody this paradigm by coordinating multiple vehicles in tasks such as search, inspection, and tracking. Recent advances in large language model (LLM) agents have strengthened natural-language task understanding and high-level planning, providing a flexible semantic interface between mission descriptions and collective behavior. While these advances expand semantic reasoning, applying LLM agents to UAV swarms raises challenges in grounding model decisions in executable capabilities, reconciling global task reasoning with distributed execution, and using mission-specific experience for continual adaptation. To address these challenges, we introduce AeroWeaver, an embodied-agent harness that weaves individual UAV skills into coordinated mission-level behavior. AeroWeaver connects semantic decisions to governed skills, organizes role-conditioned local agents for distributed coordination, and uses role-indexed state-action-reward experience to refine skill selection online. Experiments and runtime validation show that AeroWeaver maintains valid skill execution under tested conditions and supports body-local multi-UAV operation without a central agent generating joint actions from global context, while reward-guided online updates provide a training-free path for adaptive learning swarm agents from accumulated execution experience. Code: https://github.com/Admire-ljb/AeroWeaver. --- ### COMPASS-ABS: Reducing Fragmentation in Shared GPU Clusters for Deep Learning Training Workloads Authors: Yukai Zhou, Hongfan Wu Published: 2026-09-16 Categories: cs.DC, cs.LG arXiv: https://arxiv.org/abs/2609.18519v1 With the rapid advancement of deep learning technology, shared GPU clusters receive an increasing number of deep learning training (DLT) jobs. Yet resource fragmentation make such clusters underutilized and forces the DLT jobs running on them to endure long turnaround times. Extensive research has been devoted to quantifying fragmentation and developing scheduling algorithms that alleviate its impact. However, existing fragmentation measures break down in the absence of workload distribution information, while current schedulers cannot continuously maintain resource fragmentation at a low level. To tackle these problems, we first introduce Scheduler-Induced Fragmentation (SIF), a metric built on the notion of partial-nodes that is independent of historical workload knowledge. We then propose COMPASS-ABS, which employs the COMPact-ASSured (COMPASS) algorithm to confine the cluster state within a tight Anchor-Based Space (ABS), whose construction fully leverages the topological alignment between dominant workload size and node capacity. Moreover. We also prove that it ensures SIF is bounded by $\frac{2}{N}$ under a workload composition condition that matches both theory and production. Evaluations implemented on a physical cluster and a simulated cluster demonstrate COMPASS-ABS effectiveness at improving resource utilization, reducing DLT job completion time by reducing fragmentation. --- ### Align, Integrate, and Fire: Efficient Token-Level Alignment for Zero-Shot SpeechLLMs Authors: Abderrahmane Issam, Yusuf Can Semerci, Jan Scholtes, Gerasimos Spanakis Published: 2026-09-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.18516v1 While Large Language Models excel in natural language processing, efficiently extending their capabilities to spoken input remains a significant challenge. Existing methods for building SpeechLLMs often rely on computationally expensive full-model fine-tuning, or employ parameter-efficient projectors that suffer from inefficient token sequence lengths and costly full-model supervision. In this paper, we introduce Aligned Continuous Integrate-and-Fire, a highly efficient framework for zero-shot speech processing. Our method dynamically compresses continuous acoustic frames into the exact discrete token length of the target text utilizing explicit Dynamic Time Warping alignments. This allows our initial training stage to establish a robust acoustic-to-semantic bridge using lightweight distance metrics, entirely bypassing the computationally expensive LLM forward pass. For subsequent fine-tuning, we propose a memory-efficient knowledge distillation objective that targets a single LLM layer, performing competitively with full-model cross-entropy training at a fraction of the computational cost. Through extensive evaluations on Automatic Speech Recognition and Speech Translation, we demonstrate that our method achieves superior performance compared to prior parameter-efficient baselines. --- ### Beyond Routine Compliance: Cunning Data Cultivates Safety Vigilance in Large Language Models Authors: Youjia Wang, Lin Xu, Yang Sun, Yuxiao Lu, Chengfang Fang, Jie Shi Published: 2026-09-16 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.18515v1 Safety alignment teaches large language models (LLMs) to recognize harmful requests and reject risky instructions. Yet aligned models can fail when harmful intent is concealed within seemingly benign contexts. Robust safety therefore requires both knowledge of safety boundaries and \textbf{vigilance}: the ability to detect unusual premises, misleading reasoning, and latent risks beneath surface-level semantics. Vigilance requires models to scrutinize a request's underlying intent and assumptions before acting. To cultivate this capability, we introduce \textbf{cunning questions}, which are not necessarily safety-related but contain misleading premises, atypical reasoning, or subtle inconsistencies. We hypothesize that learning to look beyond such reasoning traps can transfer to safety-critical scenarios. Experiments show that Cunning training improves robustness to out-of-distribution jailbreak attacks and strengthens subsequent safety fine-tuning. Furthermore, augmenting an existing state-of-the-art safety alignment pipeline with Cunning establishes a new state of the art across our evaluated settings, reducing mean ASR across nine backbone--benchmark combinations from 17.40\% to 15.05\%. Trace analysis after matched safety fine-tuning suggests that safety judgments are more likely to govern responses before harmful planning begins. A conditional theoretical analysis further characterizes when invariance learned from cunning data can transfer to safety-related inputs. These findings suggest that cunning data can strengthen model vigilance and complement conventional safety alignment. --- ### ActiveScale: Scaling Active Perception for Robots across Model, Data, and Hardware Authors: Shuai Zhou, Kaisheng Pang, Wenxuan Song, Wenjie Zhang, Xinhu Zheng, Haoang Li Published: 2026-09-16 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2609.18514v2 Active perception is essential for robotic manipulation when fixed viewpoints leave task-relevant information occluded or unobserved. However, enabling vision-language-action (VLA) models to reason across changing viewpoints and actively acquire informative observations remains challenging. We present ActiveScale, a framework that advances active perception through coordinated model, data, and hardware designs. Our model augments a VLA with historical video observations and explicit camera-pose supervision, using per-frame pose tokens and a lightweight prediction head to associate observations across viewpoints and support a coherent understanding of the scene. To learn from the camera motion naturally present in human activity, we introduce a scalable human--robot mid-training recipe using 1000 hours of egocentric and robotic data, adapting the model to temporal inputs and pose supervision. We further introduce Active-perception Mobile-manipulation Platform (AMP), a robotic platform that supports active perception and mobile manipulation through single-operator teleoperation, enabling scalable collection of demonstrations that coordinate viewpoint changes and manipulation. Experiments demonstrate improved success rates on active-perception tasks, while ablation studies validate the contributions of camera-pose-aware modeling and egocentric mid-training. Together, these components provide an integrated foundation for studying and developing active perception in robotic manipulation. --- ### Learning from Distributed Eyes: Leveraging Collaborative Perception for Automated Model Adaptation Authors: Yanan Ma, Yihang Tao, Zhengru Fang, Zihan Fang, Yiqin Deng, Xianhao Chen, Yuguang Fang Published: 2026-09-16 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.18511v1 In autonomous driving, perception models often struggle to generalize to new environments due to domain shifts. While unsupervised model adaptation offers a feasible solution without labor-intensive manual labeling, existing methods that rely solely on the ego-vehicle's data often lead to inferior pseudo-labeling performance. To address this critical issue, we propose LDE, Learning from Distributed ``Eyes", a novel framework that transforms collaborative perception (CP) into a source of high-quality supervision for model adaptation. This pseudo-labeling approach is hyperparameter-insensitive and relatively reliable, assuming CP often outperforms single-agent's perception. However, naively implementing this approach encounters (1) the communication bottleneck of sharing rich features under time and bandwidth constraints, (2) the view discrepancy between the CP view and the learner's Field of View (FoV), and (3) the unreliability even in CP-generated labels. To address these issues, we design an adaptation-oriented feature sharing mechanism that selectively transmits the most critical information for adaptation, an FoV filtering method that meticulously eliminates mismatched labels, and a curriculum learning strategy to progressively exploit pseudo labels. Extensive experiments on 3D object detection tasks demonstrate that LDE consistently outperforms both the pre-trained models and state-of-the-art unsupervised adaptation methods. --- ### DiT-Garment: Garment Dynamics with Diffusion Transformers Authors: Antoine Dumoulin, Laurence Boissieux, Joao Regateiro, Pierre Hellier, Stefanie Wuhrer Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18510v1 We present DiT-Garment to model dynamic 3D clothing over human body models in arbitrary motion. Unlike existing methods, DiT-Garment can animate garments with unseen designs and physical materials, while allowing for direct inference of deformations for any target pose. To achieve this, we leverage a 2D diffusion transformer architecture to learn 3D deformations in a 2D UV-space. As the result is non-deterministic, our generative model learns the distribution of possible outcomes. The template garment is represented as a 3D triangle mesh spatially aligned with a 3D human body model in a standardized pose. To work with different garment designs without the need of a common template or complex graph convolution operations, the diffusion transformer is conditioned on a 3D position map of the template, represented in UV-space, which allows to implicitly learn a deformation of the 3D space around the body in standard pose. Further conditioning on body motion and physical parameters allows to physically ground the model. We quantitatively and qualitatively evaluate DiT-Garment on both synthetic and real data. While only trained on synthetic simulations of automatically generated cloth designs, our method generalizes to captured and artist-made garment designs. Code and data are available for research purposes at https://dumoulina.github.io/dit-garment/. --- ### MiST: Mid-Training LLMs for Cybersecurity Authors: Oded Ovadia, Elad Ben Zaken, Elad Guttman, Orly Moreno Kadosh Published: 2026-09-16 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2609.18496v1 Cybersecurity combines high-stakes analysis with complex technical language, making it an impactful and challenging domain for LLMs. We present MiST (Mid-trained Security Transformer), a suite of 8B and 32B models that achieve strong performance on public cybersecurity benchmarks. We use mid-training as an intermediate adaptation stage between general pre-training and cybersecurity training. Rather than performing continual pre-training over large volumes of raw domain text, we curate a compact, expert-vetted seed corpus, and transform it into high-quality domain-specific synthetic training data. The final MiST checkpoints improve mean cybersecurity accuracy by +13.1 and +8.6 absolute percentage points over the corresponding Qwen baselines for 8B and 32B, respectively, corresponding to relative gains of +27.0% and +15.8%. Ablation results further show that these cybersecurity gains arise in the mid-training and supervised fine-tuning stages through a combination of the synthetic data generation flows. Furthermore, we show that MiST provides a stronger initialization for downstream task-specific fine-tuning adaptation and reinforcement learning. --- ### Size Matters: Foundation Model for Czech HTML documents Authors: Martin Dvořák, Vít Tlustoš, Artyom Voronin, Martin Habrovec, Kateřina Podlesná, Barbora Rišová, Josef Vonášek Published: 2026-09-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.18494v1 Creating universal, high-quality representations of web documents in high-traffic industrial environments requires models that are both performant and economic. Existing approaches, however, often depend on large models, overlook the structural information inherent in HTML, or are constrained by short context windows, limiting their ability to process real-world web pages. We present HTML-LM, a compact foundation model with 154 million parameters that addresses these limitations through HTML-aware training and a ModernBERT-based architecture. It was trained on 100 million web documents using multiple objectives, including masked language modeling, bag-of-words prediction, and contrastive distillation from large language models. Consequently, HTML-LM sets a new state-of-the-art for classification and regression applications in the Czech Internet domain, surpassing both larger encoders and small-sized LLMs. The model is deployed in production, processing thousands of web documents per second, and released to the community under the CC BY-NC 4.0. https://huggingface.co/Seznam/html-lm. --- ### Semantic-ITC: A Frame-wise Indoor Mobile Laser Scanning Dataset and Benchmark for Semantic Segmentation Authors: Haiyang Wu, Muhammad Affan, George Vosselman, Ville Lehtola Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18493v1 Semantic labels for indoor mobile laser scanning (MLS) frames remain largely absent from current point cloud semantic segmentation benchmarks, which mainly focus on reconstructed indoor scenes or outdoor LiDAR perception. This paper introduces Semantic-ITC, to the best of our knowledge the first public dataset and benchmark for frame-wise indoor MLS semantic segmentation. The dataset contains 52 indoor sequences, 79,108 MLS frames, and 1.23 billion labeled points collected in classrooms, corridors, meeting rooms, offices, and study areas. Labels are attached directly to measured LiDAR points in each frame using 16 semantic classes covering structural elements, furniture, room equipment, vegetation, and other indoor objects. Semantic-ITC preserves the sparse, non-uniform, and frame-wise sampling pattern of indoor MLS, making it distinct from scene-level reconstructed point clouds and mesh-based indoor datasets. The annotations are produced by a hybrid workflow that combines predictions from a visual foundation model applied to synchronized RGB images, structural information from BIM, and manual refinement, with the final labels assigned to the original LiDAR frames. A single-frame benchmark is provided, and the best baseline reaches 79.27\% mIoU. Remaining errors are concentrated around object boundaries and ambiguous indoor classes, indicating the challenges of indoor MLS segmentation under sparse frame geometry and long-tailed class distributions. The dataset provides a public benchmark for evaluating semantic segmentation directly on measured indoor MLS frames and supports future studies on frame-wise indoor MLS semantic segmentation. --- ### Learning A Unified Template for Gait Recognition Authors: Panjian Huang, Saihui Hou, Junzhou Huang, Yongzhen Huang Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18490v1 "What I cannot create, I do not understand."Human wisdom reveals that creation is one of the highest forms of learning. For example, Diffusion Models have demonstrated remarkable semantic structure and memory in image generation, understanding, and restoration, which intuitively benefits representation learning. However, current gait networks rarely embrace this perspective, relying primarily on learning by contrasting gait samples under varying complex conditions, leading to semantic inconsistency and uniformity issues. To address these issues, we propose Origins with generative capabilities whose underlying philosophy is that different entities are generated from a unified template, inherently regularizing gait representations within a consistent and diverse semantic space to capture accurate gait differences. Admittedly, learning this unified template is exceedingly challenging, as it requires the comprehensiveness of the template to encompass gait representations with various conditions. Inspired by Diffusion Models, Origins diffuses the unified template into timestep templates for gait generative learning, and meanwhile transfers the unified template for gait representation learning. Especially, gait generative and representation learning serve as a unified framework for end-to-end joint training. Extensive experiments on CASIA-B, CCPG,SUSTech1K, Gait3D, GREW and CCGR-MINI demonstrate that Origins performs unified generative and representation learning, achieving superior performance. --- ### Butterfly Effect and the Kinetic Energy Cascade in Probabilistic Machine Learning Weather Prediction Models Authors: Jiakai Chen, Joel Oskarsson, Simon Driscoll, Sebastian Schemm Published: 2026-09-16 Categories: physics.ao-ph, cs.LG arXiv: https://arxiv.org/abs/2609.18489v1 This study analyses kinetic energy (KE) spectra, difference kinetic energy (DKE) spectra, and signatures of KE transfer across spatial scales in four state-of-the-art probabilistic machine learning weather prediction (MLWP) models: NeuralGCM-ENS, FourCastNet 3, AIFS-ENS, and GenCast. Results are compared with those from the physics-based numerical weather prediction model IFS-ENS. While NeuralGCM-ENS successfully reproduces the expected upscale transfer of KE, noise injection at its encoder stage underestimates mesoscale KE. Conversely, AIFS-ENS, GenCast, and FourCastNet 3 produce realistic KE spectral magnitudes but do not capture the expected upscale transfer of KE. In particular, AIFS-ENS and GenCast, which employ spatially uncorrelated stochastic perturbations, exhibit enhanced accumulation of KE at high wavenumbers. All examined models exhibit upscale error growth, reflected by the progressive shift of the DKE spectral peak toward larger wavelengths over time. However, the MLWP models struggle to reproduce the rapid initial growth of ensemble spread at small spatial scales associated with the butterfly effect. The results show that MLWP models can misrepresent the known scale transfer of kinetic energy despite producing skilful weather forecasts. --- ### Beyond Random Couplings: Contrastive Noise Alignment in Generative Flows Authors: Lennart Wittke, Vinicius Azevedo Published: 2026-09-16 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.18488v1 Diffusion and flow-matching models are typically trained by corrupting data through independently sampled Gaussian noise. While simple and scalable, this forward process induces arbitrary data-noise couplings, forcing the network to learn high-curvature transports between unrelated endpoints. Existing optimal-transport methods reduce this burden by reassigning fixed noise samples to data, but the source noise distribution itself remains passive. To address this, we introduce Contrastive Noise Alignment (CNA), a training-time method that creates dynamic, contrastive couplings by optimizing the noise representations directly. By modeling the noise batch as an interacting particle system, CNA employs a cross-modal InfoNCE objective to align noise particles with their paired data targets. To prevent spatial collapse, this alignment is regularized using an angular entropy term and a radial norm penalty. We show theoretically that this equilibrium asymptotically preserves Gaussian structures, maintaining tractability during inference. Empirically, CNA improves the alignment between noise and data, reduces flow curvature, and provides better generation quality with fewer required sampling steps. For few-step, pixel-space generation (2-4 NFEs), CNA reduces FID by over 50\% compared to standard rectified flow, and by at least 24\% against Optimal Transport baselines. --- ### ActionPiece: Rethinking Action Tokenization for Autoregressive Vision-Language-Action Models Authors: Shijie Lian, Bin Yu, Zhaolong Shen, Xiaopeng Lin, Yichao Du, Zhirui Zhang, Laurence T. Yang, Kai Chen Published: 2026-09-16 Categories: cs.RO, cs.AI, cs.CL, cs.CV arXiv: https://arxiv.org/abs/2609.18487v1 Action tokenizers play a central role in autoregressive vision-language-action (VLA) models, determining both the targets for policy training and the executable commands recovered from predicted tokens. Their fidelity is commonly evaluated using pointwise reconstruction metrics such as mean squared error (MSE), yet small individual errors do not fully characterize how faithfully action adjustments across demonstrations are preserved. After compression, similar actions may still cluster around a representative motion, while the adjustments needed for different contexts are diminished, distorted, or even reversed. We introduce physical rank consistency (PRC) to measure how well tokenization preserves local physical distance rankings after reconstruction. Evaluating decoded actions provides a common reference across token vocabularies and decoder architectures, complementing pointwise accuracy with a measure of relational fidelity. We further present ActionPiece, which preserves physical action relationships through joint supervision of representation learning and quantization. Physical rank preservation supervises near-far ordering in encoder and quantized feature distances, while quantization regularization applies the same ordering to codeword assignment distributions. Both objectives augment reconstruction, producing discrete action tokens for standard autoregressive policy learning and execution through a frozen decoder. Under the same Qwen3-VL-4B policy training setup, ActionPiece achieves 94.8% on LIBERO and 68.8% on unseen LIBERO-Plus, with additional evaluations reaching 71.9% on SimplerEnv and 51.5% across VLA-Arena L0-L2. Component ablations show that the two objectives jointly improve PRC and policy success, demonstrating the value of physical relationship supervision for action tokenization. --- ### Hyperbolic Graph Representation Learning for Differential Diagnosis on Biomedical Knowledge Graphs Authors: Pietro Miotto, Lucia Mellini, Tommaso Marzi, Cesare Alippi, Elena Casiraghi, Alberto Paccanaro, Giorgio Valentini, Mauricio Soto-Gomez Published: 2026-09-16 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.18481v1 Biomedical knowledge graphs combine ontology-derived hierarchies with transversal associations among heterogeneous entities such as phenotypes, diseases, genes, proteins, and patients. This hybrid structure raises the question of whether hyperbolic embeddings, which naturally capture tree-like organization, remain useful beyond purely hierarchical graphs. We present a preliminary study of hyperbolic graph representation learning for Mendelian-disease differential diagnosis on a patient-integrated biomedical graph. Experiments on isolated ontology subgraphs show that hyperbolic models achieve strong performance in substantially lower dimensions than Euclidean baselines. We then evaluate the models on a link-prediction task that ranks candidate diseases for each patient. Results suggest that hyperbolic embeddings can exploit biomedical hierarchical structure while supporting diagnostic reasoning over heterogeneous patient-level graphs. --- ### CADSplat: Sparse-View 3D Gaussian Splatting Aided by CAD Models for Robust, Photorealistic Digital-Twin Reconstruction Authors: Kristof Overdulve, Lode Jorissen, Nick Michiels Published: 2026-09-16 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2609.18473v1 We present CADSplat, a framework that reconstructs photorealistic, geometrically accurate digital twins from sparse ($<15$ views), wide-baseline posed images of an object by regularizing 3D Gaussian Splatting (3DGS) with an explicit CAD shape prior. Using such a prior requires finding a CAD model whose shape resembles the object depicted in the images and determining the pose of each camera relative to the object. We obtain both by matching segmented object silhouettes against silhouettes rendered from a CAD library and keeping the camera-to-object poses of the best-matching model. We then anchor 3D Gaussian primitives to the surface of the retrieved model and jointly optimize the 3DGS parameters, the camera-to-object registration, and a non-rigid deformation field to account for shape differences between the physical object and the CAD model. Across two real-world datasets, CADSplat outperforms unconstrained, few-shot, and mesh-texturing baselines and degrades gracefully to as few as 3 views. Our experiments show that most of the gain in rendering quality comes from how the splats are constrained---a fixed set of splats tied to a surface and moved by a single smooth deformation field---rather than from the CAD shape itself. The CAD model adds shape knowledge where views are scarcest, in the sparsest captures and on strongly self-occluded objects, and it places every camera in the object's own frame. This enables applications beyond novel-view synthesis, such as markerless augmented reality registration, per-image object pose estimation, physical simulations, and the transfer of part labels from the design to the reconstruction. --- ### First Token Matters: Understanding Safety Collapse in Large Reasoning Models Authors: Yizheng Yang, Haining Yu, Yuechen Wang, Yikai Hou, Xing Fu, Jinbo Yang, Tianqing Zhu Published: 2026-09-16 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.18471v1 Large Reasoning Models (LRMs) exhibit strong problem-solving abilities, yet their safety alignment often degrades when handling harmful queries. Existing approaches to improving safety largely rely on additional training or preference optimization, while offering limited understanding of the internal mechanisms behind safety failures. In this work, we investigate this failure through a token-level positional analysis of refusal dynamics and identify a localized vulnerability at the onset of reasoning, which we term Onset Refusal Collapse (ORC). We find that the refusal-related signal of LRMs drops sharply at the first generated token under harmful queries, which is associated with unsafe response generation. Motivated by this finding, we propose SafeToken, a lightweight inference-time intervention that injects a learned continuous safety anchor precisely at reasoning onset. Despite updating only a single token embedding, SafeToken effectively mitigates ORC, improves safety on harmful-query benchmarks, and largely preserves reasoning utility. These results suggest that safety failures in LRMs can arise from a transient breakdown at the critical transition from understanding to generation. --- ### Divide and Conquer: Mixture-of-Bottleneck Experts in Informative Ordinal Space for Video-based Multimodal Sentiment Analysis Authors: Ronghao Lin, Qiaolin He, Zefeng Lu, Yichu Liu, Li Huang, Sijie Mai, Haifeng Hu, Yap-peng Tan Published: 2026-09-16 Categories: cs.MM, cs.CL arXiv: https://arxiv.org/abs/2609.18470v1 Video-based Multimodal sentiment analysis (MSA) must handle information from text, audio, and image sequence in human speaking videos, yet current methods often fail to integrate modalities with task awareness. Most models treat video sentiment prediction as a single task, overlooking its ordinal nature, and their fusion strategies struggle to capture diverse unique and synergic cues across modalities. To address these limitations, we adopt a divide-and-conquer perspective by reformulating MSA as an ordinal regression problem and decoupling it into polarity recognition and intensity prediction. Driven by information theory, we introduce a Mixture-of-Bottleneck (MoB) framework that assigns different latents to polarity- and intensity-specific experts for different modalities. With the learning of information bottleneck, each expert learns compact and task-relevant representations while filtering out redundancy and noise. A multimodal bottleneck routing fusion module then fuses these expert latents with hard mining strategy, guiding the prediction in the ordinal sentiment space. Extensive experiments on 4 MSA datasets and 4 language models show that MoB effectively leverages informative latents from diverse modalities and captures general sentiment structure. Beyond stronger performance, MoB comprehensively captures fine-grained intra- and inter-modal dynamics, enabling more trustworthy localization of nuanced video sentiment signals. --- ### Spatially Adaptive Noise Injection Authors: Frantzeska Lavda, Maciej Falkiewicz, Van Khoa Nguyen, Alexandros Kalousis Published: 2026-09-16 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.18466v1 Diffusion samplers reverse a learned noising process using either stochastic (DDPM) or deterministic (DDIM) updates, which represent endpoints of a single family controlled by a scalar noise-injection variance that is applied identically at every spatial location. This uniform approach neglects the geometry of natural images: high-curvature regions such as edges and textures, where the denoiser is uncertain, benefit from stochastic correction, whereas smooth regions, where the score is precise, are degraded by injected noise. This work investigates whether each pixel requires stochastic correction at a given timestep and introduces Spatially Adaptive Noise Injection (SANI), a novel sampling framework that dynamically adjusts noise application on a per-pixel basis. SANI integrates a probabilistic gating mechanism with a derived spatially adaptive variance, ensuring that noise is injected precisely where needed to refine complex features while preserving well-formed structures. Experimental results and decoupling ablations demonstrate that SANI consistently improves Fréchet Inception Distance (FID) over the vanilla DDPM and DDIM endpoint samplers across diverse sampling timesteps, while remaining competitive with variance-learning baselines, highlighting the importance of spatial adaptivity in diffusion sampling. --- ### GeoCond: A Conditioning-Aware Reliability Adapter for Feed-Forward 3D Reconstruction Authors: David Ahmedt-Aristizabal, Mohammad Ali Armin, Russell Tsuchida, Lars Petersson Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18465v1 Feed-forward 3D foundation models such as VGGT predict cameras, depth, and point maps in a single pass, but can fail silently under low overlap, low parallax, and extreme relative rotation. Stratified analyses over these factors show that these failures are governed by geometric conditioning and are poorly captured by native aleatoric confidence. We introduce GeoCond, a lightweight reliability adapter for frozen feed-forward 3D backbones. GeoCond reads the backbone's predicted geometry and outputs pose-level uncertainty and a refinement gate. During training, it can be supervised by frame-permutation orbit variance, ground-truth pose error when labels are available, or cycle residuals from unlabelled independent pose graphs. At inference, the default head requires only one backbone pass and a small MLP. On VGGT, GeoCond improves out-of-distribution (OOD) AUSE (area under the sparsification-error curve; lower is better) from $0.32$ to $0.20$ over native confidence, transfers zero-shot to outdoor extreme-view scenes, and avoids the collapse caused by applying bundle adjustment uniformly. Across multiple backbones, cycle-distilled variants provide a ground-truth-free adaptation route, including cases where permutation variance vanishes on equivariant models. The same reliability signal supports gated refinement, pose-graph weighting, calibration, curation, and capture decisions. Reliable feed-forward 3D reconstruction requires not only predicting geometry, but also knowing when that geometry should be trusted. --- ### CSWAM: Better Causal Semantic Representations for Out-of-Distribution Generalization in World Action Models Authors: Tianbin Liu, Jian Zhu, Taiyi Su, Jianjun Zhang, Chong Ma, Zitai Huang, Weiyi Lu, Yi Xu Published: 2026-09-16 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.18462v3 FastWAM-style world action models enable efficient action-only inference, but generalize poorly under visual distribution shifts. Their reconstruction-oriented representations emphasize appearance-specific details, limiting generalization to unseen scenes and objects. Without observation history, the model also lacks temporal evidence for robustly identifying task-relevant state changes and motion in unfamiliar visual conditions. To address these limitations, we present the Causal Semantic World Action Model (CSWAM), which augments FastWAM with a causal semantic expert built on V-JEPA 2.1. V-JEPA provides temporally grounded representations of semantic state changes and motion with less dependence on appearance-specific details. The expert learns their future evolution from a sparse history of current and past observations and shares the history-derived context with both the video and action streams through causal attention. At inference, CSWAM conditions action denoising on the current video state and observed semantic history, retaining efficient action-only inference. We conduct simulation and real-robot experiments to evaluate generalization under distribution shifts. With embodied pretraining, CSWAM raises Randomized success on RoboTwin 2.0 Clean-to-Randomized transfer from 10.16% to 45.18%, a gain of 35.02 percentage points over FastWAM. Across two real-robot tasks and three OOD difficulty levels, CSWAM improves average success over FastWAM by 42.5 percentage points, from 27.5% to 70.0%. --- ### Disentangling Long-Term Memory via Latent Neuro-Symbolic Reasoning Authors: Cai Ke, Xinghao Chen, Xiaoyu Shen, Keyu Chen, Siyu An, Junnan Dong, Ruifeng Xu, Ruizhi Qiao, Xing Sun Published: 2026-09-16 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.18461v2 Personalized agents are required to reason over long-term history interactions to infer both explicit preferences and implicit behavioral evidence. While early flat retrieval methods score memory fragments independently and neglect the distributed information, current structured memory frameworks rely on query-agnostic static graphs that fail to capture the context-dependent relations. Crucially, raw textual memories are inherently entangled and noisy, making fine-grained personalization and cross-session reasoning computationally prohibitive. To this end, we present LGM, a novel neuro-symbolic framework that shifts long-term memory disentanglement into a continuous latent space. Specifically, (i) instead of persisting fixed graphs, we design a tailored latent graph construction with a sparse autoencoder. Subject to each query, it maps historical interactions into latent memory nodes and disentangles the memory traces into sparse concept activations, dynamically synthesizing query-aware relational edge weights. (ii) A graph encoder then treats the query embedding as a conditioning preference to direct non-linear message passing across the task-specific latent subgraph. This yields a highly expressive memory representation for effective activations. Extensive experiments on long-term personalization benchmarks demonstrate that LGM significantly outperforms state-of-the-art baselines in capturing both explicit and implicit preferences while enabling personalized responses. --- ### Collective Loss of Control in LLM Agent Systems: An Epidemic Account of Mutation, Contagion, and Recovery Authors: Xiangfan Wu, Zonghao Ying, Huiyu Wu, Xing Zheng, Huangsheng Cheng, Xiaorong Shi, Jing Guo Published: 2026-09-16 Categories: cs.AI, cs.CR arXiv: https://arxiv.org/abs/2609.18460v1 How does a multi-agent system evolve from a local deviation into collective loss of control? We propose an epidemic explanation organized around accidental mutation, contagion, and recovery. A spontaneous deviation creates a seed; communication enables other agents to adopt and retransmit its unsafe strategy; collective failure can emerge when propagation outpaces correction and containment. Thus, rare individual deviations can coexist with substantial collective risk. Motivated by reported OpenAI agent coordination incidents, we examine two ingredients of this mechanism. A deployment audit identifies implicit communication paths between nominally independent evaluation runs and verifies transport through a default Docker backend. RogueHandoff-20, a benchmark of 20 executable scenarios, tests recipient susceptibility by injecting unsafe trajectories generated by a modified Qwen-27B route. Across four native-pending routes, executed harm is 0-5% on normal tasks and 40-95% after injection, exceeding paired direct malicious requests by 5-45 percentage points. These results support low observed baseline harm alongside high conditional susceptibility; they do not establish natural rare-event rates or demonstrate an autonomous cascade. The account motivates complementary defenses: strengthen resistance and recovery alongside prevention of spontaneous deviations, and audit and restrict unintended communication paths that can turn local failures into collective loss of control. --- ### The Mirage of Calibrated Confidence: Trajectory-Independence of Verbalized Confidence in Vision-Language Models Authors: Jisoo Yang, Jaeho Han, Trung X. Pham, Junyeong Kim Published: 2026-09-16 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.18453v1 A calibrated Vision-Language Model (VLM) can repeatedly self-correct, say "Wait, I should recheck," arrive at the wrong answer, and still report high confidence. We find that this occurs because verbalized confidence is largely trajectory-independent in the VLMs and calibration methods we evaluate. We examine this through three complementary lenses: content variation, token masking, and the model's own hesitation markers. We show that confidence is insufficiently sensitive to what the reasoning trajectory actually contains, and that calibration training can paradoxically worsen this disconnect. Since existing metrics like ECE and AUROC cannot detect this problem, we propose the Trajectory-Grounding Score (TGS) in two complementary forms: TGS-self, which compares confidence with and without access to the model's own trajectory, and TGS-pair, which tests whether the model assigns higher confidence to correct trajectories than to flawed ones along the vision, reasoning, and answer axes. We propose TGS-Bench, a model-agnostic suite spanning 10 benchmarks with controlled good/bad trajectory pairs, and show that conventional calibration rankings diverge from trajectory-grounding rankings, exposing a blind spot in current evaluation practice. --- ### M-SQE: Multilingual Skill Quality Estimation for Enhancing Language Equality in Agentic Skill Use Authors: Yilun Liu, Shimin Tao, Minggui He, Chenxin Liu, Li Zhang, Chen Liu, Miao Zhang, Jiaxin Guo, Min Zhang, Liqun Deng, Xiaojun Meng, Daimeng Wei Published: 2026-09-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.18445v1 Agent skills, reusable procedural documents that extend LLM agents beyond their parametric memory, have become an important interface for deploying agents on real-world tasks. Community-maintained skill libraries built around this interface are growing rapidly. However, this ecosystem remains deeply English-centric: our audit finds that low-resource languages such as Swahili and Hindi have no in-language skill content, so retrieval often returns a skill written in a different language than the query, degrading accuracy and recall. A practical solution is to synthesize in-language skills for retrieval but the quality can be unreliable, so relevance in this setting alone often surfaces a related but unusable candidate. To address this, we propose M-SQE, a post-retrieval Multilingual Skill Quality Estimation framework that scores candidates via a Theory view for intrinsic quality and an Action view for task-grounded utility, unified into a domain-conditioned final score. We evaluate M-SQE across three skill-use domains: general, tool-use, and cultural tasks. Empirically, we build three-layer candidate skill pools mirroring today's ecosystem, where M-SQE's task success exceeds existing baseline's average by at least +3.5 points across three different retrievers. Particularly, M-SQE lifts the lowest-resource languages most (+12.9pp on Hindi and +5.6pp on Swahili) and achieves strong performance across all six culture regions, thereby moving agentic skill use toward linguistic and cultural equality. --- ### DR.WILSS: Diffusion-Based Replay for Weakly Supervised Continual Semantic Segmentation Authors: Leon Arthur Marx, Francesco Barbato, Matteo Caligiuri, Pietro Zanuttigh Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18444v1 Weakly supervised class-incremental semantic segmentation (WILSS) aims to train a segmentation model over multiple steps, each introducing new concepts to be learned with only image-level supervision. We introduce DR$.$WILSS, an innovative approach to address catastrophic forgetting in continual learning using diffusion-based generative replay. Our framework leverages language clues to guide the diffusion process, employing self-inpainting and regularization techniques to efficiently produce replay data, aiding the learning process. By generating high-quality replay data, the information from previously learned classes can be preserved during continual updates, a critical challenge in incremental learning scenarios. To further align the statistics of replay data with those of training samples, we apply LoRAs to the generative model. Experimental results demonstrate state-of-the-art performance across multiple benchmarks and generative architectures, while avoiding storage of training data and the use of additional resource-demanding tools during training. The proposed technique enables an optimal tradeoff between training complexity and inference-time accuracy, making DR$.$WILSS a promising solution for real-world applications. --- ### Risk-Aware World Modeling with Flow-Guided Occupancy Evolution for Selective Trajectory Planning in Automated Driving Authors: Rongxiang Zeng, Linsen Cai, Jiafu Zhang, Yijie Zhong, Yide Tao, Shuai Wang, Nan Zheng, Hai L. Vu, Alvaro Garcia Hernandez, Yongqi Dong Published: 2026-09-16 Categories: cs.AI, cs.ET, cs.LG, cs.RO, eess.SY arXiv: https://arxiv.org/abs/2609.18442v1 Safe motion planning in automated driving requires anticipating evolving traffic risks and deciding when to revise the current planned trajectory. We introduce RiskWorld, a risk-aware world modeling framework for shared occupancy forecasting and selective trajectory replacement. Spatial risk fields and temporal actor context are fused with visual bird's-eye-view features. Flow-guided evolution transports occupancy and scene features, while signed residuals correct occupancy after transport. One forecast is generated per planning step and reused across candidates. Each candidate is compared with a current-state persistence reference, yielding a nonnegative collision-score correction. The trajectory selected by current-world evaluation serves as the planning anchor and is replaced only when additional predicted risk triggers intervention and an alternative satisfies component-wise constraints on predicted risk and trajectory error. Candidate geometries remain unchanged. We evaluate RiskWorld for open-loop planning on nuScenes using camera features, annotation-derived current and historical actor states, and dataset-provided map context. RiskWorld achieves the lowest collision rate at a long evaluation horizon of 3 s, and the second-best average L2 error among various state-of-the-art baselines, while running at 11.5 FPS on a single NVIDIA RTX 4090 with 90.81 M parameters. Within-setting ablations show that RiskWorld achieves lower collision rates than the current-state rescoring baseline, while forecast reuse enables additional candidates to be evaluated at low marginal computational cost. --- ### Multitask Reinforcement Learning for Assisting Choice Model Specification Authors: Gabriel Nova, Stephane Hess, Sander Van Cranenburgh Published: 2026-09-16 Categories: econ.GN, cs.AI arXiv: https://arxiv.org/abs/2609.18441v1 Discrete choice model specification is a time-consuming task in which modellers often specify and estimate multiple models while balancing goodness-of-fit, parsimony, and behavioural plausibility. We present Delphos, a multitask reinforcement learning framework that learns transferable specification strategies across transport choice datasets. Delphos frames model specification as a sequential decision-making problem in which it applies a sequence of modelling actions and receives feedback from an estimation environment based on model performance and convergence. To transfer modelling decisions across datasets with different sets of variables, Delphos represents utility specifications as sets of modelling terms using a DeepSet-Q architecture, allowing a shared specification policy to learn across multiple datasets. Trained on nine transport choice datasets, Delphos consistently outperforms independently trained single-task agents, indicating that sharing modelling experience improves learning efficiency and helps identify promising sequences of modelling decisions with fewer unsuccessful estimation attempts. When applied without further training to the unseen Swissmetro and Decisions datasets, the same agent identifies competitive specifications in less than 20 minutes on a standard CPU. It achieves a higher log-likelihood per observation than the VNS metaheuristic on Swissmetro and performance comparable to a published MNL specification developed by expert modellers on Decisions. These findings show that accumulating and reusing modelling experience enables Delphos to function as an intelligent assistant for discrete choice model specification. It reduces manual trial-and-error while allowing modellers to retain control over model diagnosis, refinement, and final selection. --- ### Planning or Improvisation? Stress-Testing the Poetry Planning Site on Open Models and Open Cross-Layer Transcoders Authors: Éric Jacopin Published: 2026-09-16 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.18440v1 Lindsey et al. (2025) report that Claude 3.5 Haiku plans rhymes: features for candidate rhyme words are active on the newline before a line is written, and a suppress-and-inject intervention redirects the line only when applied there (their Figure 13). We test how far this generalizes on seven cells crossing four open models (0.6B to 2.6B parameters) with six open cross-layer transcoders (CLTs), on one consumer GPU, decomposing the claim into position specificity (C1), newline site identity (C2), and a newline-resident plan (C3). This is a stress test rather than a faithful reproduction: attribution graphs are unavailable for these CLTs, so features are found bottom-up from decoder vectors. C1 generalizes, in every cell and in all 247 of 444 prompt-by-inject pairs with a detectable effect, but the effective position is the final prompt token, adjacent to emission, and only two cells reach behaviorally meaningful probabilities. C2 and C3 are not recovered by any probe: a census of every active feature finds no rhyme-anticipating enrichment at the newline, and steering the newline while the model composes the whole line, over 36 runs and 8,640 sampled lines, shows why. That intervention is strong but one token long, making the injected word the first word of the composed line in 703 of 720 samples and leaving the rhyme six words later untouched. A final test drops the transcoder entirely: patching the newline's whole residual, at every layer, from a minimal-pair poem whose third line ends on a different rhyme moves the rhyme in 11 of 1,260 composed lines against 4 at baseline, with a design resolving 1.4%. We read this as a boundary condition rather than a refutation: at this scale and with these transcoders, the causal site is emission-adjacent. We reproduce Figure 13's shape, not its mechanism. Code and data are public (code: github.com/PCfVW/poetry-planning-site). --- ### WetRobo: A Reproducible Robot Kit for Coding Agents in Biological Laboratories Authors: Yuna Oikawa, Kei Endo, Takanori Uzawa, Yunzhe Zhang, Manan Anjaria, Lerrel Pinto, Sherry Yang, Koji Tsuda Published: 2026-09-16 Categories: cs.AI, cs.RO arXiv: https://arxiv.org/abs/2609.18435v1 Automating biological research requires general-purpose, reproducible robot systems that allow individual wet-lab researchers to delegate robot tasks without performing teleoperation or neural-network training. Vision-language-action policies have been proposed for general-purpose arms, but can lose performance when their operating environment changes. We therefore built WetRobo, a robot kit that can readily transfer between laboratories. It consists of one robot arm, laboratory equipment (an incubator, a reagent bottle with a cap, and a Petri dish), the existing code that moves the arm, teleoperation demonstrations of each task that we recorded, and a general AGENTS.md skill file. A biological experimentalist provides natural-language tasks without collecting local teleoperation training data or training a neural network. The coding agent observes the local laboratory and writes and executes programs, using external tools as needed for adaptation. We demonstrate use of WetRobo with OpenAI Codex (gpt-5.6-sol) on three successful tasks: lifting a Petri dish lid, removing a bottle cap, and opening the incubator door, all in real-world laboratories. The coding agent achieved the cap task in both laboratories, Lab X and Lab Y, whereas a VLA fine-tuned on Lab X demonstrations succeeded there but failed to transfer to Lab Y. These results point to a practical route for laboratory robotics: instead of training a policy for each laboratory, distribute a kit and let a coding agent adapt it in each laboratory. Code, demonstrations, and the evolved programs are available at https://github.com/tsudalab/WetRobo. --- ### Occluded Gait Recognition with Mixture of Experts: An Action Detection Perspective Authors: Panjian Huang, Yunjie Peng, Saihui Hou, Chunshui Cao, Xu Liu, Zhiqiang He, Yongzhen Huang Published: 2026-09-16 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.18432v1 Extensive occlusions in real-world scenarios pose challenges to gait recognition due to missing and noisy information, as well as body misalignment in position and scale. We argue that rich dynamic contextual information within a gait sequence inherently possesses occlusion-solving traits: 1) Adjacent frames with gait continuity allow holistic body regions to infer occluded body regions; 2) Gait cycles allow information integration between holistic actions and occluded actions. Therefore, we introduce an action detection perspective where a gait sequence is regarded as a composition of actions. To detect accurate actions under complex occlusion scenarios, we propose an Action Detection Based Mixture of Experts (GaitMoE), consisting of Mixture of Temporal Experts (MTE) and Mixture of Action Experts (MAE). MTE adaptively constructs action anchors by temporal experts and MAE adaptively constructs action proposals from action anchors by action experts. Especially, action detection as a proxy task with gait recognition is an end-to-end joint training only with ID labels. In addition, due to the lack of a unified occluded benchmark, we construct a pioneering Occluded Gait database (OccGait), containing rich occlusion scenarios and annotations of occlusion types. Extensive experiments on OccGait, OccCASIA-B,Gait3D and GREW demonstrate the superior performance of GaitMoE.OccGait is available at https://github.com/BNU-IVC/OccGait. ---