# Research Papers — 2026-07 · bundle page 57 of 222 > Full-text markdown of papers #2801–2850 of 11075 published in 2026-07, newest first. AI-readable mirrors of the full 11075-paper 2026-07 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-07/page/57 [← Previous bundle](https://aipod.dev/papers/archive/2026-07/page/56.md) [Next bundle →](https://aipod.dev/papers/archive/2026-07/page/58.md) ### Robustifying pathology foundation models via fine-tuning Authors: Alexandre Filiot, Oskar Thaeter, Benoit Schmauch, Lionel Guillou Published: 2026-07-24 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.22861v1 Pathology foundation models (FMs) produce powerful tile-level representations which remain sensitive to scanner and staining variability, undermining deployment across laboratories. We develop a novel fine-tuning recipe that improves the robustness of pathology FMs to acquisition factors. Applied to ten different FMs, our fine-tuning strategy consistently improves robustness for every model as well as downstream performance, with no observed trade-off. On average, it raises the PathoROB robustness index by 23% (from 0.72 to 0.87) and increases the overall cross-benchmark performance by 43% on Patho-Bench, HEST and THUNDER combined, with individual gains reaching up to 72% in robustness (Phikon-v2) and 76% in performance (Midnight-12k). We publicly release the fine-tuned versions of Phikon-v2 (Phaet) and Midnight-12k (Mascaret) at https://huggingface.co/wearewaiv/models. --- ### PatiGonit22K: A Comprehensive Dataset for Solving Complex Bengali MWPs Authors: Swastika Kundu, Azizul Hakim Fayaz, Tashreef Muhammad Published: 2026-07-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2607.22859v1 Mathematical Word Problems (MWPs) are an important benchmark for evaluating natural language understanding and quantitative reasoning. Despite recent progress in high resource languages, Bengali remains underexplored due to the limited availability of large scale annotated datasets. In this work, we introduce PatiGonit22K, an expanded Bengali MWP dataset containing 22,441 problems, developed by extending the original PatiGonit dataset with a substantially larger collection of complex mathematical problems. The dataset includes both simple and multi operation equations, providing a balanced benchmark for evaluating mathematical reasoning across different difficulty levels. Each problem is carefully translated, annotated, culturally adapted, and verified to ensure linguistic consistency and mathematical correctness. By increasing both the scale and complexity of Bengali MWPs, PatiGonit22K provides a more comprehensive resource for future research on mathematical reasoning and educational NLP applications in low resource languages. --- ### Coordinated Networking for On-Device Agent-Augmented Real-Time Communication Authors: Goodsol Lee, Juheon Yi, Jinglu Wang, Haowen Xu, Saewoong Bahk, Yan Lu Published: 2026-07-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.22854v1 AI agents are enabling a new paradigm of agent-augmented real-time communication (RTC), where humans focus on high-level collaboration, while agents autonomously retrieve, analyze, and generate information in real time to support their interactions. These apps enable new experiences across various domains: for example, when corporate employees co-author a legal document, their agents can discuss and draft on their behalf, sparing them the burden of manually reviewing each other's work. As existing cloud-based agents suffer from privacy risks and unscalable server costs, on-device agent-augmented RTC offers a promising alternative. However, this on-device paradigm introduces a new networking challenge: contention between concurrent traffic flows generated by humans (for live video streaming) and agents (for sending context files for analysis). We design HFS, a framework to ensure both high live video quality and low agent response latency in agent-augmented RTC apps. We achieve the goal through an app-guided multi-flow transport approach, where a unified app-layer orchestrator jointly controls the sending rates of live video and agent context flows based on their heterogeneous app requirements. Our prototype built atop WebRTC and llama.cpp demonstrates that HAFS outperforms baselines, achieving 1.5x higher video quality while reducing agent response time by 31%. --- ### Gaze-Anchored Social Net: Decoding Implicit Relations via Joint Modeling Authors: Yuqi Hou, Zhuo Chen, Han Hu, Je Woo Kim, Jianbo Jiao, Hyung Jin Chang Published: 2026-07-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.22847v1 Human gaze does more than point to visual targets; it serves as a subtle indicator of social intent within static images, whereas standard models typically process individuals independently, treating gaze as an i.i.d. quantity or predicting social semantics in isolation. Recent multi-person methods attempt to address this but often treat social relations as rigid, post-hoc classifications decoupled from the gaze estimation process. This oversimplification fails to capture the nuanced nature of social intent, which acts as an underlying driver of gaze behavior rather than a secondary categorical output. We address these limitations by proposing ANCHOR, a target-centric paradigm designed to decode gaze-anchored social intent by modeling the joint distribution of visual attention and latent implicit relations. Our approach surfaces these dependencies as the latent structural scaffolding of gaze behavior. The architecture utilizes a relational attention mechanism to capture fine-grained interpersonal links, leveraging feature-wise modulation for efficient multi-person parsing from a single vision backbone. To stabilize the training of this coupled formulation, we implement an optimization synergy to resolve the inherent conflicts between spatial gaze accuracy and latent social reasoning. This approach ensures robust generalization by seeking stable, flat minima while simultaneously harmonizing competing task gradients. We validate our framework on an extended benchmark featuring dense multi-person annotations and novel social influence rankings. Our results demonstrate state-of-the-art performance and provide the first quantitative evidence that implicit social hierarchies can be robustly disentangled and learned directly from static gaze patterns. --- ### Queryable Self-Organizing Maps: A Database Abstraction for Topology-Driven Data Exploration Authors: Denis Mayr Lima Martins, Gottfried Vossen Published: 2026-07-24 Categories: cs.DB, cs.LG arXiv: https://arxiv.org/abs/2607.22843v1 Self-Organizing Maps (SOMs) have long been used as exploratory tools for high-dimensional data: they organize objects into a two-dimensional topology that reveals clusters, gradients, sparse regions, dense regions, and boundaries. Yet, in modern data systems, SOMs are typically trained and visualized outside the DBMS, disconnected from the relational data they summarize. We introduce the abstraction of a queryable data map: a learned topological artifact consisting of representatives, neighborhood relations, object assignments, and derived summaries. We instantiate this idea with MapDB, a lightweight prototype that makes SOM artifacts queryable so users can explore data topology without leaving the database. Experimental study shows that SOM training is feasible at moderate analytical scale, that map queries are interactive after materialization, and that SOM regions provide meaningful targets for exploratory SQL. --- ### Language-Routed RAG and Direct Option Scoring for Multilingual Financial QA: DS@GT at FinMMEval Authors: Justice Ayela, Kabir Sahni Published: 2026-07-24 Categories: cs.IR, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2607.22841v1 We present DS@GT's submission to FinMMEval 2026 Task 1, a multilingual financial exam question answering benchmark spanning English, Spanish, Greek, Chinese, and Hindi. Financial certification exams such as the CFA, EFPA, and CPA demand structured domain reasoning that standard NLP benchmarks do not capture, and this challenge compounds across languages where retrieval and representation infrastructure is underdeveloped. We build a retrieval-augmented pipeline on LangGraph that detects query language and retrieves semantically relevant exemplars from a 30,209-entry multilingual knowledge base using BGE-M3 embeddings and FAISS indexing. The system then scores answers via Retrieval-Augmented Direct Scoring (RADS), reading next-token log-probabilities over candidate option letters rather than generating free-form output. For low-resource languages, we fuse per-language and cross-lingual retrieval indices using weighted Reciprocal Rank Fusion. Model selection is language-routed: Qwen3-14B for Arabic, Chinese, and Hindi; Qwen2.5-14B for English; and Llama-3.1-8B for Greek, a routing derived from empirical ablations that reveal substantial language-asymmetric performance gaps. Notably, chain-of-thought prompting significantly degrades Greek accuracy (90.7% to 20.9%), and enabling Qwen3's default thinking mode collapses Arabic RADS performance to near-chance levels. Our results indicate that effective multilingual financial reasoning requires language-aware retrieval, model routing, and deliberate scoring strategy selection. --- ### Frustratingly Simple Black-Box Adaptation of Language Models via Logit Bias Authors: Ofek I. Cohen, Lior Shani, Aviv Rosenberg, Ankur Samanta, Tal Wagner, Yonathan Efroni Published: 2026-07-24 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2607.22837v1 Many organizations aim to adapt language models for internal use, both to improve performance on domain-specific tasks and to address privacy concerns around sensitive data. However, such adaptation remains non-trivial: it often requires operationally challenging fine-tuning of open-source models or ad hoc prompt optimization. We study a minimal alternative based on a simple API-level control: allowing users to bias the model's logits with a user-defined vector. We develop a black-box method for learning a single context-independent logit-bias vector, added at every decoding step, without modifying model weights or requiring gradients. Starting from a KL-regularized reinforcement learning (RL) objective, we characterize when such a fixed logit-bias vector can approximate the optimal prefix-dependent correction and derive a closed-form inverse-propensity estimator from rollouts, rewards, and token probabilities. Empirically, this simple decoding-time intervention improves over base models on mathematical and reasoning benchmarks while using far fewer trainable parameters than conventional fine-tuning. Our results suggest that learned logit bias is a lightweight mechanism for adapting language models under minimal access requirements. --- ### MEMENTO: Memory-Guided Memetic Code-as-Policy Evolution Authors: Alkis Sygkounas, Victor Aregbede, Amy Loutfi, Andreas Persson Published: 2026-07-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.22832v1 Long-horizon embodied tasks require policies that execute many dependent actions before task success can be observed. Representing policies as executable control pro- grams (code-as-policy) enables their decision logic to be inspected and revised after rollout evaluation. Revised programs can then be executed and compared by rollout performance, framing policy improvement as execution-guided program search. Evo- lutionary methods driven by large language models (LLMs) provide a natural mecha- nism for this search by generating variants and selecting high-performing candidates. However, existing approaches primarily select among independently generated vari- ants and lack a sequential local improvement phase. We introduce MEMENTO, a memory-guided single-elite memetic framework for code-as-policy evolution. ME- MENTO first evolves a rollout evaluator that maps policy rollouts to scalar fitness and structured feedback metrics. Fitness selects accepted candidates and the next elite, while feedback metrics condition policy proposals generated by memory-guided hill-climbing, macro-mutation, and crossover. We evaluate MEMENTO on two long- horizon embodied domains: Robosuite Franka Tower-of-Hanoi manipulation and AI2- THOR household interaction. MEMENTO outperforms Eureka and REvolve, adapted as code-as-policy evolutionary baselines, in task success and generalization to held- out Robosuite object configurations and unseen AI2-THOR scenes. Ablations show that zero-shot generation and unevolved evaluators fail to solve either domain, and that removing policy-search branches reduces performance. Finally, we deploy the best-evolved Robosuite policy on a physical Franka robot, demonstrating the feasibil- ity of sim-to-real transfer of the evolved code-as-policy. Code, prompts, and videos are available at: https://github.com/sygkounas/MEMENTO. --- ### ID-V2V: Identity-Preserving Video Restylization Authors: Yuancheng Xu, Mingming He, Pablo Salamanca, Li Ma, Yash Kant, Emmett Steven, Paul Debevec, Ning Yu Published: 2026-07-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.22830v2 In visual storytelling, human performances are central to creative intent and narrative meaning. However, preserving human identity and performance while enabling flexible visual edits remains challenging for generative video models. We formalize this challenge as identity-preserving video restylization, which propagates scene, lighting, and style changes specified by an edited keyframe across a source video, while preserving facial likeness and performance, including expressions, eye gaze, and lip synchronization. A key obstacle is the absence of paired training data, as identity-preserving restylized video pairs are rare in real-world settings. To address this, we propose a decoupling of source-grounded identity preservation and edit-driven video synthesis. Our key insight is that facial appearance and expression should remain invariant, with illumination being the primary permissible variation. We therefore cast identity preservation as a video relighting problem, while modeling visual edit propagation as controlled video synthesis guided by the edited keyframe. Building on this formulation, we introduce ID-V2V, a video-to-video generative framework integrating complementary control signals: relit facial regions and facial normal maps tightly constrain facial likeness and performance, while edited keyframes and depth sequences enable flexible and temporally coherent generation. This design enables constructing training pairs from a single video, eliminating the need for scarce paired data. Extensive experiments demonstrate that ID-V2V significantly outperforms existing methods in preserving facial likeness and fine-grained facial performance, supports both single- and multi-subject scenarios, and delivers high visual quality, highlighting its potential as a human-centric tool for real-world content production. The code is available at: https://github.com/Eyeline-Labs/ID-V2V. --- ### ID-V2V: Identity-Preserving Video Restylization Authors: Yuancheng Xu, Mingming He, Pablo Salamanca, Li Ma, Yash Kant, Emmett Steven, Paul Debevec, Ning Yu Published: 2026-07-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.22830v1 In visual storytelling, human performances are central to creative intent and narrative meaning. However, preserving human identity and performance while enabling flexible visual edits remains challenging for generative video models. We formalize this challenge as identity-preserving video restylization, which propagates scene, lighting, and style changes specified by an edited keyframe across a source video, while preserving facial likeness and performance, including expressions, eye gaze, and lip synchronization. A key obstacle is the absence of paired training data, as identity-preserving restylized video pairs are rare in real-world settings. To address this, we propose a decoupling of source-grounded identity preservation and edit-driven video synthesis. Our key insight is that facial appearance and expression should remain invariant, with illumination being the primary permissible variation. We therefore cast identity preservation as a video relighting problem, while modeling visual edit propagation as controlled video synthesis guided by the edited keyframe. Building on this formulation, we introduce ID-V2V, a video-to-video generative framework integrating complementary control signals: relit facial regions and facial normal maps tightly constrain facial likeness and performance, while edited keyframes and depth sequences enable flexible and temporally coherent generation. This design enables constructing training pairs from a single video, eliminating the need for scarce paired data. Extensive experiments demonstrate that ID-V2V significantly outperforms existing methods in preserving facial likeness and fine-grained facial performance, supports both single- and multi-subject scenarios, and delivers high visual quality, highlighting its potential as a human-centric tool for real-world content production. The code is available at: https://github.com/Eyeline-Labs/ID-V2V. --- ### Disentangling Multi-View Scanning in Mamba for Network Traffic Anomaly Detection Authors: Xinglin Lian, Chengtai Cao, Ting Zhong, Fan Zhou Published: 2026-07-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.22829v1 Network Traffic Anomaly Detection (NTAD) is a critical task in cybersecurity, yet timely and accurate anomaly detection remains challenging. Mamba has emerged as a particularly promising backbone for NTAD due to its linear-time complexity for long-sequence modeling. It further incorporates a dedicated multi-view scanning mechanism to enhance detection precision through complementary contextual cues. However, we identify a previously overlooked structural deficiency in multi-view Mamba scanning for NTAD: redundancy accumulation. Specifically, distinct scanning branches capture substantial view-invariant information, which is repeatedly amplified during multi-view fusion; conversely, view-specific information is diluted or even suppressed, leading to representation homogenization and multi-view degradation. To address this problem, we propose DisenMamba, a novel disentangled multi-view Mamba framework. DisenMamba reformulates multi-view scanning as a two-stage disentangle-then-fuse process that explicitly separates view-invariant and view-specific components prior to fusion. This design prevents the invariant information accumulation while preserving complementary multi-view cues, yielding more discriminative representations for subtle traffic anomalies. Extensive experiments demonstrate the effectiveness of DisenMamba, establishing a new disentangled multi-view Mamba paradigm. Code is available at https://github.com/ikun0124/DisenMamba. --- ### Agentic Autoresearch for CT Reconstruction Authors: Andreas Maier, Lucas Kachelriess, Siming Bayer, Yixing Huang, Yan Xia, Amber Simpson, Moritz Zaiss Published: 2026-07-24 Categories: physics.med-ph, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2607.22824v1 Comparing CT reconstruction methods fairly is labor-intensive and largely manual, and many benchmarks use idealized data. We ask whether a large language model (LLM) agent can do the labor of reconstruction research on its own, and whether a ranking measured on ideal data predicts behavior under realistic noise. We built an agentic loop: the agent edits a solver, runs a short cluster job, reads one frozen metric, and revises. The metric is a calibrated headroom score against the FBP baseline, inside the field of view; every method shares the same differentiable fan-beam projector. We benchmarked 26 methods on Mayo low-dose CT (noise-limited) and a 128-view sparse-view breast task from the noiseless DL-Sparse-View Challenge, with validation-selected iterations scored on a held-out test set. Every trained breast model was then re-scored on noisy inputs (I_0 = 10^5 photons) without retraining, and separately retrained on matched noise. The agent independently implemented, tuned, and benchmarked all 26 methods, and recombined them into a compact solver of 969 parameters that ties the top Mayo tier at the 1% level using 0.4% of the champion's parameters. Benchmarking gives a tier of statistically indistinguishable top methods, not one winner. Mild input noise nearly inverts the breast ranking: the noiseless champion (a supervised image denoiser, hr 0.89) collapses to 0.00, while a learned primal-dual method rises to champion (0.72 to 0.93). An ideal-data leaderboard therefore does not predict robustness. The inversion is a transfer effect, not a permanent deficit: retraining on matched noise restores much of the clean ranking (Spearman rho 0.04 to 0.61). Noise is only the easiest confounder in an open-ended set (beam hardening, scatter, anatomy, disease), so no single-factor challenge certifies generality. Benchmarks should model a broad spectrum of realistic factors at once. --- ### Practical advantage beyond the quadratic speedup limit with fully-quantum walks Authors: Massimiliano Incudini, Guglielmo Mazzola Published: 2026-07-24 Categories: quant-ph, cond-mat.dis-nn, cs.LG arXiv: https://arxiv.org/abs/2607.22818v1 We introduce a new class of fully-quantum Metropolis walks in which both the proposal and acceptance steps are intrinsically quantum. Unlike standard quantum walks obtained by quantizing classically efficient Markov chains, our algorithm employs Hamiltonian simulation as a quantum-native proposal mechanism, enlarging the class of quantum walks beyond classical counterparts. We target the problem of sampling from the low-temperature Gibbs distribution of classical dense Ising models, within a fixed error in total variation distance. This approach achieves about a cubic polynomial asymptotic advantage over previous quantum-walks, resulting in a total sixth-degree polynomial queries speedup compared to the best classical walk. This shows that speedups beyond the widely assumed quadratic limit are possible within the quantum walk formalism. We perform a complete fault-tolerant compilation of all algorithmic primitives and benchmark against CPU, GPU, and FPGA implementations of the best classical Markov chain. Under identical hardware assumptions, the resulting advantage runtime crossover is reduced from approximately $10^3$ years for conventional quantum walks to less than one day. These results identify fully-quantum Markov chains as a promising route toward practical quantum advantage. --- ### From Hybrid Mechanistic--Data-Driven Modeling Toward Neuro-Symbolic AI: What, Why, and How Authors: Moein E. Samadi, Andreas Schuppert Published: 2026-07-24 Categories: cs.LG, cs.AI, cs.LO, stat.ML arXiv: https://arxiv.org/abs/2607.22811v1 Hybrid mechanistic/data-driven models, which combine first-principles with learned components, are increasingly used in process engineering and scientific machine learning. Common hybrid modeling designs are specified primarily through their architectures and training losses, which offers a limited basis for a shared semantic interface to compare or verify them across domains, with comparatively little attention paid to epistemic uncertainty in the mechanistic part. We bridge hybrid modeling and neuro-symbolic (NeSy) AI by reconstructing these designs as instances of NeSy interface. The resulting translation, Hybrid-to-NeSy (H2N), places mechanistic knowledge on the language side, learned modules on the belief side, and validity domains together with constraints on the logic side. For each design, H2N then yields an explicit NeSy inference functional and a logic-belief decomposition. From this decomposition we derive two metrics: structural violation rate (SVR), measuring whether the learned belief respects the mechanistic structure; and belief dispersion (BD), measuring how concentrated the learned plausibility is, serving as a hybrid model's epistemic uncertainty in its mechanistic part. We instantiate H2N on a case study of a structured hybrid model for binary classification under label noise and show that models with higher SVR and BD exhibit greater variability in held-out accuracy. Under structural distribution shift, H2N further quantifies a model's uncertainty during extrapolations, whereas test accuracy reveals the same shift only post hoc. --- ### Hybrid Semantic and Spectral Ensemble for Robust Synthetic Image Source Attribution Authors: Md. Ajwad Hossain Published: 2026-07-24 Categories: cs.CV, cs.AI, cs.GR, eess.IV arXiv: https://arxiv.org/abs/2607.22808v1 The rapid advancement of text-to-image (T2I) models has necessitated robust Synthetic Image Source Attribution (SIA) methodologies. A critical challenge in SIA is the distribution shift between pristine training images and real-world deployed images, which undergo unknown post-processing operations such as JPEG compression and blurring. In this work, proposed for the DLMMDD Challenge at ICANN 2026, we introduce a dual-branch ensemble framework fusing Semantic Deep Learning with Mathematical Forensic Feature Extraction. The semantic branch employs EfficientNet-B0 regularized with Exponential Moving Averaging (EMA) and Label Smoothing. The forensic branch extracts 126 mathematical features -- including SVD spectral profiles and Local Binary Patterns -- from high-pass noise residuals, compressed via Truncated SVD and classified with XGBoost. Evaluated on a dataset of 10 generators where 55% of the test set is degraded, our approach achieves a private leaderboard accuracy of 95.60%. Furthermore, the entire pipeline is highly computationally efficient, requiring no GPU acceleration and executing end-to-end on a standard CPU in under 6.5 hours, highlighting the practicality and scalability of mathematical forensics for real-world deployment. --- ### The Best Programming Language for Tokenmaxxing: An Investigation of Coding Agent Behavior Across Programming Languages Authors: Zixuan Wu, Carolyn Jane Anderson, Arjun Guha Published: 2026-07-24 Categories: cs.SE, cs.CL arXiv: https://arxiv.org/abs/2607.22807v1 Although coding agents are now very effective in a variety of programming languages, this paper first shows that the cost (in tokens) can very significantly by programming language. We evaluate five recent models on programming problems in Python, Java, Rust, and OCaml. We carefully control for problem difficulty, and show that there can be stark variation in token consumption that is consistent across models. To understand why, we analyze both the structure and content of agent trajectories. First, we re-execute every intermediate solution and abstract each trajectory as a sequence of test-outcome vectors, then label the work between successive solutions. This reveals agents repeatedly producing noncompiling solutions in unfamiliar languages and revising solutions that already pass. Second, we analyze trajectory text, finding that agents plan solutions in code comments, distrust the provided tests in favor of inputs they invent, and sidestep unfamiliar target languages by prototyping in Python. Our results show that by-language token efficiency is a metric that should be considered when benchmarking and developing multilingual agents, and, for the tokenmaxxer, a guide to the most expensive language to work in. --- ### SLA-Constrained Carbon-Aware Routing in Geo-Distributed Serverless Clouds Authors: Anmol Chaudhary, Rahul Mishra Published: 2026-07-24 Categories: cs.LG, cs.DC arXiv: https://arxiv.org/abs/2607.22806v1 Modern cloud deployments distribute applications across multiple geographic regions, yet standard routing mechanisms prioritize latency while ignoring the fluctuating carbon intensity of local power grids. Latency-driven routing incurs avoidable carbon emissions, particularly when cleaner regions are within acceptable latency bounds. The proposed model formulates the carbon-aware serverless routing problem as a constrained optimization over geo-distributed cloud regions and introduces an SLA-constrained carbon-aware routing policy that achieves optimal carbon reduction within the SLA-feasible region, evaluated using real carbon intensity measurements across 5 primary AWS deployments. Experimental results show that the proposed policy achieves up to 46.8% carbon reduction while maintaining zero SLA violations across all evaluated thresholds. The system reduces carbon by an average of 27.4% under mixed workloads, and the routing overhead is very low (less than 0.02% of total request latency). A scalability study across 12 AWS regions spanning 6 continents demonstrates that average carbon savings increase from 27.4% to 47.5% as routing flexibility expands under mixed workloads. The proposed work contributes to SDG 13 (Climate Action) and SDG 7 (Affordable and Clean Energy) by enabling low-carbon routing decisions. These results indicate that cloud systems can achieve significant carbon savings without compromising user experience. --- ### OrchNAS: Orchestrated Neural Architecture Search Service for Personalised Federated Edge Intelligence Authors: Keya Patel, Sajib Mistry, Sheik Mohammad Mostakim Fattah, Aneesh Krishna Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22805v1 We propose OrchNAS, an energy-aware, personalised, federated edge intelligence framework that leverages a Neural Architecture Search Service to automatically design service-adaptive models for heterogeneous edge environments. The framework orchestrates the architecture search process on a server-side NAS service, enabling edge services to derive personalised architectures under device-level energy, computation, and memory constraints. We introduce an energy-aware global architecture search mechanism that learns a compact global representation across heterogeneous services. We develop an energy-efficient architecture selection mechanism that enables each service to derive a personalised subnet that satisfies its resource constraints via a progressive, greedy, energy-aware pruning strategy. We propose an energy-efficient personalised model optimisation scheme that updates service-adaptive parameters while preserving global representations, where a primal-dual optimisation mechanism enforces strict energy budgets during architecture adaptation. Experiments on real-world and benchmark datasets demonstrate the effectiveness of the proposed approach. --- ### LithoFormer: A Robust Framework for Stratigraphic Inference via Transformers Authors: Shwetha Salimath, Francesca Bugiotti, Sylvain Wlodarczyk, Sohaib Ouzineb Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22804v1 Accurate geological characterization of subsurface reservoirs from well log data is essential to support projects such as carbon capture and storage (CCS), geothermal development, and extraction of natural resources. Existing automated techniques for geological characterization primarily use sliding-window classification, which limits their ability to understand broader geological contexts, often leading to misaligned formation layers. To overcome these limitations, we introduce LithoFormer, a robust framework for stratigraphic inference using a Seq2Seq transformer model that ingests entire multivariate well logs in a single pass. The framework utilizes a channel-independent PatchTST backbone enhanced with rotary positional embeddings (RoPE) to capture long-range geological dependencies across entire multivariate well logs. A decoupled multi-task head is employed to jointly predict geological zonation and precise boundary probabilities, while a geology-informed loss function enforces physical constraints such as the Law of Superposition. Validated and deployed on three real-world datasets, LithoFormer demonstrates a 90% reduction in median boundary error and eliminates stratigraphic order violations compared to traditional sliding-window baselines. It also achieves a 80% reduction in manual expert labor and eliminates stratigraphic inconsistencies, providing a scalable and reliable solution for large-scale subsurface modeling. --- ### Learning Dense 2D-3D Correspondence for X-ray-to-CT Registration of Knee Bones Authors: Rembert Daems, Jonas Grammens, Caro Roten, Andrew Meyer, Thomas Luyckx, Matthias Verstraete Published: 2026-07-24 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2607.22803v1 Recovering the 6-DoF pose of the knee bones from a plain radiograph, given the patient's segmented pre-operative CT, turns a routine low-dose image into a quantitative measurement of joint geometry, without the added dose of a repeat CT or a fixed biplanar rig. Classic solutions align a rendered bone silhouette to image edges; recent alternatives refine pose by backpropagating an image-similarity loss through a differentiable X-ray renderer. Both operate one patient at a time and are fragile under a single view. Silhouettes are depth-ambiguous, and differentiable-rendering refinement has a narrow capture range at substantial per-iteration cost. We instead learn an amortized, subject-agnostic dense 2D-3D correspondence, supervised solely by projection geometry. One shared-weight model per bone, trained across 758 patients, registers patients unseen during training. The pose then follows in closed form from a global, initialization-free, render-free PnP+RANSAC solve. Because X-ray formation is transmissive, our correspondence target is transmission-aware rather than tied to a single surface. Though trained only to register, the representation is anatomically semantic: a simple classifier reads a landmark's anatomical region from its embedding across held-out patients, and the same features separate the knee's bones into a 2D-3D-consistent identity learned without any bone label. On a large single-institution cohort the model generalizes well to held-out patients. --- ### DY-LUT: Depth-Aware YCbCr Lookup Tables for Real-Time Underwater Image Enhancement Authors: Cunhao Zhu, Xiangtao Kong, Dongliang Xu, Zhiheng Zhang, Tianyu Wang, Yue Yao Published: 2026-07-24 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2607.22801v1 Underwater image enhancement is challenged by spatially non-uniform, wavelength-dependent attenuation. Propagation distance and wavelength govern this degradation, while YCbCr separates luminance from chrominance for restoration. We propose DY-LUT, a depth-aware YCbCr lookup-table framework for real-time enhancement. A dual-branch encoder predicts image-level fusion weights and a joint pair of pixel-wise degradation indices from image and depth features. These quantities condition learnable 4D LUTs, followed by lightweight local refinement. DY-LUT preserves traditional LUT efficiency while enabling depth-conditioned, spatially adaptive restoration. With externally supplied depth, its 3.56M-parameter enhancement network achieves competitive quality on UIEB-90 and LSUI and runs $9$--$304\times$ faster than representative high-capacity baselines. Adaptive inference further maintains real-time performance ($\sim7$ ms) for 4K UIQAD images. DY-LUT also benefits downstream detection and feature matching. Ablations show that YCbCr is a more effective basis than RGB for depth-conditioned lookup, while the jointly learned indices further improve adaptive querying. These results provide a physically grounded route to efficient UIE on practical platforms. --- ### StateAct: Program State, before Pixels, for Long-Horizon Computer-Use Agents Authors: Yan Yang, Xiangru Jian, Ziyang Luo, Zirui Zhao, Yutong Dai, Ziji Shi, Hanshu Yan, Jun Hao Liew, Silvio Savarese, Junnan Li Published: 2026-07-24 Categories: cs.SE, cs.CV arXiv: https://arxiv.org/abs/2607.22798v1 Computer-use agents are usually improved by strengthening perception: better models for reading a screenshot and choosing where to click. Yet a screenshot is only a lossy rendering of the underlying program state, e.g., the files, application backends, and DOM that hold the task data. Different states can produce the same pixels, while code can inspect and modify that state directly. StateAct is a code-first, multi-agent harness built around this distinction. Its main agent works directly with program state by using code, while a dedicated GUI subagent handles screenshot-and-click interaction on the few subgoals that need it, just 28 of 108 tasks and 1.1% of main-agent steps. The same direct access to program state also supports verification: an independent finish gate double-checks the saved result for structural failures, e.g., output that is missing, unsaved, or written to the wrong path. To stay on track over hundreds of steps, the main agent hands subgoals to fresh subagents, keeping its own context focused. On OSWorld 2.0, StateAct lifts Claude Opus 4.8 from 20.6% to 26.9% on binary success, and from 54.8% to 61.6% on partial success, at ~ 9x lower cost per task than the same model driven by screenshots alone; a code-only variant with no GUI subagent reaches only 45.9% partial, below that screenshot-based baseline's 54.8%. In general, grounding action, verification, and memory in state, what we call state-grounding, shifts the main bottleneck from perception toward reasoning: failures depend more on what the agent thinks than on what it sees. --- ### A Multi-level Information Integration Framework for Physically Verifiable Fault Diagnosis of Rotating Machinery Authors: Yuntong Chen, Jianyu Liu, Yingqi Li, Guobin Zhao, Ziang Wang, Chao Chen, Xitian Tian, Lijiang Huang Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22797v2 Integrating multi-level information, from physical models through data-driven diagnostics to natural language reasoning, into verifiable decision chains is a growing need in intelligent manufacturing. In bearing fault diagnosis, taken here as a representative testbed, the standard output is a class label and a confidence score derived from the classifier's own distribution, offering limited means of comparison against independent physical knowledge. Meanwhile, language models increasingly used for maintenance communication may introduce unsupported content. This work addresses both limitations from the output side. The proposed Diagnostic Evidence Network (DENet) is an encoder-agnostic multi-task framework that extends the output to a structured evidence record: the classification, a predicted characteristic frequency comparable against the theoretical value determined by bearing geometry and shaft speed, and a temporal localization of transient impulses inspectable on the raw waveform. Across four encoders and three public datasets, this evidence incurs no statistically significant accuracy cost, with a frequency error of about 6 Hz on 1,024-point segments. The deviation between predicted and theoretical frequency constitutes a label-free, inference-time validation signal. It detects misclassifications with AUROC of 0.970 and 0.871, and retains separation within the high-confidence subset. Finally, a QLoRA-adapted language model renders DENet's evidence into traceable maintenance reports without contributing diagnostic decisions, reducing unsupported-claim rates from 10-12% to 2% with no fabricated quantities observed. --- ### Physically Verifiable Evidence and LLM-Based Reporting for Bearing Fault Diagnosis Authors: Yuntong Chen, Jianyu Liu, Guobin Zhao, Ziang Wang, Chao Chen, Ju Huang, Xitian Tian, Lijiang Huang Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22797v1 Trustworthy deployment of AI-based diagnosis in safety-critical mechanical systems hinges on validation: whether a prediction can be checked against physical reality before it is acted upon. Current intelligent fault diagnosers fail this standard in two ways. Their standard output, a class label with a softmax confidence score, is an internal statistic of the classifier, offering nothing checkable against independent physical knowledge; and the growing use of generative language models in maintenance reporting adds a second risk: hallucinated content entering reports on which decisions rest. Taking bearing fault diagnosis as the testbed, this work addresses both problems from the output side. The proposed Diagnostic Evidence Network (DENet) is an encoder-agnostic multi-task framework extending the output to a structured evidence record: the classification, a predicted characteristic frequency comparable against the theoretical value determined by bearing geometry and shaft speed, and a temporal localization of transient impulses inspectable on the raw waveform. Across four encoders and three public datasets, this evidence incurs no statistically significant accuracy cost, with a frequency error of about 6 Hz on 1,024-point segments where spectral estimation is structurally inapplicable. Centrally, the deviation between predicted and theoretical frequency constitutes a label-free, inference-time validation signal: it detects misclassifications with AUROC values of 0.970 and 0.871, and remains discriminative in the high-confidence regime where confidence-derived detectors are blind. Finally, a QLoRA-adapted language model is constrained to translate, but never generate, diagnostic content, reducing unsupported-claim rates from 10-12% to 2% and eliminating fabricated quantities. --- ### Multimodal Domain Generalization for Depression Detection: An Attention-Based BiLSTM Network with Domain-Adversarial Training Authors: Ali Tabaraei, Federico Simonetta, Stavros Ntalampiras Published: 2026-07-24 Categories: cs.LG, cs.AI, cs.CL, cs.SD arXiv: https://arxiv.org/abs/2607.22794v1 Automatic depression detection with deep learning has shown promise but often suffers from limited generalization due to domain shift arising from inter-speaker variability. To address this critical issue, we present the first patient-independent multimodal depression detection framework that incorporates domain generalization (DG), jointly leveraging both acoustic and textual modalities. The proposed model integrates bidirectional Long Short-Term Memory (BiLSTM) with intra- and cross-modal attention mechanisms, accompanied by segment-level fusion for decision-making. Generalization is further enhanced by applying a gradient reversal layer inspired by Domain-Adversarial Training of Neural Networks (DANN), which promotes domain-invariant representations by adversarially limiting the model's ability to identify individual speakers, effectively reducing patient-specific bias. Conducting experiments on the Androids-Corpus dataset with a 5-fold cross-validation (CV) protocol, various pairings of audio and text feature extractors were evaluated over different segment durations, determining MelSpec and ItalianBERT as the optimal baseline at a 30-second segment duration. The addition of DG to this baseline yields a 2.5% increase in accuracy and 3.3% in F1-score, achieving 93.2% accuracy, 93.2% precision, 96.2% recall, and 94.2% F1-score, surpassing all existing benchmarks. Extensive ablation studies assess the impact of multimodal fusion, deep architectural choices, and DG, highlighting their combined contribution to robust and generalizable depression detection. --- ### Small, Bias-Free, Blind and Convolutional Denoiser: A compact ConvNeXt U-Net for blind Gaussian color-image denoising Authors: Nikolas Markou Published: 2026-07-24 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2607.22793v1 We describe and evaluate BF-ConvUNeXt, a compact bias-free ConvNeXt U-Net for blind additive-white-Gaussian-noise color image denoising, combining four existing ingredients so a single property survives end to end: a frozen depthwise Gabor stem (oriented band-pass, zero trainable parameters), a Laplacian-pyramid encoder routing the high-frequency residual into each skip connection, a ConvNeXt-V1 U-Net body, and bias-free construction throughout (no additive bias, linear head, LeakyReLU, variance-only batch norm). Together these make the 0.82M-parameter network exactly degree-1 homogeneous at inference, D(alpha y) = alpha D(y), licensing a Miyasawa/Tweedie score reading of the residual and blind generalization across noise levels from one model. We train a single blind model on a noise-sigma curriculum (sigma approximately 6.4 to 64, 0-255 scale); it extrapolates past that ceiling without a cliff, degrading smoothly to 22.8 dB at sigma=150 and 20.0 dB at sigma=200. Evaluated unchanged on four standard color sets (CBSD68, Kodak24, McMaster, Urban100) at sigma in {15,25,50}, it matches or beats DnCNN and FFDNet on every set and level, averaging about +0.7 dB over DnCNN. Against heavyweight CNN/transformer state of the art it trails by a small margin (roughly 0.3-1.7 dB depending on set) at 1/15 to 1/39 of their parameters. The homogeneity is inference-only and checkpoint-specific, and the learned residual is a local, not global, score (non-conservative Jacobian), so plug-and-play/RED guarantees do not transfer; it still drives stochastic sampling and linear inverse problems (inpainting, super-resolution, deblurring, compressive sensing). --- ### Learning-based Hierarchical Tracheal Anatomy Understanding from Sparse Surgical Demonstration Annotations for Ultrasound Robots Authors: Hiu Ching Cheung, Wenchao Yue, Zhengran Han, Mingcong Chen, Guanglin Cao, Hongbin Liu, Hongliang Ren Published: 2026-07-24 Categories: eess.IV, cs.CV, cs.RO arXiv: https://arxiv.org/abs/2607.22789v1 Tracheostomy requires precise localization of the tracheal incision site; however, conventional manual palpation is subjective and often unreliable, while ultrasound utility remains operator-dependent. This work presents a learning-based framework for hierarchical tracheal anatomy understanding, designed specifically for ultrasound-guided robotic systems. We propose a two-stage perception pipeline integrating a YOLOv8n localization backbone with a sparse, prompt-optimized SAM2 decoder to achieve high-fidelity segmentation from sparse surgical annotations. Our hybrid training strategy, bridging curated laboratory data with unconstrained sequences, ensures clinical robustness. Experimental benchmarks demonstrate that this decoupled architecture effectively balances generalization, precision, and efficiency. The YOLOv8n and SAM2 framework achieves a consistent Mean Dice Similarity Coefficient (DSC) of 0.777 across both controlled and generalized domains. This significantly outperforms U-Net baselines, which often suffer from anatomical fragmentation and performance degradation (Generalization DSC $\le$ 0.494). By constraining mask decoding to targeted, sparse regions of interest, our model achieves a throughput of 6.92 FPS, which is vital for closed-loop robotic teleoperation. This study confirms that a robust hierarchical understanding of tracheal anatomy can be derived by coupling lightweight localization with foundation-scale visual models. Our framework establishes a scalable foundation for standardized, autonomous surgical assistance, effectively navigating the variability of real-world ultrasound to enhance the safety and precision of robotic-assisted tracheostomy. --- ### FMOPF: Latent Flow Matching with Constraint-Aware Interaction Priors for AC Optimal Power Flow Authors: Zhilin Huang Published: 2026-07-24 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2607.22788v1 AC optimal power flow determines the minimum-cost generation dispatch under nonlinear power balance constraints and is solved thousands of times daily in electricity market operations. Learning a direct mapping from load conditions to OPF solutions can accelerate this computation, yet with deepening renewable penetration, a single optimal dispatch is no longer sufficient. Operators require a characterization of the distribution of feasible near-optimal solutions for risk quantification, sensitivity analysis, and multi-objective trade-off assessment. Supervised neural networks provide fast point predictions but cannot capture this conditional distribution. Diffusion-based generative models can sample diverse solutions in principle, yet existing methods operating in the raw state space exhibit degraded solution quality and fail to scale beyond medium-sized systems. We identify the root cause as the conflation of two distinct tasks within a single model. Compressing the high-dimensional OPF solution manifold is one task, and learning the conditional mapping from loads to that manifold is another. This paper presents FMOPF, a framework that resolves this conflation by decoupling compression from generation through latent flow matching and by explicitly modeling load-state coupling through a Constraint-Aware Interaction Prior Network. Experiments on four IEEE test systems demonstrate that FMOPF provides the most effective Newton-Raphson warm starts, achieves the lowest tail risk among generative methods, and is the first such method to scale to systems with several hundred buses while preserving full feasibility. Ablation studies confirm that the latent generation pipeline is a necessary condition for physical feasibility and that the interaction prior functions as a late-stage tail-risk controller. --- ### Optimizing Transformer Neural Network for Real-Time Outlier Detection on FPGAs Authors: Ilia Sobakinskikh, Paul Alexander Bilokon Published: 2026-07-24 Categories: cs.LG, cs.AI, cs.AR, cs.DC, cs.PF, stat.CO arXiv: https://arxiv.org/abs/2607.22786v1 In this work, we explore how the inference time of a Transformer Neural Network can be efficiently optimized with applications to real-time anomaly detection in financial time series. The financial time series are price series such as asset prices. Unfortunately, the data is often with errors or outliers that make the downstream data processing tasks useless, unstable or even harmful. Moreover, the amount of financial time-series data has been significantly increasing. Hence, there is a need for better data-cleaning methods in terms of accuracy and in terms of processing speed. Transformers as a neural network architecture have achieved superior performances in many tasks such as Natural Language Processing and Computer Vision. Time series modelling and especially anomaly detection tasks can benefit from the features of transformers architecture in multiple ways, including the capacity to capture long-range dependencies and interactions. Increasingly powerful hardware, such as field-programmable gate arrays (FPGAs), have seen increasing usage in recent years due to their reconfigurability and high performance. They can be efficiently utilized to speed up the computations of the Transformer architecture. We explore different Transformer architectures for time series modelling and how they can be efficiently implemented on an FPGA board (PYNQ-Z2). In particular, we examine the application of Transformers to detect anomalies in time series and we show how they can be efficiently implemented on an FPGA board to minimize latency. The code is available at https://github.com/thxi/icl_thesis --- ### FusionML: Prefill, Not Decode - Mechanism and Boundaries of CPU+GPU Co-Execution on Unified-Memory Apple Silicon Authors: Om Mohite Published: 2026-07-24 Categories: cs.AR, cs.DC, cs.LG, cs.PF arXiv: https://arxiv.org/abs/2607.22785v1 Apple-Silicon SoCs share CPU, GPU, and Neural Engine over one unified memory system, raising the question of whether transformer inference can be accelerated by splitting single operators across units. Prior attempts, including our own, failed or produced precision-confounded wins. We identify the cause: MLX's lazy-graph scheduler \emph{serializes} cross-stream work whenever a CPU-stream operation consumes an unmaterialized GPU result inside one evaluation graph, so a row-split matmul that runs \x{1.38} faster with materialized inputs runs \x{0.66} slower than GPU-only inside a lazy graph; an eager materialization boundary restores concurrency (\x{1.34}). \sys{} implements a per-layer, contention-aware CPU+GPU row split for transformer prefill built on this fix. Evaluated across five chips and three Apple-Silicon generations, community-replicated, the split accelerates Llama-shaped decoder-block prefill by \x{1.15}--\x{1.38}, unchanged at full 32-block depth, and reaches \x{1.18}--\x{1.25} faster time-to-first-token on a real Qwen2.5-7B checkpoint served through stock MLX-LM, with token-identical outputs and unchanged decode throughput. We characterize the boundaries equally carefully: decode cannot benefit, bound by shared bandwidth co-execution does not add; precision-matched training loses \x{0.86}--\x{0.97} on all five chips; ANE dispatch overhead excludes it at layer granularity; and a no-regression runtime gate becomes self-defeating under memory pressure, where probing an alternative mode evicts the active mode's working set. Code, raw results, and generation transcripts are released. --- ### JPEG AIC2026: A large-scale dataset for fine-grained assessment of image coding Authors: Mohsen Jenadeleh, Jon Sneyers, João Ascenso, Thomas Richter, Alexander Karabutov, Panqi Jia, Elena Alshina, Osamu Watanabe, António Pinheiro, Touradj Ebrahimi, Dietmar Saupe Published: 2026-07-24 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2607.22783v1 Recent advances in conventional and learning-based image coding have increased the demand for benchmark datasets that support fine-grained assessment of compressed image quality, particularly for learning-based image compression methods. This paper introduces Assessment of Image Coding 2026 (AIC2026), a large-scale dataset for high-fidelity image compression containing 70 source images selected from 2,787 candidates using semantic clustering, inter-metric disagreement among objective image quality assessment (IQA) methods, and manual inspection and refinement. The dataset covers a wide range of compression artifacts produced by eight conventional and four learning-based codecs across 17 coding configurations. Each source image is encoded using seven codecs. For each source-codec pair, decoded images are provided at 20 perceptually spaced distortion levels, corresponding approximately to 0.2-4.0 just-noticeable difference (JND) units using the ColorVideoVDP (CVVDP) metric for distortion estimation, yielding 9,618 distorted images. This fine-grained sampling enables analysis of rate-distortion behavior and objective metric evaluation for subtle quality differences across a wide range of compression artifacts. We report an extensive objective analysis using 24 conventional and 12 learning-based IQA methods. The results show substantial disagreement among current IQA methods for fine-grained quality differences, particularly for artifacts introduced by learning-based codecs. The complete dataset is publicly available at https://doi.org/10.18419/DARUS-6156. --- ### What Softmax Throws Away: Mass-Aware Attention for Evidence Accumulation Authors: Minwoo Yu, Young-guk Ha Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22781v1 High task performance does not show whether a model retains prediction-relevant structural information in its internal representation. Temporal graph models, for example, can achieve high future-link AUC while basic graph statistics remain difficult to recover from the same representation. We identify one source of this gap in the weighted averaging used by standard attention: when an evidence pattern is repeated, the numerator and denominator grow at the same rate, so inputs with different amounts of accumulated evidence can produce the same aggregate. We propose Mass-Aware Attention (MAA), which generalizes standard L1 normalization to an Lp family. Under repetition, MAA makes the numerator and denominator scale at different rates, retaining the effective number of contributing inputs in the representation magnitude. It adds no supervision, parameters, hidden dimensions, or explicit count features, and recovers standard attention at p=1. Across four continuous-time dynamic graph models and three datasets, MAA improves future-link AUC in 11 of 12 model-dataset cells. Linear recovery from the same hidden representation increases by 4.49% on average, and preferential-attachment recovery improves in all 12 cells after family-wise correction. We also observe consistent evidence in marked temporal point processes, temporal knowledge graphs, retrieval-augmented generation, and spatio-temporal point processes. Information accessibility and task utility remain distinct: NLL improves in MTPP, ranking is largely preserved in TKG, additional information in RAG does not improve the diagnostic head, and downstream LayerNorm can erase the signal in STPP. These results position MAA as a general normalization principle for improving predictor-facing representation informativeness by controlling repetition invariance in standard attention. --- ### Inter-Reflective Gaussian Splatting for Robust and Efficient Inverse Rendering Authors: Chun Gu, Xiaofei Wei, Zixuan Zeng, Yuxuan Yao, Li Zhang Published: 2026-07-24 Categories: cs.GR, cs.CV arXiv: https://arxiv.org/abs/2607.22780v1 Faithful inverse rendering requires visibility and indirect radiance to explain secondary illumination and inter-reflection, yet rasterization-oriented Gaussian representations do not naturally support the secondary-ray queries needed to recover them. We present IRGS++ (Inter-Reflective Gaussian Splatting), a unified robust and efficient Gaussian inverse rendering framework. During transport-aware optimization, IRGS++ employs differentiable 2D Gaussian ray tracing on surface-oriented Gaussian primitives to query visibility and indirect radiance on the fly and evaluate the full rendering equation for inter-reflective transport. This physical core makes Gaussian inverse rendering physically grounded beyond rasterized appearance modeling. To make this backbone useful beyond low-gloss dielectric scenes, the framework incorporates metallic-aware material modeling and robust reflective initialization for glossy, specular, and metallic materials. To make it practical, multiple importance sampling and denoising stabilize finite-sample rendering, while mesh-based secondary-attribute queries reduce the cost of relighting under novel illumination. Quantitative evaluations on low-gloss and glossy benchmarks show improved decomposition and relighting quality together with favorable quality--speed trade-offs under the reported configurations, while real-world studies illustrate plausible relighting under novel illumination. --- ### Multimodal Surface EMG Hand Gesture Recognition Using Query-Based Transformers for Prosthetic Control Authors: Federico Del Pup, Elisa Tentori, Manfredo Atzori Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22779v1 Hand gesture recognition via surface electromyography (sEMG) is fundamental to prosthetic control. In this field, deep learning approaches have become the gold standard. However, current architectures struggle to scale; model performance typically decreases as the number of hand movements increases. Performance degradation is tied to the increased statistical complexity of decoding expanded gesture sets and compounded by the limitations of state-of-the-art methods, which primarily rely on low-latency unimodal convolutional architectures. Convolutions operate locally, limiting model's ability to capture long-range sequential patterns. Unimodal setups cannot leverage complementary information from coordinated signals characterizing movement execution, such as inertial and eye-tracking data. These limitations motivate architectures that integrate local and global features across multimodal physiological sequences. To bridge this gap, this study introduces EMG-CrossFormer, an end-to-end hybrid convolutional-transformer for seamless multimodal integration. EMG-CrossFormer combines representations from an arbitrary number of unimodal encoders through cascaded cross-attention fusion layers, and decodes the fused representations using learnable gesture queries. EMG-CrossFormer was evaluated on four NinaPro datasets (DB2, DB3, DB7, and DB10) and benchmarked against six state-of-the-art models using an increasing number of modalities. Using only sEMG, EMG-CrossFormer achieved mean accuracies of 72.33%, 52.48%, 79.16%, and 73.49% on DB2, DB3, DB7, and DB10, respectively. Incorporating inertial signals improved performance to 90.66%, 80.40%, 92.79%, and 92.06%. These results show that joint local-global feature modeling improves sEMG-only decoding and that multimodal fusion substantially amplifies this benefit, underscoring the value of both design principles for complex hand gesture recognition. --- ### Subject-Level Heterogeneity in EEG Motor Imagery Decoding: A Large-Scale Benchmark and Portfolio-Based Reduction of the Search Space Authors: Xavier Vasques, Paul Barbaste, Olivier Oullier Published: 2026-07-24 Categories: q-bio.NC, cs.LG arXiv: https://arxiv.org/abs/2607.22778v2 Robust EEG motor imagery decoding remains limited by strong inter-individual variability, making it difficult to identify pipelines that generalize across users. We present a large-scale, standardized within-session benchmark of decoding pipelines across three public datasets: Cho2017 (52 subjects), PhysionetMI (109 subjects), and Zhou2016 (4 subjects). Using a common MOABB LeftRightImagery setting, two frequency bands (8-15 Hz and 8-30 Hz), and a broad combination of feature extraction, preprocessing, and classification steps, we analyzed 216,714 raw evaluation rows, which after structured aggregation yielded 44,928, 109,000, and 4,192 subject-level observations respectively. Covariance tangent-space projection (cov-tgsp) and Common Spatial Patterns (CSP) consistently defined the strongest methodological families, though their relative ordering was dataset-dependent. On Cho2017, the best family-level mean accuracy came from cov-tgsp in 8-30 Hz (0.712 +/- 0.140), whereas Zhou2016 favored CSP (0.832 +/- 0.121 in 8-15 Hz). These aggregate rankings concealed substantial subject-level heterogeneity: 42 distinct winning pipelines across 52 Cho2017 subjects, and 93 across 109 PhysionetMI subjects. We then used the benchmark as an empirical performance landscape for building compact portfolios of pipelines of size K. Several construction procedures were compared, including a ranking-based Top-K Mean heuristic and search-based strategies. Results were broadly consistent, with Top-K Mean giving the best trade-off. A single best global pipeline already retained 94.2% of the oracle in Cho2017 and 81.8% in PhysionetMI; at K = 12, oracle retention rose to 96.5% and 90.0%. The landscape is therefore subject-dependent, and this heterogeneity can be exploited through compact portfolios that make personalization more feasible. --- ### LC-SEPLM: long-range contact-supervised adaptation for sequence-only protein representation learning Authors: Chen Wang, Boming Kang, Qinghua Cui Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22777v2 Protein language models learn transferable sequence representations. However, because they primarily model contextual dependencies along amino-acid sequences, their training objectives do not explicitly constrain the model to learn three-dimensional residue contacts formed after folding . Here, we introduce LC-SEPLM (Long-range Contact-supervised ESM Protein Language Model), which adapts ESM2 with LoRA and long-range residue-pair contact supervision while retaining sequence-only downstream inference. Pair-specific queries use cross-attention over the complete sequence to extract global sequence context associated with long-range spatial contacts. To expose the model to diverse structural information, we trained LC-SEPLM on 500,000 AlphaFold Swiss-Prot proteins. In downstream evaluation, LC-SEPLM improved all eight protein-level tasks relative to ESM2. The largest gain occurred in remote-homology recognition, where macro-F1 increased from 0.6122 to 0.6769 (+0.0647, or 6.47 percentage points). On the official ESM-S EC benchmark, LC-SEPLM also outperformed ESM-S with a maximum absolute gain of 0.1771. These results support residue-pair contact supervision as a bounded route for introducing structural information into protein sequence representations while preserving sequence-only inference. --- ### Predicting the Outcome of rTMS Depression Therapy using EEG Signals and CNN Authors: Wael Korani, Md Fahimul Kabir Chowdhury, Sadam AlQadi, Priyan Malarvizhi kumar, Reza Rostami, Reza Kazemi Published: 2026-07-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.22776v1 Repetitive transcranial magnetic stimulation (rTMS) is a non invasive therapy for Major Depressive Disorder (MDD). In this study, we generate images using two time frequency methods to represent EEG signals: Fourier-Bessel Series Expansion with Euclidean Distance (FBSE-ED) and Discrete Wavelet Transform (DWT). We propose an efficient deep learning classifier to predict the outcome of rTMS depression therapy. In this study, we use a private rTMS databases to train a lightweight custom Convolutional Neural Network (CNN) using 10-fold cross validation strategy in order to avoid any bias in our results. The results show that the FBSE-ED representation achieves the highest classification accuracy of 93.60\%, outperforming traditional time-frequency technique (DWT). In addition, the proposed architecture with FBSE-ED image representation technique outperforms more complex EEG-Specific deep learning models (EEGNet, DeepConvNet, SleepEEGNet) by 3.62-10.72% and pretrained models (Xception, DenseNet201, and MobileNetV2) by 23.03-27.35%. For more experiments, we utilize another private rTMS database as test database to show the robustness of the proposed model. Our results suggest that integrating advanced signal decomposition with deep learning can facilitate early prediction of rTMS treatment response and support more targeted clinical decision-making. The proposed framework is interpretable, computationally efficient, and well-suited for deployment in real-world local psychiatric clinics. --- ### CC-AOS: Cost- and Horizon-Conditioned Amortized Backward Induction for Finite-Horizon Optimal Stopping Authors: Tianwei Yu Published: 2026-07-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.22774v1 Finite-horizon optimal stopping is a central problem in early time-series classification, where a system must decide at each sequence prefix whether the expected benefit of another observation justifies its acquisition cost. Existing data-driven backward-induction methods typically solve each cost-horizon operating point separately, so changing operating conditions requires repeated optimization and separate model stacks, making continuous cost adaptation and multi-horizon deployment inefficient. We propose CC-AOS (Cost- and Horizon-Conditioned Amortized Optimal Stopping), a structured amortized solver for a family of finite-horizon stopping problems with continuous costs and multiple horizons. CC-AOS learns a shared continuation-value model conditioned on the current state, absolute time, remaining horizon, and acquisition cost through joint amortized fitted backward induction. We establish that the exact value and continuation functions are nondecreasing, concave, and horizon-dependently Lipschitz in cost, encode these properties in the model architecture, and derive residual-based bounds on value and policy errors. Experiments on controlled Gaussian and time-varying non-Gaussian processes and the FordA engine-noise time-series benchmark compare CC-AOS with representative per-operating-point backward-induction solvers and tuned static stopping rules. At six unseen FordA cost-horizon pairs, one CC-AOS checkpoint achieved a lower terminal-risk-plus-sampling-cost objective than independently fitted Convex Function Learning at all six pairs, with an average reduction of 15.75 percent, while matching the tuned static thresholds on average. --- ### Metric Surface Reconstruction of Neurosurgical Scenes from Monocular Operating Microscope Images and Microscope Pose Authors: Thomas Bucher, Didier Neuenschwander, Thomas Petutschnigg, Michael Murek, David Bervini, Andreas Raabe, Manuela Eugster Published: 2026-07-24 Categories: eess.IV, cs.CV, physics.med-ph arXiv: https://arxiv.org/abs/2607.22773v1 Objective: We evaluated whether metric 3D geometry of neurosurgical operative exposure can be recovered from standard monocular operating-microscope images combined with microscope pose data. Methods: In a phantom-based laboratory study, two aneurysm training phantoms were imaged with a ZEISS Pentero 800 microscope integrated with Brainlab Cranial Navigation. Microscope images from the standard composite video output were stored with synchronous microscope poses. After intrinsic and extrinsic calibration, depth was estimated with the pretrained Depth Anything 3 model without task-specific fine-tuning. Fused point clouds were converted to meshes using Poisson surface reconstruction. Reconstructions were compared with reference surfaces from structured-light scanning and fine-slice CT. Results: For phantom A, representing a deeper surgical corridor, reconstruction accuracy ranged from 1.95 $\pm$ 1.70 mm to 2.33 $\pm$ 2.15 mm. For phantom B, representing a directly exposed surface, accuracy ranged from 1.02 $\pm$ 0.93 mm to 1.52 $\pm$ 1.21 mm. Larger image sets mainly improved completeness, while accuracy remained within a narrower range. Corridor analysis showed preservation of overall geometry with local deviations in incompletely reconstructed regions. Conclusions: Standard monocular microscope images combined with navigation-derived pose data can reconstruct millimeter-range 3D surfaces using a foundation-model-based pipeline. These results show technical feasibility in a controlled phantom setting and support further development toward objective quantification of operative exposure, image fusion, and characterization of working spaces for future surgical instrumentation. --- ### Generative Video Compression with Adaptive Score Distillation Authors: Naifu Xue, Zhaoyang Jia, Haosen Li, Zihan Zheng, Jiahao Li, Bin Li, Xiaoyi Zhang, Qi Meng, Yuan Zhang, Yan Lu Published: 2026-07-24 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2607.22772v1 Diffusion models provide strong generative capabilities for video compression at ultra-low bitrates. Existing diffusion-based video codecs adapt base models originally developed for text-conditioned generation, whereas diffusion models designed and trained specifically for compression remain unexplored. To fill this gap, we introduce our Generative Video Codec (GenVC), built on a video diffusion model trained from scratch for compression. To our knowledge, this is the first compression-oriented video diffusion model. We realize this model directly in pixel space with a global-to-local hierarchy that recovers fine spatio-temporal details, enabling high-quality generative reconstruction from compressed representations. To accelerate inference, we distill the multi-step model into one step using distribution matching distillation (DMD). Applying DMD directly, however, drives the student toward motion-stalled reconstructions. We trace this to a teacher-side guidance failure: once student-induced perturbations leave the frozen teacher's training region, its guidance can become misleading, causing DMD updates to reinforce rather than correct the student drift. To break the resulting feedback loop, we propose Adaptive Score Distillation, which gates DMD updates according to their alignment with the ground-truth direction, enabling high-quality reconstruction with coherent motion. Experimental results show that GenVC achieves state-of-the-art perceptual quality at ultra-low bitrates, with average bitrate savings of 62.5% at matched LPIPS and 71.3% at matched FID over GLVC. Unlike prior codecs that inherit billion-scale pretrained backbones, our diffusion model has only 478.0M parameters and decodes 1080p video in a single step at 15.1 fps on an A100 GPU. --- ### When Do Cheap Probes Predict Expensive Training? Probing 3D-CT Encoders for Text Generation Authors: Renjie Liang, Zijian Xu, Jinqian Pan, Chengkun Sun, Zhengkang Fan, Shawn Li, You Qin, Mei Liu, Jie Xu Published: 2026-07-24 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.22771v3 Building a 3D CT vision language model begins with a choice of which image encoder to build on. Today that choice is made by fine-tuning every candidate through the full language model and comparing downstream scores, an enormously expensive search. A cheap probe on the encoder's representation promises a way out, but whether it forecasts the expensive outcome has never been tested. We test this with CheapCT on report generation and on MeasureVQA, a new VQA dataset we build. MeasureVQA scores the outcome one capability at a time, its answers measured from segmentation masks and Hounsfield units. Report generation scores the whole report at once and reflects mostly disease. CheapCT forecasts expensive training across every capability. The forecast survives changing the probe readout and the language-model backbone. The rank agreement between CheapCT and fine-tuning stays high throughout, from rho = 0.90 to 0.97. Used to choose an encoder, CheapCT picks one nearly as good as the best while fine-tuning a single candidate, at orders of magnitude less compute. We release the code and MeasureVQA at https://github.com/renjie-liang/CheapCT. --- ### Cheap Probes Predict Expensive Training in 3D-CT Vision--Language Models Authors: Renjie Liang Published: 2026-07-24 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.22771v1 Picking the frozen image encoder for a 3D~CT vision--language model (VLM), together with the token-compression scheme on top of it, is a search over many candidates. There are several encoders, several ways to compress their tokens, and several token budgets, and the combinations grow fast. Comparing them the usual way means fine-tuning a large language model (LLM) on each combination, and running the whole sweep this way needs far more compute than most groups can spend. We ask whether a cheap probe on the encoder's cached embeddings can stand in for that comparison. We build an image-grounded probing benchmark over (encoder $\times$ compression) cells, with clinical attribute families and two validation gates, scale-sanity and probe-separability, that keep each attribute well-scaled and decodable. These gates are the main methodological contribution. On this benchmark we compare a range of read-out heads, and in a preliminary study we pair each probe with its matched downstream task. The early signal is encouraging: the cheap probe orders the candidates in close agreement with expensive fine-tuning, at about $r\approx0.95$ on the cells measured so far. We read this as an ordinal claim, a ranking predictor rather than an exact estimate, and we are explicit about where it stays preliminary. If it holds up, encoder and compression choices can be screened in minutes with frozen-token probes, with full training spent only on the finalists. --- ### Dementia Etiology Diagnosis via Collaborative Meta Knowledge Enhancement Authors: Siyuan Du, Mengxi Chen, Xinyang Jiang, Zilong Wang, Jiangchao Yao, Dongsheng Li, Ya Zhang, Lili Qiu, Yanfeng Wang Published: 2026-07-24 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2607.22770v1 Although artificial intelligence (AI) has shown promising performance in several medical tasks, accurate dementia etiology diagnosis with AI remains challenging due to complex overlapping symptoms among diseases. Scaling up the dataset size by combining the cross-center samples may bring a gain in the pursuit of performance, while the inherent data heterogeneity across centers or populations induces the conflict. Conventional multi-task learning paradigms offer a promising framework; however, they fail to consider critical meta information (e.g., site-specific acquisition and modality availability) to combat the heterogeneity. To address this challenge, we propose a Collaborative Meta Knowledge Enhancement (COME) framework for dementia etiology diagnosis, which injects multi-center acquisition semantics, source identifiers, and modality indicators as heterogeneity-aware embeddings into a unified Transformer architecture for scale-up training, enabling explicit modeling of heterogeneity. Besides, a trust-region constrained optimization scheme is designed to regularize the model from spurious correlations during training through a reference model. Across seven independent cohorts, our method achieves state-of-the-art in-domain performance with a mean macro-averaged AUC of 85.62% and a 4.29-point gain over the strongest baseline, while maintaining superior out-of-domain generalization under both cross-center and cross-sequence evaluations. Extensive validation also confirms the alignment between model predictions and established biomarkers (amyloid, tau) and clinical severity, highlighting the potential of COME to enable robust and interpretable dementia diagnostics in real-world settings. --- ### DomainPilot: Domain-Level Loss-Guided Two-Stage Data Mixture Optimization for Efficient Language Model Fine-Tuning Authors: He Zhang Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22769v1 The training efficacy of large language models (LLMs) is fundamentally constrained by the quality and composition of training data. Existing dynamic data scheduling methods face critical limitations in industrial-scale pretraining and supervised fine-tuning (SFT): data selection incurs prohibitive O(N) costs on terabyte-scale corpora, mixture optimization schemes introduce severe I/O bottlenecks or require training auxiliary reference models, and sample-level reweighting strategies rely on loss signals that conflate noise, difficulty, and novelty. We present DomainPilot, a domain-level loss-guided two-stage data mixture optimization framework. DomainPilot introduces token-level domain loss monitoring to capture per-domain learning dynamics during training without halting the data pipeline. Building on these signals, we propose a Scaling Law guided coarse optimization stage that fits domain-specific convergence curves and derives a principled prior for mixture adjustment. A subsequent Mixing Law guided fine optimization stage refines the mixture by modeling cross-domain interaction effects through controlled sweep experiments. The entire mechanism is realized via a patch-based architecture that injects domain-aware loss computation into existing training frameworks (e.g., MindSpeed/Megatron-LM) with only ~30 lines of framework-specific adapter code. We validate DomainPilot on the Qwen3-1.7B model during SFT. Compared to the original data mixture, our optimized mixture achieves improvements of +2% on MMLU-Redux, +1.8% on AIME24, +3.8% on LiveCodeBench v5, and +3.6% on BFCL v3, without increasing total data volume or training cost. These results demonstrate that domain-level training signals provide an effective, lightweight alternative to expensive data selection or auxiliary model training for mixture optimization. --- ### Beyond Shapley: An Influence-Based Data Auditing Pipeline for LLM Alignment and Evaluation Authors: Yunting Song, Matthew Watson, Peter Grabowski, Jun Qin Published: 2026-07-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.22766v1 The alignment of Large Language Models (LLMs) is increasingly bottlenecked by data quality. As datasets scale, massive preference and instruction-tuning corpora inevitably accumulate hidden structural contradictions, safety risks, and systemic human annotation errors. Standard dataset auditing methods, such as semantic deduplication or LLM-as-a-judge, struggle to capture the actual predictive impact of individual records and often miss deep functional rule clashes. To address this, we introduce a scalable, inference-only data valuation pipeline that approximates the Shapley value without iterative model retraining. By mapping semantic k-NN neighborhoods into a directed graph, our framework evaluates data utility directly through a reference LLM's probability distribution using zero-shot and one-shot conditional log-likelihood shifts. Our pipeline then translates these predictive influence scores into localized advantage metrics to isolate gradient-conflicting records. We demonstrate the pipeline's efficacy in sanitizing two heavily vetted alignment datasets. First, applying our pipeline to the HelpSteer2 dataset reduced the manual audit search space by 99.1%, successfully uncovering falsely-labeled records across diverse failure modes. Second, applying our automated audit strategy to Anthropic's HH-RLHF training and evaluation splits identified thousands of hidden safety and factual preference inversions. Crucially, by extending this audit to the evaluation split, we expose severe vulnerabilities in current benchmark integrity: highly capable models frequently predict the safer or more helpful response, only to be penalized by objectively flawed human ground-truth labels. Overall, our work provides a mathematically grounded, highly efficient diagnostic tool to uncover human label failures, sanitize evaluation benchmarks, and ensure the integrity of LLM alignment data. --- ### Frequency-Aware Dual-Stream Learning for Balanced Realism and Fidelity in Electron Microscopy Imaging Authors: Longmi Gao, Zhengkai Zhao, Pan Gao, Manoranjan Paul Published: 2026-07-24 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2607.22765v1 Electron microscopy enables nanoscale cellular visualization but faces a trade-off between imaging resolution and acquisition speed. Existing learning-based methods rely on single-stream architectures that struggle to balance perceptual realism and quantitative fidelity, either over-smoothing details or generating unrealistic hallucinations. This work introduces a frequency-adaptive dual-stream architecture to resolve this conflict. Using discrete wavelet transform, we decompose images into low-frequency structures and high-frequency details, then employ a conditional diffusion model for realistic global synthesis and a transformer network for precise detail recovery. Experiments on the EMDiffuse dataset show the method achieves superior LPIPS and resolution ratio, substantially outperforming existing approaches. The method also shows strong generalization across diverse biological samples, supporting fast and reliable electron microscopy imaging for structural biology and nanotechnology applications. The source code and associated dataset are publicly available to facilitate further research. --- ### Robot-Factored World Models via Robot Rendering Authors: Byungjun Kim, Taeksoo Kim, Hyunsoo Cha, Hanbyul Joo Published: 2026-07-24 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2607.22535v1 Action-conditioned video world models predict future observations from an initial observation and an action signal. In robotics, actions influence future observations through two distinct processes: they are first realized into robot motion by the robot body and controller, and the scene then responds through contact and object motion. Conditioning directly on action commands asks the world model to learn the realization process itself, while conditioning on logged future states leaks the interaction outcomes it is meant to predict. We propose robot-factored world models, which move two robot-specific factors outside the world model. First, action realization: each command is rolled through the robot's own controller and kinematics into a deployment-available nominal trajectory, a middle signal that avoids both action-realization learning and future-state leakage. Second, robot rendering: this nominal trajectory is rendered through the robot URDF, factoring the robot's geometry, kinematics, and appearance out of the model and into explicit rendered robot geometry. To resolve depth ambiguity, we pair end-effector depth with scene depth, giving geometric cues for contact and occlusion beyond image-plane overlap. Together, camera-aware static RGB/depth context and rendered robot geometry form a shared visual world-model interface that stays consistent across viewpoints and robot embodiments, so the model sees the action only as visible robot geometry and learns how objects respond to it. Our experiments show that the rendered interface outperforms vector-conditioned baselines and generalizes to unseen robot embodiments at inference. We further demonstrate that our model generates robot manipulation videos from human demonstrations by retargeting and rendering the hand motion as robot geometry. --- ### SM4RT: Learning Structured Motion Geometry for 4D Reconstruction Authors: Shing Ho J. Lin, Wenzhao Zheng, Dong Zhuo, Yuqi Wu, Jie Zhou, Jiwen Lu Published: 2026-07-24 Categories: cs.CV, cs.AI, cs.RO arXiv: https://arxiv.org/abs/2607.22534v1 Geometry Foundation Models (GFMs) have substantially advanced monocular 3D reconstruction, yet extending this capability to 4D dynamic understanding remains a fundamental challenge. Most existing motion perception methods (e.g., sparse tracking, dense point-wise flow) treat motion as independent point-wise displacements, ignoring the structured nature of physical motion. However, real-world objects usually obey rigid-body kinematics, and points thus usually move collectively, not in isolation. Motion itself possesses geometric structure: physical objects undergo a set of rigid-body transformations governed by SE(3), rather than unstructured point-wise displacements. Building on this insight, we propose SM4RT, a Structured Motion 4D Reconstruction Transformer for end-to-end 3D reconstruction and structured motion perception. SM4RT introduces Structure-of-Motion to represent scene dynamics, where scene motion is decomposed into a compact set of motion bases, each represented as a temporal sequence of 6D twists in SE(3). Dense scene motion is then recovered by sparse, time-shared per-pixel assignment weights over these bases, ensuring points on the same object share a common rigid-body motion trajectory. SM4RT introduces a parallel motion geometry encoder and decoder that jointly infer 3D geometry, world-coordinate motion, and scene kinematic structure in a single forward pass from monocular RGB video. SM4RT achieves strong motion reconstruction performance while preserving the geometric structure of scene motion. --- ### Twins: Learn to Predict Unified Representations with Focal Loss Authors: Kaixiong Gong, Xin Cai, Bin Lin, Hao Wang, Yunlong Lin, Mingzhe Zheng, Bohao Li, Jian-Wei Zhang, Miles Yang, Zhao Zhong, Liefeng Bo, Xiangyu Yue Published: 2026-07-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.22531v1 Unified multimodal models seek a shared visual token space that supports both multimodal understanding and image generation. Discrete methods unify the interface via a shared codebook, whereas continuous pipelines often rely on two disparate representations -- semantic features (e.g., ViT) for understanding and low-level latents (e.g., VAE) for synthesis -- resulting in mismatched latent spaces. We propose Twins, a unified continuous token space formed by channel-wise concatenating ViT and VAE features on the same token grid, so the sequence length is unchanged and attention cost does not increase. However, jointly modeling Twins in a Diffusion Transformer exposes a severe optimization imbalance: the model fits the ViT component well but struggles to match the VAE latent distribution. We trace this imbalance to three sources of heterogeneity: frequency bias, intrinsic dimensionality, and condition-aligned vs condition-independent uncertainty. To address it, we adapt a focal regression objective for flow matching that upweights large-error VAE dimensions, better balancing optimization across the ViT and VAE components. On ImageNet, this yields up to 10.57 gFID gain over naive MSE loss without classifier-free guidance. Twins also performs competitively on multimodal understanding benchmarks and improves reconstruction fidelity, narrowing the gap between understanding- and generation-oriented representations. --- ### Skill Self-Play: Pushing the Frontier of LLM Capability with Co-Evolving Skills Authors: Siyuan Huang, Pengyu Cheng, Haotian Liu, Tao Chen, Yihao Liu, Jingwei Ni, Shijie Zhou, Ziyi Yang, Gangwei Jiang, Mengyu Zhou, Yu Cheng, Xiaoxi Jiang et al. Published: 2026-07-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2607.22529v1 LLM training is shifting from manual design and annotation to interaction-driven self-evolution. However, existing self-evolutionary methods face a fundamental dilemma between task diversity and verification reliability: environment-bound methods obtain precise feedback but confine learning to narrow domains, while open-ended self-generation broadens the task space but lacks reliable verification, allowing misleading rewards to pollute the training loop. We identify agent skills as a powerful middle ground to reconcile this tension: each skill ensures deep, verifiable execution in a specific scenario, while dynamic routing across skills maintains open-ended task variety. Leveraging this insight, we introduce Skill Self-Play (Skill-SP), a co-evolutionary framework comprising a proposer, a solver, and a dynamic skill controller. Orchestrated via a reinforcement learning loop, these components co-evolve in a continuous self-play loop: the proposer generates challenging tasks conditioned on dynamically sampled skills; the solver explores candidate solutions to push its capability boundaries; and the skill controller collects execution feedback to update and expand the skill library. This interactive co-evolution effectively bridges the gap between structured verification and open-ended exploration. Empirical evaluations on tool-use and reasoning benchmarks demonstrate that Skill-SP, serving as a robust evolution engine, consistently pushes the performance ceiling of competent backbones while catalyzing striking turnarounds for initially misaligned models. Our code is available at https://github.com/Qwen-Applications/skill-self-play. ---