Sound
An Initial Study of Physiological Markers for Deepfake Speech Detection
Vojtěch Staněk, Michal Frič, Kamil Malinka et al. (+1 more)
Conventional deepfake
speech detectors primarily target synthesis artifacts and often degrade under unseen attacks or recording conditions. Human
speech, however, contains physiological markers associated with respiration, phonation, and temporal coordination arising from the
speech-production system that synthesizers may not reproduce faithfully. This motivates the hypothesis that deepfake
speech can be detected by identifying inconsistencies in these patterns, which we test using 36 handcrafted physiological markers organized into respiratory, vocal, and temporal categories. We evaluate the marker vectors with logistic regression and SVM classifiers on pooled English recordings. The best in-domain system achieves a 21.40% equal error rate (EER), while logistic regression achieves 32.77% on the out-of-domain
dataset pair. Since the markers describe specific
speech-production properties, the logistic-regression coefficients indicate whether higher marker values push the classifier toward bona fide or deepfake
speech. These results provide initial evidence that physiological markers offer a useful, interpretable signal for deepfake
speech detection.
Read the full abstract ▾
All authors Vojtěch Staněk, Michal Frič, Kamil Malinka, Anton Firc
Artificial Intelligence
The Agentic Scientist: Orchestrating Coding Agents for Automated Scientific (Co-)Discovery
Robert Tjarko Lange, Luca Grillotti, Yutaro Yamada et al. (+2 more)
Modern coding
agents are fundamentally reshaping the role of human input for computational research: they write code, launch in silico experiments, analyze their results, and autonomously operate with ever-increasing time horizons. Despite these developments, fully autonomous research
agents suffer from a limited ideation radius and myopic hill-climbing search dynamics. To bridge this gap, we introduce The Agentic Scientist, an end-to-end distributed pipeline for human researchers to conduct computational research alongside modern coding
agents. Building on its predecessor, our system decomposes research into task drafting, idea generation, planning, sandboxed experimentation, result aggregation, baseline and ablation construction, post-hoc analysis, and report writing. Importantly, each step in the pipeline is powered by the next generation of coding
agents and leverages agentic self-verification. The overall research process is driven by a meta-controller that manages parallel research
agents and enforces rubric verification at each stage. This design enables the same orchestration layer to be reused across heterogeneous domains with different execution environments, agentic coding harnesses, and evaluation metrics. Our system can be used across a spectrum of autonomy: either as a co-scientist tool or to perform end-to-end scientific investigation. Within the co-scientist setting, we leveraged a development version of our system to submit a single research paper, which got accepted for publication at ALIFE 2026. We further apply The Agentic Scientist fully autonomously to open problem-solving tasks in machine learning, and demonstrate the system's capability of scientific discovery across different domains of computational research. Our ablations suggest that reliable scientific autonomy depends not only on base model capabilities, but also on the cross-
agent orchestration layer.
Read the full abstract ▾
All authors Robert Tjarko Lange, Luca Grillotti, Yutaro Yamada, Rujikorn Charakorn, David Ha
History and Overview of Mathematics
From a Student's Question to the Half-Derivative: A Constructivist Teaching Model - Part I
Davit Kapanadze
This paper presents a conceptual, cognitive-constructivist model for teaching the half-derivative. The model starts from a student's question and frames it as the problem of generalizing the familiar operation of differentiation. On a suitably chosen space of monomials, under the assumptions of linearity and of lowering the exponent by one half, the requirement that the operator applied twice give the ordinary derivative produces a recurrence for the coefficients. The freedom of normalization becomes visible, and an additional convention yields a formula compatible with the monomial rule of the Riemann-Liouville half-derivative. The domain of the operator and its relation to the Riemann-Liouville and Caputo operators are treated separately. The mathematical results rest on the author's earlier preprints; the contribution of this paper is their transformation into an instructional pathway organized in phases, with diagnostic questions, an analysis of typical errors, and comparative problems. The model is a design proposal: its effectiveness has not yet been established by systematic empirical research. The interactive Socratic implementation and the detailed diagnostics are the subject of Part II.
Read the full abstract ▾
Computational Physics
A Unified Framework for Bell-Shaped Gaussian Quadratures in Deterministic Boltzmann Solvers
Shanshan Dong, Lu Wang and Guanqing Wang
In this work, we develop a unified theoretical framework that constructs Bell-shaped Gaussian quadratures for deterministic Boltzmann solvers. The proposed framework establishes a general class of bell-shaped weight functions together with the corresponding integral transformations, revealing the common mathematical structure underlying existing generalized Gaussian quadratures. Based on this framework, a unified procedure for constructing discrete velocity sets is developed by combining radial quadratures with appropriate directional quadratures in different velocity-space dimensions. Several existing generalized Gaussian quadratures are naturally recovered as special cases, while the proposed framework provides a systematic route for constructing new quadrature rules based on different bell-shaped weight functions. Extensive
benchmark simulations over a broad range of flow regimes demonstrate that the proposed framework achieves high integration accuracy and computational efficiency with significantly fewer discrete velocities than conventional quadrature strategies. Overall, the proposed framework not only unifies existing generalized Gaussian quadratures under a common mathematical foundation but also establishes a general design principle for the systematic development of efficient velocity-space quadratures in deterministic Boltzmann solvers.
Read the full abstract ▾
Information Theory
On the Investigation of Antenna Beam Pattern Models for Fast Beam Placement in LEO Satellite Networks
Ninh Thi-Thanh Tam, Nguyen Viet Anh, Nguyen Minh Quan et al. (+1 more)
In this paper, we investigate the impact of
antenna beam pattern models on the beam placement problem in Low Earth Orbit (LEO) satellite networks. Unlike conventional beam placement studies that often rely on a fixed
antenna pattern or a predefined half-power beamwidth, this work compares four representative gain models, namely the Bessel $J_1$ model, the combined Bessel $J_1+J_3$ model, the Gaussian model, and a 3GPP-TR 37.842 model. For each
antenna pattern, the corresponding half-power angle is obtained and used to construct a user compatibility graph. The beam placement problem is then transformed into a Minimum Clique Cover problem and solved using a graph coloring-based
algorithm. Numerical results compare the considered
antenna models in terms of the number of active beams, load balancing, statistical channel gain-to-noise ratio (SCGNR), and signal-to-interference-plus-noise ratio (SINR). We analyze the impact of increasing the number of active beams on SINR-based spectral efficiency to explain the importance of carefully controlling the number of active beams. The results show that
antenna beam pattern models affect both beam resource usage and link quality, highlighting the need for
antenna-pattern-aware beam placement in LEO satellite networks.
Read the full abstract ▾
All authors Ninh Thi-Thanh Tam, Nguyen Viet Anh, Nguyen Minh Quan, Trinh Van Chien
Dynamical Systems
Packing intermediate dimensions of measures and applications
Nicolas Angelini
We introduce a spectrum of dimensions of measures, the packing intermediate dimensions $\dim_{P,\theta}\mu$, $\theta\in(0,1]$, of a Borel probability measure $\mu$ on $\mathbb{R}^n$ and the quasi-Hausdorff dimension $\dim_{qH}\mu$. We show that this spectrum encodes the information on $\mu$ that governs the packing dimension of its projections and Hölder images. First, $\dim_{qH}\mu$ is the exact threshold for the packing dimension profiles of Falconer and Howroyd: for every real $0<s\leq n$, $\dim_P^s\mu=s$ if and only if $s\leq\dim_{qH}\mu$. Second, we prove a lower bound for the profiles in terms of $\dim_H\mu$, $\dim_{P,\theta}\mu$ and the Assouad dimension of $\operatorname{supp}\mu$, which improves the bound of Falconer and Mattila, the best previously known general lower bound for the packing dimension of projected measures. The improvement is strict whenever $\dim_H\mu<\dim_P\mu$ and $\dim_A(\operatorname{supp}\mu)<n$, and when $\dim_A(\operatorname{supp}\mu)=n$ it occurs exactly when $\dim_{P,\theta}\mu$ exceeds an explicit universal lower bound for some $\theta$ in an explicit range. We present examples of measures with $\dim_H\mu<\dim_P\mu<\dim_A(\operatorname{supp}\mu)=n$ for which our bound gives the exact value $\dim_P\mu_V=\min\{\dim_P\mu,m\}$ for almost every $m$-dimensional subspace $V$, whereas the bound of Falconer and Mattila is strictly smaller. As applications, for a large family of random $\alpha$-Hölder maps $f_\omega:\mathbb{R}^n\to\mathbb{R}^m$, we prove that, for compactly supported $\mu$, $\dim_P\mu_{f_\omega}=m$ almost surely if and only if $\alpha m\leq\dim_{qH}\mu$, and that the packing analogue of Kahane's formula holds for fractional Brownian motions if and only if $\dim_{qH}\mu=\dim_P\mu$. Analogous results are obtained for analytic sets.
Read the full abstract ▾
Number Theory
Orthogonality relation between $GL(3)$ coefficients and additive characters with linear and non-linear phases
K. Venkatasubbareddy
We study the distribution and oscillatory properties of
prime-supported
Fourier coefficients attached to a Hecke–Maass form for $\mathrm{SL}(3, \mathbb{Z})$ twisted by additive characters. For linear phases involving irrationals of finite type, we construct an automorphic variant of Vaughan's identity and exploit bilinear cancellation to obtain an unconditional power-saving estimate. For non-linear phases $e(\alpha n^\theta)$ with $0 < \theta < 1$, we refine the Type~II bilinear forms via Weyl differencing and modern exponent pair methods. Combined with the analysis of untwisted $\mathrm{GL}(3)$ sums, this framework breaks the $2/3$-barrier of Jiang and Lü and widens the admissible range of non-trivial cancellation.
Read the full abstract ▾
Artificial Intelligence
Why Static Skill Inconsistencies Do Not Equate to Runtime Risks: A Behavioral Analysis of LLM Agent Execution
Fengrui Fan, Huahao Wang, Jingyi Fang et al. (+3 more)
LLM
agent frameworks heavily rely on external skill assets to standardize tool usage, encode operational workflows, and supply executable runtime resources. Despite their widespread adoption, multi-component skill packages frequently suffer from cross-artifact inconsistencies, where documentation, scripts, and configuration files specify conflicting behavioral constraints. Although static auditing techniques can uncover such contractual discrepancies, static defect volume cannot reliably reflect real-world
agent performance degradation. This gap leaves practitioners without rigorous evidence to assess the practical risk of flawed
agent skills. This paper presents a systematic empirical study to demystify the runtime implications of skill inconsistencies. We conduct large-scale auditing over 575 community and
benchmark skills, yielding 1,182 manually validated contract inconsistencies. We further build a controlled evaluation pipeline with 49 executable
agent tasks and three representative LLMs across NoSkill, OriginalSkill, and FixedSkill experimental settings. Through fine-grained behavioral analysis of 441 paired execution trajectories, we quantify the actual performance impacts of resolved defects.
Read the full abstract ▾
All authors Fengrui Fan, Huahao Wang, Jingyi Fang, Yutao Hu, Yueming Wu, Deqing Zou
Image and Video Processing
Personalising Deformable Image Registration with Clinical Context and Deformation History for Adaptive Radiotherapy
Caiwen Jiang, Songlin Zhao, Yuzhen Ding et al. (+23 more)
Adaptive radiotherapy must account for anatomical changes during treatment, but doing so depends on reliably mapping targets, organs at risk and dose across serial images. Existing registration methods treat each image pair independently, without using radiotherapy-specific context or the patient's preceding anatomical trajectory. We developed Clinical-context-Aware and Patient-adaptive Deformable Image Registration (CAP-DIR), integrating source contours, dose, clinical text and prior-registration features. We evaluated 16,361 CT examinations from 3,473 patients across four institutional cohorts and five public
benchmarks. CAP-DIR increased internal-test organ-at-risk-region normalized cross-correlation from 84.76% to 96.92% relative to the strongest learning-based comparator, with low deformation folding, and remained superior across seven external cohorts. In a blinded review of 20 longitudinal cases, clinically acceptable contours increased from 45.0% to 82.5% for clinical target volumes and from 22.5% to 62.5% for organs at risk, while estimated editing time decreased. In a controlled
benchmark, CAP-DIR reduced global dose-mapping error by 34.5% relative to the strongest comparator. These findings show that clinically informed, longitudinal registration can improve contour usability and dose mapping for adaptive radiotherapy.
Read the full abstract ▾
All authors Caiwen Jiang, Songlin Zhao, Yuzhen Ding, Jia Mi, Samir H. Patel, Alexandra Bennett, Keldon K. Lin, Terence T. Sio, Jonathan B. Ashman, Lisa A. McGee, Jean-Claude M. Rwigema, William G. Rule, Sameer R. Keole, Sujay A. Vora, William W. Wong, Nathan Y. Yu, Michele Y. Halyard, Steven E. Schild, Safia K. Ahmed, Daniel J. Ma, Robert L. Foote, Michelle A. Neben Wittich, Thomas J. Whitaker, Haibo Lin, Dinggang Shen, Wei Liu
Hardware Architecture
CacheSmith: Accurate RAG Serving under Latency SLOs with Adaptive KV Reuse
Xinyue Ma, Damin Lee, Aashaka Shah et al. (+1 more)
This paper presents CacheSmith, a KV cache reuse system for RAG prefill that maximizes quality-gated SLO attainment. While context caching lowers prefill latency by reusing precomputed KV states for retrieved chunks, it misses cross-chunk
attention that a full prefill would compute over concatenated inputs. Prior selective recomputation schemes tackle this problem but enforce static, one-size-fits-all configurations that fail to effectively handle request heterogeneity or dynamic latency budgets. CacheSmith decomposes selective recomputation into a three-phase workflow of WarmUp, Probe, and Patch that can be dynamically configured to exploit a wide range of cost-accuracy trade-offs. At serving time, an SLO-aware scheduler in CacheSmith forms batches to mitigate inter-request interference during layer-by-layer prefill, while a budget-aware planner configures recomputation phases across the batch to maximize the number of requests satisfying both generation quality and latency targets. Both components make fast, solver-free decisions over profiled cost-accuracy bands. On four
benchmarks across varying arrival rates and burstiness levels, CacheSmith improves quality-gated SLO attainment by up to 27.5%, over four state-of-the-art baselines.
Read the full abstract ▾
All authors Xinyue Ma, Damin Lee, Aashaka Shah, Myeongjae Jeon
Robotics
ChunkAct: Action Chunking for Agentic Robotics
Bingxin Xu, Wenrui Bao, Xinxin Liu et al. (+3 more)
Agentic robotics places a large
language model (LLM)
agent in the
robot control loop, where it composes perception, motion, and
vision-
language-action (VLA) tools under the guidance of a
memory and thereby generalizes to new tasks without retraining. This generality comes at a steep price in efficiency. Every primitive requires a round of LLM
reasoning over the full interaction history, so the cost of an episode grows with the number of primitives. We find this
reasoning largely redundant: once a sub-goal is chosen, its primitives follow a routine already prescribed by
memory. We present ChunkAct, which introduces action chunking to agentic robotics. Online, a single call launches a chunk of chained primitives, which executes autonomously and returns control only upon failure. Offline, ChunkAct compiles the
memory into these chunks, so execution is
memory-free at test time and the planner is reserved for decisions that require
reasoning. LLM computation therefore scales with the number of decisions rather than the number of primitives. On LIBERO-Pro, ChunkAct improves task success by 5.0 points over the state of the art in agentic robotics while requiring 2.9x fewer LLM calls, 4.4x fewer input
tokens, and 2.3x fewer output
tokens.
Read the full abstract ▾
All authors Bingxin Xu, Wenrui Bao, Xinxin Liu, Jiawei Zhou, Yuzhang Shang, Emilio Ferrara
Robotics
LoopWAM: Closed-Loop Generalist Policy Improvement from Deployment Rollouts
Wenxuan Guo, Yingping Liang, Zhaoguo Wang et al. (+11 more)
Scaling expert demonstrations alone does not enable generalist
robot policies to improve from deployment. Real-world rollouts contain successes, suboptimal behavior, and failures: imitating every recorded action can reinforce poor behavior, while filtering out failures discards evidence of physical interactions and their consequences. We introduce LoopWAM, a closed-loop world-action model that learns directly from this full quality spectrum, using rollouts from both itself and other deployed policies. Source-dependent noise sampling unifies behavior learning and action-conditioned world modeling, while quality steering and quality-weighted supervision retain full
video supervision across qualities and regulate action imitation.
Video and action
tokens interact throughout parallel denoising, allowing learned consequences to inform action generation directly. On RoboChallenge Table30-V2, three deployment-and-training rounds raise success from 23.67% to 43.33%, a 19.7-percentage-point gain, with a final score of 56.80. LoopWAM also achieves 94.5% average success on RoboTwin 2.0 and 49.5% overall success on RoboCasa365 without additional embodied pretraining, obtaining the best overall results among the compared methods across all three
benchmarks. On challenging shirt folding task, two rounds of deployment-driven improvement after single-task post-training raise success from 13.20% to 55.56%. These results demonstrate closed-loop improvement from mixed-quality experience under both multi-task learning and single-task post-training. Project page: https://loop-wam.github.io/
Read the full abstract ▾
All authors Wenxuan Guo, Yingping Liang, Zhaoguo Wang, Ziyuan Li, Angyuan Ma, Zhenyu Wu, Chengkun Wang, Bingyao Yu, Ze Chen, Borui Zhang, Xiuwei Xu, Meng Zhang, Erjin Zhou, Jiwen Lu
Robotics
EgoExplorer: A Vision-Motion-Touch Observatory for Human Loco-Manipulation In Situ
Yunsong Zhou, Xuekun Jiang, Haiqi Shen et al. (+11 more)
Learning humanoid loco-manipulation from human demonstrations rests on observing how locomotion, posture, and object interaction evolve together as tasks unfold. Scaling collection across diverse environments, in turn, favors lightweight, portable capture systems. Yet, existing
datasets provide only partial views of this coordination. Thus, we introduce EgoExplorer, a 100-hour
vision-motion-touch observatory of human loco-manipulation built with a wearable system. With an emphasis on professional activities, its continuous recordings align egocentric
video with reconstructed whole-body motion, fine-grained finger articulation, and tactile observations, accompanied by step-level descriptions. To achieve this coverage with sparse body sensing, we develop the Calibrate-Complete-Constrain pipeline. Concretely, reusable calibration fixes participant geometry and sensor attachments before motion inference to disentangle mounting offsets from articulation. A temporal prior then completes unobserved joint motion, while layered constraints promote measurement agreement, temporal continuity, and
geometric consistency, properties that motion plausibility alone does not guarantee. Against tracker-fitted references, the method reduces body-point error by 71.9% relative to the strongest baseline; matched evaluations also show gains in hand stability and contact estimation. Preliminary studies explore
multimodal generation and conditioning alongside humanoid retargeting with whole-body control in simulation.
Read the full abstract ▾
All authors Yunsong Zhou, Xuekun Jiang, Haiqi Shen, Hangxu Liu, Chaojian Shi, Yujie Liu, Jingyuan Li, Haoran Yang, Jiafei Cao, Dongzhan Zhou, Dong Wang, Yao Mu, Chunhua Shen, Weinan Zhang
Computer Vision
Cepheid: Decoupling Scale from Structure for Cross-Family Motion Completion
Chih-Hai Su, Tsung-Wei Ke and Yu-Lun Liu
Sparse-to-dense motion completion recovers dense mesh motion from sparse vertex displacements. Residual priors trained on one geometry family frequently fail to beat parameter-free harmonic extension off-family. Moreover, cross-family training meets a scaling barrier: residual amplitude varies by more than an order of magnitude across geometry families and collapses as observations become denser. Since this scale depends on unknown target motion, it cannot be measured directly at test time. We decouple residual structure from physical scale by learning a dimensionless prior, estimating amplitude from observables and a lightweight correction fitted only on training sources. Our unified model surpasses harmonic extension across all evaluated families, including unseen cloth and legged bodies, without skeletons, registration, or cross-mesh correspondence. The same prior applies without retraining to point-tracked animal
video and a
robot grasping a soft body. Project page: https://su-terry.github.io/cepheid-website/
Read the full abstract ▾
Computer Vision
MiniDog: Dog-Breed Text-to-image Flow Matching from Pretraining to Fine-Tuning on 4 Consumer GPUs
Reyhaneh Esmailizadeh, Xingjian Leng, Zhanhao Liang et al. (+1 more)
We generally perceive that training text-to-image
diffusion models requires heavy compute: it is true if we aim at a comprehensive frontier model. This understanding prevents people from training their own models when getting started in the field. We develop MiniDog, a framework that enables pretraining
diffusion transformers in 55 minutes and fine-tuning in another 15 minutes on four 3090 GPUs. For data, we use 26k dog images from ImageNet and caption them for pretraining and use 2k images generated by FLUX.1-schnell for fine-tuning. We precompute the image and text features and cache them for efficiency. For architecture, we use E2E-INVAE for its semantic latent space, a small
transformer with 22M parameters, a small Qwen model for text encoding, and in-context conditioning. For training, we use iREPA and an EUPE perception encoder for acceleration. For evaluation, our repository supports FID, HPSv2 score, PickScore, and various VAE reconstruction metrics. The project is open-sourced at https://github.com/Reyhaneesmailizadeh/MiniDog. We do not claim technical novelty.
Read the full abstract ▾
All authors Reyhaneh Esmailizadeh, Xingjian Leng, Zhanhao Liang, Liang Zheng
Robotics
What Matters in Quality-Aware VLA Pretraining: A Benchmark and Empirical Study
Mingxuan Yan, Litian Gong, Dihong Huang et al. (+19 more)
How should demonstration quality be used during
vision-
language-action (VLA) pretraining to improve performance after target-task fine-tuning? We compare weighting, filtering, and conditioning across reward definitions and label granularities using 5,000 naturally suboptimal human demonstrations and 24 target tasks. Within each VLA, we match demonstrations and training budgets. Additional comparisons examine pretraining data size and the roles of quality tags during fine-tuning and inference. Across three VLAs, we find that approach-aware episode conditioning improves mean success over plain behavioral cloning; conditioning and filtering favor episode and chunk labels, respectively, on average, with model-dependent reversals; and constant-tag fine-tuning improves mean success after conditioned pretraining with either tagged or untagged inference. We introduce AXIS-Bench to support these controlled comparisons. Built on crowdsourced simulation demonstrations, it provides recorded object and contact states for recomputing quality labels on shared trajectories. Replayable trajectories support workspace-aligned rendering and user-defined real-world target tasks. In our real-world experiment, simulation pretraining raises success from 9.6% to 31.7%. Project page: https://axis-bench.github.io/.
Read the full abstract ▾
All authors Mingxuan Yan, Litian Gong, Dihong Huang, Zhixu Li, Tianyu Zhang, Wenqian Zhang, Hefeifei Jiang, Peihao Li, Jian Zhang, Mengfei Zhao, Yikai Tang, Hai Zhai, Zehao Wang, Ruijian Liang, Yanjia Huang, Lin Shao, Changliu Liu, Kaiyu Hang, Zhiwen Fan, Masayoshi Tomizuka, Jianfei Yang, Jiachen Li
Computation and Language
Tev: Scaling Long-Horizon Agentic Environments with Dense Progress Rewards
Wenbo Hu, Shraman Pramanick, Krishna Kumar Singh et al. (+3 more)
Long-horizon
agents have drawn growing
attention for their ability to work autonomously – from hours to days – on economically valuable tasks, a capability essential to automated discovery and recursive self-improvement (RSI). Yet open-source models still struggle on long-horizon tasks, largely for lack of both training data and a working recipe, and scaling both is hard: long, complex tasks are difficult to curate and verify, and their rewards are sparse. We address both. First, we introduce Tev-Gym, the first long-horizon
agent training
dataset with verifiable, dense, progressive rewards – 1,170 verified task environments across 7 domains. It is built on a latent-world-first approach that first constructs a latent world: a complete and correct hidden specification of the task domain, workflow, and expected state after each step. Each task is then progressively rendered from this specification into an
agent-facing environment, enabling controllable complexity and a natural progress metric. Second, we turn this metric into a training recipe: via progress-potential reward shaping (PPRS), we convert progress into a dense per-step reward that improves training efficiency and makes otherwise-intractable tasks trainable. Training with this recipe yields Tev-27B, which achieves substantial gains on the
Agents' Last Exam computing and math sciences domains (ALE-CM), matching GPT-6 Astra and Claude Opus 5 at Pass@5. Despite training only on ALE-related data, it generalizes to the Terminal-Bench (TB) 3.0 and 4.0 subsets, surpassing GLM 5.2 and Gemini 3.7 Flash, and reaches Claude Opus 4.8-level performance on middle-horizon tasks such as TB 2.1.
Read the full abstract ▾
All authors Wenbo Hu, Shraman Pramanick, Krishna Kumar Singh, Nanyun Peng, Kai-Wei Chang, Yong Jae Lee
Robotics
EmbodiedEvo: Self-Evolving Coding Agent For Physical Intelligence
Wenyi Li, Yizhe Chi, Youjie Zheng et al. (+8 more)
Coding
agents can construct executable programs and revise them through feedback, offering a foundation for embodied control without target-task fine-tuning. Through a programmatic
robot interface, the
agent observes the scene, executes control programs, and adapts them to the resulting physical state, serving as the closed-loop policy. However, this within-episode adaptation does not automatically preserve experience across scenes. Our trajectory analysis shows that
agents spend interaction and inference budgets establishing effective operations and identifying their preconditions, while fresh
agents cannot reuse these discoveries. This motivates EmbodiedEvo, a self-evolving coding
agent that distills physical feedback into reusable code Skills and diagnostic Lessons. Experience is acquired on one Playground scene per task, then frozen for independent
agents in unseen scenes, without updating model weights. On RoboDojo-Sim, EmbodiedEvo achieves an average score of 44.48 and a success rate of 37.82%, surpassing all officially verified results. Paired tests show the largest gains on
memory-dependent tasks and successful reuse of coordinated procedures, while fine contact control remains difficult. Experience from stronger models also improves weaker executors: Astra's experience raises Sol and Luna's success rates by 8.86 and 5.50 percentage points over no experience. Compared with raw interaction history, Skills and Lessons achieve comparable task performance at 53.0% lower execution cost.
Read the full abstract ▾
All authors Wenyi Li, Yizhe Chi, Youjie Zheng, Chunyu Liu, Xiaoqiu Wang, Mingju Gao, Zheng Jiang, Deyao Hong, Calvin Xiao, Kaisen Yang, Qinhuai Na
Robotics
AutoAgent0: An Agentic Runtime for Safe Closed-Loop Driving
Jiageng Liu, Seth Z. Zhao, Zewei Zhou et al. (+9 more)
Safe autonomous driving requires more than accurate planning at a single frame: it demands reliable closed-loop execution in dynamic environments. Although end-to-end driving policies have advanced rapidly, the accumulated closed-loop errors, especially in unseen scenarios, can push vehicles into states from which the same policy struggles to recover. We address the challenge by treating closed-loop driving safety as a test-time execution problem. We introduce AutoAgent0, an agentic runtime and harness framework that augments existing driving experts with complementary detection skills, rule-based action verification, and feedback-driven recovery. To address inaccurate scene perception, an independent detection module constructs scene context using complementary perception skills, and a rule-based verifier checks proposed trajectories against this context before execution. When the nominal trajectories are unreliable, the orchestrator of the runtime activates Recovery Planning, where an agentic loop constructs alternative action proposals from action primitives and iteratively refines them using verification feedback. This design combines independent scene assessment with feasibility checks and online recovery, addressing both perception errors and the limited recovery capabilities of individual experts. Evaluations on Fail2Drive and NavSafe-$\infty$ demonstrate improved closed-loop robustness over single-expert execution in long-tail and safety-critical scenarios. These results demonstrate that runtime supervision can improve closed-loop robustness across multiple driving experts without retraining the underlying policies.
Read the full abstract ▾
All authors Jiageng Liu, Seth Z. Zhao, Zewei Zhou, Alejandro Huerta, Aidan Bayer-Calvert, Yuxin Bao, Haotian Lin, Robert Spataru, Zijun Cui, Yining Hong, Bolei Zhou, Jiaqi Ma
Computer Vision
PLAYGROUND: Progressive Layout Generation with Render-Grounded Planning
Jaeho Lee, Junhwan Heo, Junghyun Park et al. (+5 more)
Generating editable graphic designs requires arranging visual and textual elements into a coherent composition while preserving their content and editability. Although recent MLLM-based approaches employ iterative generation and visual refinement to reflect the real-world design workflows, their high-level plans are predetermined prior to execution, limiting their adaptability as the design evolves. To address this limitation, we propose PlayGround, a progressive layout generation framework that dynamically adapts to the current rendered canvas. PlayGround establishes a progressive design loop comprising a Progressive Planner, a Layout Generator, and a Visual Verifier, guided by a Structure-Aware Retriever. Given an initial set of unordered elements, the Structure-Aware Retriever initially infers a coarse layout structure to retrieve structurally similar reference designs. Subsequently, at each turn, the Progressive Planner examines the current rendered canvas alongside the remaining elements, forms the next functional group, and plans its composition. The Layout Generator then predicts editable layout attributes, while the Visual Verifier evaluates rendered candidates and provides feedback when refinement is necessary. To train the Progressive Planner with step-wise supervision, we construct a two-stage data generation pipeline that extracts functional groups and turn-level planning trajectories from completed designs. Experiments on Crello and LICA demonstrate the effectiveness of PlayGround over existing methods. Our code and weights will be publicly released.
Read the full abstract ▾
All authors Jaeho Lee, Junhwan Heo, Junghyun Park, WonJun Moon, Eunju Yang, Seungho Jang, Seongchan Kim, Seungryong Kim
Earth and Planetary Astrophysics
Simulations of the Evolving Ozone Layer: Implications for the Early Earth, Exoplanets, and the Search For Life
Akhil Kumar and Gregory J. Cooke
Over 4 billion years, Earth's atmosphere has been transformed from a composition devoid of molecular oxygen (O$_2$) to one where O$_2$ is now 21% of our modern atmosphere by volume. With high quantities of O$_2$, an ozone (O$_3$) layer can form and partially shield the surface from biologically harmful ultraviolet radiation (UV). The
evolution of the O$_3$ layer has previously been simulated under the modern Sun by independent groups with separate models. Here, we present simulations of Earth's atmosphere from four open-source one-dimensional (1D) photochemical models and a 3D chemistry-
climate model, focussing on O$_3$. We find that the O$_3$ layer thickness exhibits a clear deviation between the 1D models and the 3D model due to several factors, including treatments of O$_2$ photolysis, chemical networks, H$_2$O microphysics and 3D transport. We find that 1D models introduce a bias for unobserved
climate states and show how these biases may propagate to calculations of biological UV damage and surface habitability. We show that when applied to specific
exoplanet atmospheres, these discrepancies can become more pronounced, propagating through to observational predictions for the next generation of
telescopes, such as HWO and LIFE, that will search for habitable and inhabited
exoplanets.
Read the full abstract ▾
Lattice Field Theory
Data Free Training of Diffusion Models for Lattice Field Theories
Dennis Grinwald, Ankur Singha, Khaled Kahouli et al. (+2 more)
Sampling
equilibrium field configurations remains a major computational bottleneck in
lattice field theory, motivating the development of generative samplers that can produce independent configurations.
Diffusion models are promising for this purpose, but their standard denoising score matching objective requires samples from the target distribution, seemingly reintroducing the original sampling problem. Here, we show that
diffusion models can instead be trained from their own generated samples using the target score identity, provided access to the unnormalized probability density. Importance-weighted formulations of this identity yield a consistent target throughout training, while unweighted variants define a self-consistent fixed-point training scheme. We
benchmark the different approaches in two-dimensional \(\phi^4\) theory and identify effective choices that enable efficient training and sampling across the symmetric, critical, and broken phases. Ultimately, our data-free training approach achieves comparable performance to training on HMC reference data.
Read the full abstract ▾
All authors Dennis Grinwald, Ankur Singha, Khaled Kahouli, Karl Jansen, Shinichi Nakajima
Computation and Language
DoubtLess: Training-Free Contrastive Decoding for Efficient Reasoning
Minghao Luo and Liang Chen
Reasoning models achieve strong performance through long chains of thought (CoT), but often spend more
tokens than needed to solve a problem. Existing ways to shorten the CoT require additional training for each model, add a fixed steering vector to the model's hidden states, or use decoding heuristics such as stopping early or banning words like "wait", which can reduce accuracy. We propose DoubtLess, a training-free decoding method in which a much smaller
reasoning model steers the large one towards brevity. The small model reads the large model's CoT so far twice, once prompted to be brief (the brief thinker) and once prompted to be verbose (the verbose thinker), and the difference between its two logit vectors, scaled by a strength, is added to the large model's logits at every step of the CoT. On seven
benchmarks of competition mathematics, graduate-level science and competitive programming, DoubtLess shortens the CoT of Qwen3-32B by 31% with unchanged accuracy, outperforming five inference-time baselines. We also show that DoubtLess reduces thinking
tokens per turn by 20% in a self-debugging coding
agent with unchanged task performance. It also generalizes to two other model families with unchanged accuracy. When combined with training-based compression, it further shortens the CoT of a model already trained for brevity. Ablations and behavioral analysis show that DoubtLess reduces repeated checking, with most
token savings coming after the correct answer first appears in the CoT. Our code is available at https://github.com/leoluolol/DoubtLess.
Read the full abstract ▾
Computation and Language
Jiaru Zou, Yunzhe Qi, Sirui Chen et al. (+5 more)
How can we build an
agent that thinks beyond human
language, yet still acts through it in the external environment? Modern LLM
agents still reason and interact through text, while latent
reasoning offers an efficient internal medium but lacks native tool access and
reasoning transparency. We bridge these two directions with Hybrid
Reasoning Agents (HybridAgents), which dynamically navigate between latent
reasoning and text-based interaction throughout a unified generation process. Architecturally, a lightweight HybridLink enables seamless transitions between text and latent modes, while the
agent's continuous working
memory preserves lossless hybrid context across multiple turns. These designs preserve tool-integrated agentic capabilities while making latent thoughts directly inspectable. Building on this architecture, we further develop dedicated post-training
algorithms for hybrid
reasoning: HybridSFT stabilizes tool-integrated hybrid generation by teaching the backbone to coordinate with HybridLink, then HybridRL scales
reasoning by learning when to think in latent space and when to act through text. We evaluate HybridAgent at 3 model scales across 11
benchmarks covering mathematics, science, medicine, search, and agentic
reasoning. Compared with frontier LLM
Agents and agentic frameworks spanning both text-based and latent methods, HybridAgent consistently improves performance by 10.4%, while delivering 35.5%-54.7%
token reduction, 1.8x-2.6x faster inference, and 2.7x-3.7x cost savings. Code and data are available at https://github.com/togethercomputer/HybridAgent.
Read the full abstract ▾
All authors Jiaru Zou, Yunzhe Qi, Sirui Chen, Federico Bianchi, Shang Zhu, Pan Lu, Jingrui He, James Zou
Artificial Intelligence
AesCode: Aesthetic Code Generation with Decoupled Cross-Modal Rewards
Xin Gao, Yifan Yang, Xin Li et al. (+13 more)
Generating information-rich visual artifacts such as slides, posters, and dashboards as code is appealing because the outputs are structured, editable, and verifiable. Yet code models cannot see how their choices of layout and color come together on the canvas, often limiting spatial planning and aesthetic quality. Image generators excel at visual design but can misrender requested content, making their outputs unreliable targets for direct reproduction. We therefore introduce AesCode, which generates code grounded in the content
prompt while using an image generated from the same
prompt as an aesthetic guidance. To do so, AesCode represents the target as a design graph that describes the whole canvas and makes each property attributable. From this graph, we derive complementary cross-modal rewards: deterministic verifiers score properties parsable from code and rendering, while a VLM judge assesses the non-parsable ones with a sample-specific Visual Graph Rubric (VGR) tied to graph elements and relations. We optimize these decoupled rewards through reinforcement learning with per-channel normalization to jointly improve spatial planning, aesthetic quality, and verifiable correctness. To evaluate this setting, we build a
benchmark that scores verifiable correctness and visual quality separately, with visual scores highly consistent with human judgments. We find that reference images substantially improve the visual quality of code generated by existing models, and ablations show that AesCode benefits from references both as model input and as a source of visual rewards. Relative to same-scale Qwen3-VL baselines given the same references, AesCode-8B and AesCode-32B improve visual quality by 31.1 and 22.4 percentage points, reaching scores comparable to those of closed-source frontier models such as GPT-5.5 and Claude Opus 4.8.
Read the full abstract ▾
All authors Xin Gao, Yifan Yang, Xin Li, Pan Wang, Yan Li, Muzhao Tian, Mingxi Cheng, Bei Liu, Kai Qiu, Qi Dai, Dong Chen, Yue Dong, Leandros Tassiulas, Rex Ying, Ji Li, Chong Luo
Combinatorics
Szemerédi's theorem with random differences and arithmetic Kakeya
Gabriel Dahia
We prove that there exists an absolute constant $c > 0$ such that the following holds for every $k \ge 3$ and $p \ge n^{-c / \log \log k}$. Letting $\delta > 0$ and $D \sim [n]_p$, with high probability every ${S \subset [n]}$ with $|S| \ge \delta n$ contains a $k$-term arithmetic progression whose common difference lies in $D$. This improves the best known bounds for the threshold of Szemerédi's
theorem with random differences for all sufficiently large $k$, previously of the form ${p \ge n^{-2/k+o(1)}}$, due to Briët and Castro-Silva. We obtain our improvement via a small cover lemma, in the sense of Alon and Pham's work on random Cayley graphs, and its proof crucially relies on a construction of Green and Ruzsa in the context of the arithmetic Kakeya
conjecture. The usage of the latter was suggested to us by GPT 6 Astra.
Read the full abstract ▾
Computer Vision
Visual Tokens on Demand: Visual Budgeting for Scalable and Efficient Multi-Image Reasoning in Multimodal LLMs
Aimon Rahman, Joymallya Chakraborty, Vidit Bansal et al. (+1 more)
Multimodal large
language models have advanced rapidly, but most remain optimized for single-image inputs. Multi-image
reasoning is costly because each image adds many visual
tokens, increasing encoding and
attention demands. Existing techniques that address the above issue, including visual
token pruning,
token merging, visual
token compression, learned resampling, hierarchical compression,
retrieval-augmented visual
reasoning, and
memory-based architectures, have largely been designed for single-image scenarios, where redundancy is primarily spatial and local. Multi-image
reasoning introduces additional challenges: redundancy may occur across images, task-relevant evidence may be sparse or distributed, and aggressive per-image compression can remove details needed for comparison, counting, temporal
reasoning,
retrieval, or evidence grounding. Moreover, existing
token-reduction methods are often either (i) training-free but inference-inefficient, typically relying on accessing
attention matrices to get
token-importance that may be impractical to access with
memory-efficient
attention implementations such as FlashAttention, or (ii) training-dependent, requiring the base MLLM to be retrained or adapted to a new compressed-
token paradigm. In this work, we study scalable multi-image
reasoning and understanding without retraining the base model and without accessing
attention matrices at inference time. To this end, we propose mmVB, a lightweight plug-and-play module that assigns query-conditioned per-image
token budgets via cross-image context, adding only 1.6M trainable parameters while reducing visual input
tokens by 9x. This reduction yields 5x faster time-to-first-
token and 7x fewer prefill FLOPs, while maintaining 94.4% of full-resolution accuracy across eight multi-image
reasoning benchmarks.
Read the full abstract ▾
All authors Aimon Rahman, Joymallya Chakraborty, Vidit Bansal, Abhishek Dan
Computer Vision
PixelMoE: Scaling Pixel-Space Diffusion with LLM-Style Mixture-of-Experts
Ruixiao Dong, Jianmin Bao, Bozhi Luan et al. (+6 more)
Pixel-space
diffusion models provide an appealing alternative to latent-space
diffusion models by directly modeling raw images without relying on a learned tokenizer. However, scaling pixel-space models has predominantly followed dense strategies, tightly coupling model capacity with active computation. In this work, we show that standard LLM-style Mixture-of-Experts (MoE) architectures can be effectively transferred to pixel-space
diffusion. Building on this finding, we introduce PixelMoE, a pixel-space generative framework that sparsely scales model capacity with standard LLM-style MoE, without
diffusion-specific routing mechanisms or auxiliary routing objectives. Specifically, PixelMoE combines standard
Token-Choice (TC) routing with fine-grained routed experts for input-adaptive sparse computation, a shared expert for capturing common representations, and a bias-based load-balancing strategy. To complement sparse capacity scaling, PixelMoE further introduces Pixel-Space Hierarchical Flow Matching, which varies spatial resolution along the denoising trajectory from global structure formation to fine-grained detail refinement. At comparable active parameter counts, PixelMoE consistently outperforms its dense counterparts across model scales, demonstrating effective sparse capacity scaling in pixel space. On class-conditional ImageNet-1K, PixelMoE achieves FID scores of 1.47 at 256x256 and 1.53 at 512x512 without representation
alignment or perceptual losses, comparing favorably with leading pixel-space and latent-space generative models. Code and models will be released soon.
Read the full abstract ▾
All authors Ruixiao Dong, Jianmin Bao, Bozhi Luan, Zijie Meng, Xin Tao, Pengfei Wan, Kun Gai, Li Li, Houqiang Li
Machine Learning
Exact Topological Compliance: Generating Persistence-Equivalent Graphs
Mattie Ji, Indradyumna Roy and Vikas Garg
Persistent homology summarizes the topological
evolution of a graph as persistence diagrams (PDs). Generating graphs whose
topology matches a target descriptor is increasingly important, yet existing
topology-aware generative models rely on soft regularization, producing graphs whose PDs may only be approximately similar. We initiate a principled investigation of the generation problem: produce graphs that realize a target PD exactly. We propose two approaches guided by the implicit local and global constraints a target PD imposes. The local approach builds degree-based PD-equivalent graphs iteratively, and we prove that every output realizes the target PD and that every degree-based PD-equivalent graph is reachable. However, as shown thereafter, local generation necessarily trades off between sample diversity and reaching infeasible states. We therefore recast generation as a global constraint satisfaction problem with a complete encoding solvable by modern constraint programming solvers. This is applicable to any vertex-based permutation equivariant filtration scheme and is minimal under the degree filtration. Experiments on four standard graph generation
benchmarks confirm that both methods realize the target PD exactly. Together, our methods provide a principled treatment of strict
topology-compliant graph generation. Code is available at https://github.com/Aalto-QuML/ExactCompliance.
Read the full abstract ▾
Robotics
Interaction Latents as Actions for Contact-Rich Manipulation
Yuhang Zheng, Yujie Zang, Yuchen Liu et al. (+9 more)
Contact-rich manipulation requires
robots to coordinate not only where to move, but also how contact and force should evolve throughout an interaction. Yet most
robot policies represent actions primarily as motion trajectories, using tactile signals as inputs or auxiliary prediction targets without explicitly encoding contact dynamics in the action representation. We propose ILA (Interaction Latents as Actions), which uses decodable interaction latents that jointly capture the
evolution of motion, contact, and force application as policy action units, providing a shared representation for action generation, evaluation, and offline improvement. We learn these latents through visually conditioned joint action-tactile reconstruction with a time-conditioned implicit decoder. Using them as flow-matching generation targets supports post-training of pretrained
vision-
language-action models (VLAs) and world action models (WAMs), as well as training flow-matching policies from scratch. A shared decoder maps generated latents to executable actions. For offline improvement, we learn a tactile-conditioned expert reference model from interaction latents encoded from expert demonstrations. The resulting expert-consistency rewards guide weighted learning from successful rollouts, while observation-compatible expert references guide correction of low-reward predictions from failed rollouts. Across three policy families, ILA improves success rates averaged over four real-world tasks by 21.0-36.5 percentage points relative to tactile-conditioned direct-action baselines. Ablations highlight the benefit of encoding contact
evolution in action targets beyond tactile conditioning, while the same latent space supports expert-guided offline improvement. Project website: https://mrsecant.github.io/ILA/.
Read the full abstract ▾
All authors Yuhang Zheng, Yujie Zang, Yuchen Liu, Yupeng Zheng, Zhennan Jiang, Zebin Xing, Shuai Tian, Songen Gu, Jingxiang Guo, Chen Gao, Zhongpu Xia, Shuicheng Yan