Robotics
Reinforcement Learning from Real-Robot Interaction without High-Quality State Estimates
Ryan Gao, Ashvin Anilkumar, Nicholas Corrado et al. (+2 more)
A longstanding goal for
robot reinforcement learning (RL) is that
robots learn from their own physical experience, as humans and animals do. A key obstacle to learning directly from physical experience, especially outside instrumented lab environments, is that the state of the world is usually unobservable to the learner and must be estimated from on-board sensing. Errors in that estimate corrupt both the policy input and potentially the reward signal. We study this challenge using the contact-rich task of a bipedal
robot learning to score soccer goals. We estimate all key state variables on-board with an unscented Kalman filter (UKF) and no external instrumentation (e.g., motion capture or overhead cameras). Because the UKF may incorrectly estimate state when the ball is unobserved, we learn a new state representation with a recurrent state-space model (RSSM) that, critically, only corrects the RSSM's latent state when the ball is observed. Building on RLPD, we introduce RSSM-RLPD and apply it to the task of a NAO
robot learning to dribble a soccer ball into a goal. We show that the
robot learns to score from multiple starting positions after 30 minutes of real-world interaction, reaching 100% success against 0% for a feed-forward policy under the same training budget. We complement our physical
robot experiments with a simulation study and show that RSSM-RLPD learns policies that achieve a higher success rate than behavior cloning, a feed-forward, and GRU baseline. Taken together, our results show that the addition of a learned state representation to existing on-
robot learning methods enables RL from real-
robot data even with a high degree of partial observability.
Read the full abstract ▾
All authors Ryan Gao, Ashvin Anilkumar, Nicholas Corrado, William Cong, Josiah Hanna
Machine Learning
A Geometric Perspective on Reward Function Updates in Inverse Reinforcement Learning
Anish Diwan, Jan Peters and Oleg Arenz
Inverse Reinforcement Learning (IRL) is classically solved using Lagrangian optimisation, by jointly optimising a primal-dual problem with gradient descent to obtain the optimal reward function parameters and corresponding optimal policy. While this algorithmic view of IRL has been highly influential, gradient-based reward updates only paint a partial picture, largely ignoring the geometry of the problem in occupancy space. In this paper, we present an occupancy-space view of reward updates in IRL. We leverage the insight that both the optimal solution and its geometry in occupancy space are known, and show that the geometry along the e-geodesic (straight line in occupancy space) toward the optimal occupancy directly affects the gradient of the IRL dual problem in reward parameter space. This perspective naturally suggests preconditioning the dual gradient with the average Fisher information along the occupancy trajectory. We show that this preconditioned gradient is just a convex interpolation of the old reward and a new target reward, yielding efficient natural-gradient-style updates with little additional computational overhead. We then analyse these occupancy-space reward updates from a Mirror Descent (MD) perspective, and show that policy search over the resulting rewards induces approximate entropic MD on the policy. Finally, we conclude with convergence results of such geometry-aware IRL
algorithms.
Read the full abstract ▾
Cryptography and Security
Symmetric Cryptography via Dirichlet Convolution
Artur Sowa, Elias Hassani Sangani, Francis Bui et al. (+2 more)
We discuss a new symmetric-key cipher for digital data encryption. Its implementation is fast,
memory efficient, and resilient against classical, AI-assisted, and
quantum attacks. Let $x$, $k$, and $c$ be elements of a finite abelian group $G$ with operation $+$ and neutral element $0$. Suppose one is given ciphertext $c=x+k$. Retrieving the plaintext $x=c-k$ from the ciphertext is trivial when the key $k$ is known. Without the key, however, recovering $x$ amounts to a blind search. Even if an efficient criterion were available for identifying the plaintext among all group elements, the average number of trials would remain prohibitively large when the group is sufficiently large. The challenge in constructing a practical cryptosystem is therefore to identify a sufficiently rich group $G$, together with efficient
algorithms that enable an immersion of real digital data into $G$ and efficient evaluation of the group operations. The proposed cipher is based on generalized Dirichlet convolution over finite fields. We investigate its security against plaintext attacks from an information-theoretic perspective. The nonlocal structure of Dirichlet convolution naturally permits plaintext salting, providing protection against plaintext attacks while allowing controlled key reuse. We establish perfect masking properties over finite fields and prove that the resulting cipher achieves information-theoretic security under key reuse when the underlying arithmetic is carried out in $\mathbb{F}_p$ with $p>2$. For arithmetic over $\mathbb{F}_2$, some information leakage remains, leading to a finite key-reuse budget.
Read the full abstract ▾
All authors Artur Sowa, Elias Hassani Sangani, Francis Bui, Grant Harris, Jonathan Norton
Robotics
RAIN: Robotic Task Generalization with Region-Aware Interaction Networks
Seunghun Lee, Kyuhwan Shim, Yeogyeong Song et al. (+5 more)
Existing
vision-
language-action (VLA) models benefit from large-scale
vision-
language pretraining, but tend to memorize training trajectories, limiting their ability to adapt learned interactions to new task and scene configurations. To address this problem, we propose Region-Aware Interaction Networks (RAIN), which formulates a task as a sequence of goal-conditioned subtasks, each specified by an action type and a target-region mask. This formulation allows learned skills to be reused across different objects and scene layouts and composed into task sequences not seen during training. Its Target-adaptive Cross-View Encoder (TCE) conditions multi-view features on target masks and exchanges information across views, while PlanDiT predicts target-directed trajectories and actions. A Transition Head estimates subtask completion to enable automatic transitions. We also introduce a Reference-based Retargeting Strategy (RRS), which uses recorded trajectories to synthesize smooth approaches to reference interactions with alternative target regions, without collecting additional data. Experiments on LIBERO-Analogy, our new
benchmark for robotic task generalization, and real-world tabletop tasks demonstrate that RAIN generalizes more robustly than existing VLA baselines.
Read the full abstract ▾
All authors Seunghun Lee, Kyuhwan Shim, Yeogyeong Song, Seung Hwan Kim, Sihaeng Lee, Byoung-Tak Zhang, Sunghoon Im, Gunhee Lee
Robotics
From bits to atoms: A survey of cross-layer safety in Embodied AI
Tuo Feng, Yue Zhang, Shiji Zhou et al. (+35 more)
Embodied Artificial Intelligence (AI) introduces cyber-physical risks because digital failures and attacks can cause physical harm. This survey organizes
embodied AI safety through a four-layer reference architecture – comprising the physical interaction, system and middleware,
algorithm and decision, and cloud-edge and data layers – and uses it to connect vulnerabilities, propagation paths, and safeguards across the system stack. It traces the
evolution of safety paradigms and examines how faults and attacks cross architectural boundaries to produce physical consequences. Furthermore, the survey systematizes composite attack chains, outlines multi-layered defense strategies, and reviews representative
embodied safety
benchmarks. Ultimately, we advocate for unifying software
alignment with rigorous hardware constraints to build trustworthy systems capable of operating safely within human-centric environments.
Read the full abstract ▾
All authors Tuo Feng, Yue Zhang, Shiji Zhou, Zhaoxin Fan, Wenguan Wang, Xiaowei Chi, Tianyu Shen, Xiangtai Li, Xuesong Bai, Zheng Zhu, Sirui Han, Boyu Zhou, Zhaosong Wang, Shizhe Chen, Zilong Zheng, Chenjia Bai, Jiangmiao Pang, Yaodong Yang, Hao Zhao, Di Hu, Li Yi, Yan Huang, Qi Wu, Hanwang Zhang, Fei Gao, Shanghang Zhang, Hongsheng Li, Jianye Hao, Yuan Zhang, Yingjie Xia, Jian Shen, Cewu Lu, Hui Zhang, Kai Chen, Wenwu Yu, Badong Chen, Jungong Han, Yike Guo
Metric Geometry
Sharp Antipode-Neighbor Bounds in Every Dimension
Samuel Korsky
Let $P\subset\mathbb R^d$ be finite with $\operatorname{diam}(P)\le1$. For $0<\varepsilon<1$, let $\mathcal N_\varepsilon(P)$ denote the number of ordered pairs at distance at most $\varepsilon$, including the diagonal, and let $\mathcal A_\varepsilon(P)$ denote the number of ordered pairs at distance at least $1-\varepsilon$. We prove the sharp bound \[ \mathcal N_\varepsilon(P)\gtrsim_d \varepsilon^{(d-1)/2}\mathcal A_\varepsilon(P) \] conjectured by Steinerberger (2025). More generally, the same inequality holds for finite positive Borel measures. The main input is a uniform $L^2$ estimate, at cap thickness comparable to
lattice spacing, for a supporting-direction incidence operator. In gnomonic coordinates, this operator is controlled by an incidence estimate for small Fenchel–Young gaps.
Read the full abstract ▾
Probability
Functional limit theorems for weighted occupation times of supercritical superprocesses
Ting Yang
Suppose $X=(X_t)_{t\ge0}$ is a supercritical superprocess with a non-local branching mechanism on a Lusin space $E$, whose mean
semigroup admits a positive leading
eigenvalue $\lambda_1$ with a bounded positive eigenfunction $\varphi$ and a finite eigenmeasure $\widetilde{\varphi}$. Let $Y_t:=\int_0^t X_s\mathrm{d}s$ be its occupation measure. We first prove a law of large numbers, that is, for every bounded measurable function $f$, $e^{-\lambda_1 t}\langle f,Y_t\rangle$ converges in $L^2$ to $\lambda_1^{-1}\langle f,\widetilde{\varphi}\rangle W^\varphi_\infty$, where $W^\varphi_\infty$ is the limit of the martingale $W^\varphi_t:=e^{-\lambda_1 t}\langle\varphi,X_t\rangle$. We then establish functional central limit
theorems for the fluctuations of $\langle f,Y_t\rangle$ around this random centering. We show that the normalisation and the limit are determined by the spectral decay rate of $f$ compared with the critical value $\lambda_1/2$, and there are three possible regimes: after suitable normalisation, the fluctuation process converges in distribution in $\mathcal{C}[0,+\infty)$ either to a mixed Gaussian process $\sqrt{W^{\varphi}_{\infty}}G(f)$, or to a constant Gaussian process with random variance depending on $W^{\varphi}_{\infty}$, or to a non-Gaussian limit given by the limit of a martingale associated with the spectral projection of $f$. These results are the occupation-time counterparts of the fluctuation
theorems for instantaneous linear functionals of supercritical non-local branching superprocesses.
Read the full abstract ▾
Complex Variables
Kobayashi visibility of strictly convex domains
Nikolai Nikolov and Amar Deep Sarkar
We show that any bounded strictly convex domain in $\Bbb C^N$, endowed with the Kobayashi distance, is visible.
General Mathematics
Unilateral and Equitransitive Tilings by Squares of Sizes 1, 2, 3, and 4
Layan Arrabi, Casey Mann, Fatima Muniz et al. (+2 more)
In this article we examine a problem stated in Grübaum and Shephard's classic book on tiling theory, Tilings and Patterns, which asks the reader to find all unilateral and equitransitive tilings by squares of sizes $1, 2, 3, \ldots, n$ where $n \in \mathbb{N}$. The unilaterality condition requires that two squares of the same size in such a tiling may not intersect along an entire side, and the equitransitivity requires that two tiles of the same size in such a tiling must be in the same transitivity class with respect to the
symmetry group of the tiling. These two conditions on tilings by squares allows for surprisingly interesting and complex tilings by squares. In this work we provide the answer to Grübaum and Shephard's question for the case $n = 4$, as well as developing methods and computer
algorithms that may be extended to the cases where $n \geq 5$.
Read the full abstract ▾
All authors Layan Arrabi, Casey Mann, Fatima Muniz, Harmony Vargas, Rachel Villafranca
General Physics
Inflationary Dynamics and CMB Constraints in $f(T,L_m)$ gravity
Suchita Patel, Sai Swagat Mishra and P.K. Sahoo
We investigate slow-roll inflation in a modified teleparallel framework with a non-minimal coupling between torsion and the matter sector. Focusing on Starobinsky-type plateau inflation, we derive the modified background dynamics and the corresponding slow-roll observables, and examine how the gravitational couplings affect the scalar spectral index, tensor-to-scalar ratio, and running of the spectral index. The model continuously recovers the standard Starobinsky scenario in the appropriate weak-coupling limit, while the matter–geometry coupling produces a non-monotonic modification of the inflationary predictions. We find that weak couplings leave the standard predictions essentially unchanged, whereas stronger couplings can significantly shift the scalar spectral index and eventually become incompatible with current
cosmic microwave background constraints. The quadratic torsion correction has a comparatively mild impact within the controlled low-energy regime, primarily affecting the tensor amplitude. We confront the resulting predictions with the Planck and BICEP/Keck constraints and find that the model remains viable over a broad range of the inflationary parameter space. Our results demonstrate that the matter–geometry coupling can leave a potentially testable imprint on
primordial observables while retaining the successful plateau-inflation phenomenology of the Starobinsky scenario.
Read the full abstract ▾
Robotics
Does Dynamic-Point Filtering Help When Texture Is Scarce? A Controlled Study of ORB-SLAM2 Front-Ends in Synthetic Indoor Scenes
Zekui Xue
Dynamic-point filters are routinely added to feature-based visual SLAM, and several recent systems argue that removing dynamic features can leave too few static features in low-texture regions. So far, these systems have been evaluated only on texture-rich
benchmark sequences. We present a controlled study that isolates this interaction. We render synthetic indoor sequences in which surface texture (four levels, quantified by FAST-corner density and image-gradient
entropy) and scene dynamics (three levels) are varied factorially along identical camera trajectories, with stereo, RGB-D, ground-truth poses and dynamic masks. On this grid we compare ORB-SLAM2 without filtering, with an optical-flow and epipolar-residual filter (FLOW), and with a multi-view depth-consistency filter (GEOM), and report trajectory error, tracking completeness and surviving static features over five runs. Because the masks give per-keypoint ground truth, we also measure each filter's dynamic-point precision and recall and its static-feature false-removal rate, so that mechanistic explanations can be tested directly. We do not propose a new filter. On 720 runs over 24 sequences, filtering helped mainly in the most dynamic
cells; the benefit did not decline monotonically with texture, but at the lowest level filtering reduced tracking completeness, and ORB-SLAM2 never initialised in static L3 scenes. Contrary to our hypothesis, GEOM discarded more static keypoints than FLOW (median FRR 6.8% vs. 1.5% for RGB-D, 19.6% vs. 1.5% for stereo); its RGB-D advantage tracked dynamic-point recall and vanished in stereo mode. Data and code are available at https://github.com/felixxxue/texture-dynamics-slam.
Read the full abstract ▾
Image and Video Processing
Abdominal Ultrasound Simulation from Semantic Labels using Paired Label-to-Physics-Based Image Translation
Santiago Vitale, Duilio Deangeli, Ignacio Larrabide et al. (+1 more)
Purpose: Current abdominal ultrasound (US) simulation methods often require CT-based anatomical references for ray-casting, limiting deformation and pathology variability. We propose a learning-based pipeline trained to predict physics-based images derived from CT scans from semantic labels, enabling controlled simulations without patient-specific CT volumes at inference time. Methods: We introduce a two-stage pipeline that maps anatomical
segmentations to realistic US images through a simplified US image. Stage~I synthesizes this image from semantic labels using models trained on CT-based ray-casting outputs. Stage~II refines it into a realistic US scan using anatomically guided unpaired translation. Deformations and pathologies are generated by editing anatomical maps. Results: We evaluated Pix2Pix and the Semantic
Diffusion Model (SDM) in Stage~I, followed by
segmentation-guided CycleGAN (SG-CycleGAN) refinement in Stage~II. SDM significantly outperformed Pix2Pix in morphological metrics, including MAE (19.85 vs. 21.65), SSIM (0.28 vs. 0.24), and mIoU (0.43 vs. 0.29), whereas Pix2Pix yielded better perceptual point estimates (LPIPS: 0.17 vs. 0.19; FID: 0.32 vs. 0.37; KID: 0.25 vs. 0.48). Conclusion: Training paired generative models with physics-based supervision enables approximation of CT-derived ray-casting outputs at inference time directly from semantic labels. Although the pipeline does not require patient-specific CT volumes at inference time, CT-derived
segmentations and ray-casting simulations remain necessary to train Stage~I. Once trained, the framework enables controllable healthy and pathological simulations through semantic-map modification.
Read the full abstract ▾
All authors Santiago Vitale, Duilio Deangeli, Ignacio Larrabide, José Ignacio Orlando
Image and Video Processing
STRIDE: Spatial-Temporal Representation for Interval-conditioned Disease Evolution in Longitudinal Glioblastoma MRI
Wenhao Guo, Changchang Yin, Pierre Giglio et al. (+3 more)
Glioblastoma (GBM), an aggressive primary
brain tumor, is routinely monitored with longitudinal MRI after treatment. Distinguishing stable disease (SD), pseudoprogression (PsP), and true progression (TP) remains challenging because these states can show overlapping MRI appearances despite different temporal trajectories. Existing longitudinal methods still face challenges in modeling scan-specific spatial variability, variable follow-up intervals, and complementary information from the observed follow-up state and its longitudinal change. We propose STRIDE, a framework for spatial-temporal representation of interval-conditioned disease
evolution that takes paired post-treatment MRI scans and their inter-scan interval as input and predicts SD, PsP, or TP. The lesion-prior-guided spatial representation combines SoftGate and an Adaptive-window Hierarchical
Transformer (AWHT) to emphasize lesion-related regions while preserving surrounding context. The time-conditioned latent transition uses pair-level context and the actual inter-scan interval to estimate interval-dependent representation changes between visits. The observed–transition fusion integrates the transition-estimated follow-up representation with the directly observed follow-up representation to jointly characterize the follow-up state and its longitudinal change. BraTS2024 is used to develop and evaluate the lesion-prior generator, while longitudinal pretraining on LUMIERE supports transfer before downstream adaptation to Burdenko. On the Burdenko three-class task, STRIDE achieves a macro ROC–AUC of 0.816 and a macro F1-score of 0.796. These results support its potential for more reliable longitudinal post-treatment GBM state assessment.
Read the full abstract ▾
All authors Wenhao Guo, Changchang Yin, Pierre Giglio, Weidan Cao, Ping Zhang, Golrokh Mirzaei
Probability
Complexity of flows with interactions and stochastic analysis
Yifan Shi
We study
stochastic flows with interaction through an infinite-dimensional
semigroup approach. By means of a kernel lifting procedure, we represent the solution to measure-valued
stochastic equation as a
Hilbert-valued Markov process. We identify its generator, construct the associated Markov
semigroup, and establish a dense range condition using measure
polynomial functionals. The results provide a framework for the analysis of stability of
stochastic equations with interaction.
Read the full abstract ▾
Artificial Intelligence
RadOnc-Agent: An Integrated Framework for AI Workflows Across the Radiotherapy Care Pathway
Caiwen Jiang, Shuoyang Wei, Songlin Zhao et al. (+8 more)
Artificial intelligence has advanced radiotherapy tasks, but these capabilities remain fragmented across the care pathway. Here we present RadOnc-
Agent, an integrated framework connecting 17 phase-specific functions and 9 cross-cutting utilities across four phases through a conversational interface. A
language-model controller maps clinical intent to structured calls, maintains longitudinal context and validates dispatch. We evaluated request-expression diversity, function coverage, cross-stage composition, information transfer and failure recovery using 2,600 single-function requests (7,800 executions), 200 cross-stage scenarios (600 executions), and 120 workflows from 60 de-identified patient records (360 executions). RadOnc-
Agent selected the intended function in 98.79% of single-function executions and completed 96.50% of scripted and 96.67% of real-patient workflows. Removing longitudinal state reduced cross-stage completion from 96.50% to 84.00%; bypassing schema and identity validation increased mismatched dispatch from 0% to 95.28% in replay/test execution. Single-function performance did not by itself predict workflow-level capability. Information continuity and cross-stage dependencies emerged as central constraints. These results establish the technical feasibility of longitudinal integration across radiotherapy workflows and provide a foundation for prospective clinical evaluation.
Read the full abstract ▾
All authors Caiwen Jiang, Shuoyang Wei, Songlin Zhao, Junyu Li, Jingyuan Chen, Yuzhen Ding, Daniel J. Ma, Samir H. Patel, Steven E. Schild, Sujay A. Vora, Wei Liu
Instrumentation and Methods for Astrophysics
An Assessment of the Triangulation Capabilities of a Multi-Site All-Sky Infrared Camera Array
Abraham Loeb, Alex Delacroix, Ezra Kelderman et al. (+2 more)
The Galileo Project operates three autonomous multi-camera observing sites near Las Vegas, Nevada, separated by baselines of $2$ and $10\,$km, with the goal of characterizing the populations of natural and human-made objects that cross the sky and thereby isolating any event that fits none of them. Each site carries an array of eight long-wave infrared (LWIR) cameras arranged in a fixed dome-like structure, an IR-Dalek, providing $360^{\circ}$ azimuthal coverage. We report the commissioning of the three arrays and present the triangulation pipeline built on them. Because our targets are not assumed to follow any particular trajectory, we depart from the parametric fits standard in meteor work and estimate the instantaneous three-dimensional position at each time step as the point that minimizes the weighted sum of squared distances to the simultaneous lines of sight, weighting each camera by its own pointing uncertainty and by its range to the object. We describe the intrinsic and extrinsic calibration procedure, which uses Automatic Dependent Surveillance-Broadcast (ADS-B) aircraft as reference sources in the absence of
stars or fixed landmarks in the LWIR field of view, and the multi-sensor track association scheme that decides which detections belong to the same object. Using one week of data (2026 May 24-30; $1{,}650$ camera-hours, $2.1\times10^{8}$ detections and $5.3\times10^{6}$ tracks), we validate the recovered positions, speeds and accelerations against ADS-B records, recovering distances to within $5$% for $99$% of the aircraft solved from three sites and for $83$% of those solved from a single site pair and speeds to within $5$% for $78$% and $63$% of them respectively. We characterize the identification completeness of the array, which is set by apparent size: the $50$% identification threshold falls at $1.8$ pixels of apparent wingspan.
Read the full abstract ▾
All authors Abraham Loeb, Alex Delacroix, Ezra Kelderman, Eric Masson, Regina Sarmiento
Robotics
Does a Learned Corrector Beat a Simple Retreat? Evidence from a Frozen VLA
Chenchao Sheng, Zhuang Jiang, Liuhaichen Yang et al. (+2 more)
Before deploying runtime recovery for a frozen
vision-
language-action (VLA) policy, one must establish that an intervention improves success beyond ordinary run-to-run variation and that its complexity adds value over a simple action. We evaluate these questions on frozen $\pi_{0.5}$ across four RoboTwin tasks. For each test seed, we pair rollouts with and without correction and include a same-seed base-policy re-run as a placebo. Seed-cluster intervals and prespecified comparison rules assess net gains against
stochastic outcome changes. Across 3,888 paired episodes, the full pipeline raises success on beat_allowbreak block_allowbreak hammer by $+13.5$\,pp (95% interval $[+9.4,+17.7]$), with no detectable gain on the other three tasks at the deployed weight. Among failed base episodes on the responsive task, $43.2\%$ succeed on a plain re-run, compared with $63.5\%$ after correction; many nominal rescues therefore reflect the base policy's own variability. A fixed-time trigger and scripted return to an earlier joint configuration produce a net gain with no detected difference from the learned pipeline across two rounds, although our prespecified equivalence criterion is not met consistently. Pausing and a constant-action control do not yield comparable gains. On this
benchmark, the decision to intervene depends strongly on the task, and a paired placebo plus a simple retreat baseline are needed to establish what learned correction contributes.
Read the full abstract ▾
All authors Chenchao Sheng, Zhuang Jiang, Liuhaichen Yang, Ningwei Bai, Zezhi Tang
Sound
Extending Music Annotation Schemas: Zero-Shot Prediction or Few-Shot Adaptation?
Christos Plachouras, Emmanouil Benetos and Johan Pauwels
Automatic music annotation is typically tackled under the assumption of a fixed annotation schema. In practice, commercial music catalogs often need to accommodate new musical attributes as needs evolve. Given that expert music annotation is expensive, it is not evident which methodological approach is most effective at accommodating new attributes and backfilling existing tracks; audio-
language models promise zero-shot prediction, but at what annotation budget does supervised adaptation become more compelling? We propose a
benchmark based on the MGPHot popular music annotation
dataset for simulating music schema extension across different annotation budgets. We investigate zero-shot prediction with audio-
language models, learning new attributes from pretrained representations, and adapting models trained on existing annotations. Our results suggest that supervised adaptation is more effective than zero-shot prediction even with small annotation budgets, while frozen representation reuse remains the most effective approach for modest budgets without the tuning required by deeper adaptation.
Read the full abstract ▾
Artificial Intelligence
Anchor Divergence for Semantic Geometry in Contrastive Learning
Akash Kannan, Kiho Park and Victor Veitch
This paper concerns how semantic context determines geometry in learned vector representations. Similarity is typically measured using cosine similarity, which provides a single fixed geometry. Semantic similarity, however, is inherently context dependent: two images may be similar because they depict the same object, share a visual style, or are relevant to the same clinical finding. We show that contrastive representations naturally encompass a family of geometries that can be specialized to particular semantic structure. The key idea is to use an interplay between contrastive learning, exponential families, and information geometry to establish a correspondence between probability distributions over "anchors" and Bregman geometries on the representation space. We use this correspondence to define "Anchor Divergences", a method for specifying context-specific semantic geometries on fixed representations. Under this correspondence, modeling the anchor distribution models the geometry itself. Experiments on
retrieval show that anchor divergences provide an effective and efficient way to specify context-specific semantic similarity.
Read the full abstract ▾
Machine Learning
Learning from Unreliable Trajectories: Adversarially-Robust Federated Q-Learning
Sreejeet Maity and Aritra Mitra
We study
federated reinforcement learning in which multiple
agents interact with a common Markov decision process and communicate through a central server to collaboratively learn the optimal state-action value function. Our goal is to understand whether the sample-efficiency benefits of collaboration can be retained when a fraction of the
agents behave adversarially and transmit arbitrarily corrupted information. To address this problem, we introduce Robust Async-Fed-Q, an epoch-based
federated learning
algorithm that combines variance-reduced estimation of the Bellman optimality operator at the
agents with robust aggregation at the server. We establish high-probability finite-time guarantees showing that the proposed method preserves the statistical gains of collaboration among the honest
agents while tolerating
adversarial corruption. In particular, the effect of the
adversarial agents decreases as the amount of data collected by each honest
agent grows and eventually vanishes in the infinite-sample limit. We complement these guarantees with information-theoretic lower bounds that characterize the unavoidable statistical cost of
adversarial corruption, leading to the first nearly matching upper and lower bounds for adversarially robust
federated reinforcement learning. We further extend our framework to accommodate single-trajectory Markovian sampling and heterogeneous partial coverage, where different
agents may explore different regions of the state-action space and learning relies on their collective coverage. Finally, our epoch-based design substantially improves the best known communication complexity for
federated Q-learning under asynchronous sampling.
Read the full abstract ▾
Artificial Intelligence
FluidPD: In-Place Elasticity for SLO-Aware Prefill-Decode Disaggregated LLM Serving
Kartik Ramesh, Kaidi Fu, Zihan Zheng et al. (+4 more)
Prefill-decode disaggregation is becoming a common architecture for LLM serving because it separates two phases with distinct execution patterns and SLO objectives. Existing systems typically combine a fixed prefill/decode worker ratio with request routing across workers. However, real-world workloads exhibit both short bursts and sustained shifts in the prefill-to-decode demand ratio. As a result, a configuration that is well provisioned at one time may quickly become mismatched, causing latency SLO violations even when idle capacity exists elsewhere. Existing autoscaling mechanisms can add capacity, but they react slowly, require spare GPUs, and do not directly address short-timescale phase imbalance. We present FluidPD, a P/D-disaggregated serving system that provides SLO-aware in-place elasticity. FluidPD introduces two complementary mechanisms. FluidToken handles transient imbalance by offloading a bounded portion of prefill computation to decode workers when decode-side slack is available. FluidRole handles sustained imbalance by reassigning running workers between prefill and decode roles in place, avoiding model reload and engine restart. Both mechanisms are guided by lightweight pressure indices that expose prefill and decode-side resource pressure before they appear as SLO violations. Across production Azure trace workloads, FluidPD improves overall SLO attainment over static SGLang by up to 94.6 percentage points, demonstrating that SLO-aware in-place P/D elasticity improves service quality without provisioning additional workers.
Read the full abstract ▾
All authors Kartik Ramesh, Kaidi Fu, Zihan Zheng, Jiahuan Yu, Fabio Oliveira, Carlos Costa, Minjia Zhang
High Energy Astrophysical Phenomena
Classification of X-ray Binaries in NGC 6946 with the use of Multiwavelength Data
Senay Avdan, Hasan Avdan, Eda Sonbas et al. (+1 more)
We identified X-ray binary (XRB) candidates in the
galaxy NGC 6946 using archival Chandra data. A total of 98 XRB candidates were detected within the 0.3$-$8 keV energy range, with luminosities between 10$^{36}$ and 10$^{40}$ erg~s$^{-1}$. X-ray color-color diagrams were constructed for these sources, classifying 57 as high-mass X-ray binaries (HMXBs) and 8 as low-mass X-ray binaries (LMXBs). The X-ray temporal variability of the sources was analyzed, revealing that 40 sources exhibit variability and 26 are classified as transient sources. Possible optical counterparts for 57 sources were identified using Hubble Space
Telescope (HST) images. Based on the color-magnitude diagrams (CMDs), 40 sources were found to have masses greater than 8M$_{\odot}$ and were classified as HMXBs. Seventeen sources with masses between 5 and 8M$_{\odot}$ were classified as intermediate-mass X-ray binaries (IMXBs). The remaining sources, for which optical counterparts could not be identified, were assumed to be LMXBs. The X-ray luminosity functions (XLFs) derived for HMXBs and LMXBs, identified through both X-ray and optical classification, were best described by a single power-law model ($\alpha \sim$ 1.3$-$1.6). We supplemented traditional classification methods with a machine learning (ML) approach based on multiwavelength photometry. While traditional methods identified no AGN candidates, the ML approach detected a significant AGN population. The spatial resolution of the observing instruments and the limits of the training database hinder the reliability of the ML classifications, thus highlighting a key challenge in combining these methods for extragalactic X-ray source
Read the full abstract ▾
All authors Senay Avdan, Hasan Avdan, Eda Sonbas, Kalvir S. Dhuga
Strongly Correlated Electrons
Disorder into Correlation: A Microscopic Mechanism for Sub-McIntyre Avalanche Noise
Shafat Shahnewaz, Hannaneh Karimi, Joe C. Campbell et al. (+1 more)
Avalanche multiplication is intrinsically
stochastic, yet experiments reveal excess noise below the minimum predicted by McIntyre's local-field theory. The microscopic origin of this noise suppression remains incompletely understood. We show that polar-optical-
phonon emission, confined to a narrow forward cone, competes with elastic
alloy scattering that diffuses field-collimated hot carriers out of the cone. Suppressed
phonon cooling narrows the ionization path beyond the dead space and synchronizes successive ionizations into a dynamic staircase. This explains the low noise of Sb-based
alloys and HgCdTe and makes
alloy disorder a design parameter.
Read the full abstract ▾
All authors Shafat Shahnewaz, Hannaneh Karimi, Joe C. Campbell, Avik W. Ghosh
Artificial Intelligence
A Governed Agent Architecture for Natural-Language Dashboard Generation over Enterprise DataBrain
Yiou Wu, Zezhi Tang, Ningwei Bai et al. (+1 more)
We present a DataBrain-specific prototype that turns natural-
language analytic requests into inspectable dashboards. An installable Codex plugin and standalone
Agent Runtime combine schema-constrained model decisions with typed tools, persistent state, and deterministic Hooks for approval, audit, checkpointing, recovery, and failure handling. The pipeline resolves entities, discovers metadata, enforces read-only SQL, composes dashboards, and applies static checks, dynamic preflight, and browser inspection. The model proposes actions while deterministic software controls execution and records state transitions. We evaluate the workflow on frozen real-DataBrain tasks and controlled Hook faults. Strict success was 6/8 on metadata and SQL tasks: metadata selection passed 4/4, all four SQL tasks met semantic criteria, and 2/4 met the exact output-column contract. The final release passed 4/4 single-panel dashboard tasks, one two-panel task, and one existing-dashboard refinement; a parameterised task exceeded its step limit. All ten fault scenarios met their specified outcomes without unapproved external side effects. Model inference accounted for over 97% of observed runtime in every reported group. These small, DataBrain-specific results do not establish production readiness, general text-to-SQL accuracy, or an efficiency advantage over manual dashboard construction.
Read the full abstract ▾
All authors Yiou Wu, Zezhi Tang, Ningwei Bai, Liuhaichen Yang
High Energy Astrophysical Phenomena
Relaxation of a Vlasov gas to an inhomogeneous state due to phase space mixing in an axisymmetric potential: A Newtonian analogy of the Kerr orbital motion
Ana Laura García-Perciante, Merce Guerrero, Darío Núñez et al. (+1 more)
We explore the dynamics of a Vlasov gas propagating in an external axisymmetric potential consisting of a central potential with an additional quadrupolar component which gives rise to two potential wells along the
symmetry axis. Employing independently $N$-particle simulations and statistical methods, we show that an initially homogeneous and isotropic configuration evolves to an inhomogeneous final state with overdensities located at the wells. The quadrupolar component is chosen such that the equations of motion form an integrable
Hamiltonian system, which allows one to compute the final state of the gas analytically. On the one hand, this allows one to compare the late-time behaviour of the simulations with an analytic prediction and on the other hand to understand the relaxation process through phase space mixing. Our model constitutes a Newtonian analogue of the particle motion in a Kerr spacetime, and we discuss possible applications to recently observed astrophysical phenomena.
Read the full abstract ▾
All authors Ana Laura García-Perciante, Merce Guerrero, Darío Núñez, Olivier Sarbach
Strongly Correlated Electrons
Correlated Low-Energy Model of Monolayer 1H-NbS$_2$: A cRPA+DMFT Study
Karollaine C. Leite, Jorge L. B. de Faria, Luiz Antonio Ribeiro Junior et al. (+3 more)
Monolayer $1\mathrm{H}\text{-}\mathrm{NbS}_2$ hosts a well-separated metallic band of predominant Nb-$d_{z^2}$ character, providing an ideal platform to probe electronic correlations within a single-band setting. In this work, we present a comprehensive study of $1\mathrm{H}\text{-}\mathrm{NbS}_2$ by combining first-principles Wannier interpolation, constrained random-phase approximation (cRPA), and DFT+DMFT. We show that an effective single-band model accurately captures the low-energy electronic structure, reproducing key experimental features from ARPES and STS $dI/dV$ spectra, including a characteristic Van Hove singularity. While static DFT+$U$ fails to describe the correlated metallic nature of the system, DFT+DMFT successfully accounts for the spectral weight redistribution and dynamic correlations. By defining the cRPA target subspace directly within the isolated Nb-$d_{z^2}$ band, we obtain a renormalized local interaction of $U_{\mathrm{cRPA}} = 1.138~\mathrm{eV}$. This interaction stabilizes a strongly correlated metallic state that remains robust upon cooling. Remarkably, this minimal single-band model captures a spectral broadening similar to that previously reported in more elaborate treatments - incorporating larger local interactions ($U \sim 1.8~\mathrm{eV}$, derived from a multiorbital cRPA construction), intersite Coulomb terms ($V$), and electron-
phonon coupling - demonstrating that a consistently downfolded local interaction alone can already capture key aspects of the experimentally observed low-energy behavior of monolayer $1\mathrm{H}\text{-}\mathrm{NbS}_2$.
Read the full abstract ▾
All authors Karollaine C. Leite, Jorge L. B. de Faria, Luiz Antonio Ribeiro Junior, Alexandre C. Dias, Helena Bragança, Alberto S. de Arruda
Networking
EDAP: An Efficient Data-Gathering Protocol for Wireless Sensor Networks
Arash Nasiri Eghbali, Nastooh Taheri Javan and Mehdi Dehghan
Directed
Diffusion (DD), uses mechanisms such as data aggregation and in-network processing to suppress the additional data overhead however there is no guarantee that paths from nearby sources join after a few hops. Also for path construction, each source floods a distinct exploratory data (ED) packet through the network which is not necessary for nearby sources. In this paper an efficient data-gathering protocol (EDAP) is proposed to address the problems of late-aggregation and distinct ED-flooding through using a virtual sink (VS) node nearby the sources which plays the role of sink node and broadcasts local interest messages and routed gathered data toward destination. Also in this protocol, multiple paths are constructed between virtual sink and the sink node which leads to load-balancing and increase in the life-time of the network nodes. Our routing protocol can produce more than one disjoint or braided path and spread the gathered data properly between constructed paths according to their length and remaining energy. Simulation results shows that in this way a significant amount of energy can be saved and the network lifetime will be increased considerably (up to 75% better than DD).
Read the full abstract ▾
Artificial Intelligence
GAMEGO: Training Game-Dev Agents with Synthetic Trajectories Anchored in Real-World Assets
Haoyue Yang, Jingyao Li, Zhengfan Wu et al. (+3 more)
Recent advances in Large
Language Models (LLMs) have demonstrated remarkable capabilities in web front-end execution, with browser-based game generation emerging as a particularly prominent frontier. While previous efforts frequently rely on complex multi-turn workflows or focus on static game evaluation
benchmarks, this work targets direct end-to-end real-world game synthesis driven by coding
agents. However, generating complex games directly from sparse user queries often forces coding
agents to make underspecified assumptions, yielding incomplete mechanics, disconnected gameplay flows, and limited visual aesthetics. To resolve this issue, this paper presents GameGo, a scalable framework that systematically transforms brief game seeds into comprehensive Product Requirements Documents grounded in industry game-development practices. To retain core gameplay constraints without restricting design exploration, GameGo uses task-specific dynamic compression to maximize information density while preserving instruction following. Based on this pipeline, GameGoData is constructed with 55,060 development trajectories across 2D, 2.5D, and 3D games, alongside GameGoBench, a
benchmark comprising 124 diverse game queries. Training GameGoCoder on GameGoData yields a model that outperforms matched baselines and is comparable to frontier models across gamedev
benchmarks. All code,
datasets, and models will be made publicly available.
Read the full abstract ▾
All authors Haoyue Yang, Jingyao Li, Zhengfan Wu, Jing Liu, Xuanle Zhao, Kang Liu
Group Theory
Unipotent automorphisms of modular group algebras and the homology of the Lazard Lie algebra
Diego Tuzzolo
Let $p$ be an odd
prime and $H$ a finite $p$-group. Hertweck and Soriano parametrize the central Frattini extensions of $H$ by subspaces of a relation module $V$, matching isomorphism classes of extensions with orbits of $\mathrm{Out}(H)$ and isomorphism classes of the associated small group rings with orbits of $\mathrm{Out}(\mathbb{F}_p H)$. The discrepancy between the two is measured by the group $U(H)\leq GL(V)$ of displacements induced by the normalized unipotent
automorphisms of $\mathbb{F}_p H$; it is the mechanism behind the counterexamples to the modular isomorphism problem in characteristic $2$. We prove that for $H$ of exponent $p$ and class at most $p-2$ every such
automorphism acts on $V$ exactly as an
automorphism of $H$ does, and that $U(H)=\exp\Omega\bigl(\mathrm{Der}_{>0}(\mathfrak{l}(H))\bigr)$, where $\mathfrak{l}(H)$ is the Lazard Lie
algebra of $H$, $\mathrm{Der}_{>0}$ its derivations raising the lower central filtration, and $\Omega$ the natural action of derivations on $H_2(\mathfrak{l}(H))$. Hence $U(H)=1$ precisely when every derivation of positive degree acts trivially on $H_2(\mathfrak{l}(H))$. We compute $U(H)$ for all groups of class $2$ and exponent $p$, obtaining $p^{2n(n-1)}$ for the Heisenberg groups $\mathrm{Heis}(p^n)$.
Read the full abstract ▾
Computation and Language
Trajectory-Derived Confidence for Reliable, Resource-Aware Clinical Text-to-SQL Agents
Mincheol Daniel Song, Joshua Ward, Jake Jung et al. (+1 more)
LLM
agents for clinical text-to-SQL applications reason autonomously over multiple steps but cannot assess whether their own
reasoning or outputs can be trusted. In high leverage applications such as healthcare, this presents a critical risk where system mistakes can be costly. These reliability failures are also resource failures: an incorrect
reasoning trajectory spends computation budget on outputs that must be discarded. We introduce Sentinel, a trajectory-derived, classifier-based confidence layer that analyzes an
agent's
reasoning, code and database outputs to decide at three points whether to stop: refusing unanswerable questions before the
agent runs, halting doomed trajectories mid-run, and withholding untrustworthy answers at delivery. Here, utilizing Chow's rule, we optimize decisions under the EHRSQL shared task's Reliability Score, which penalizes incorrect answers given a utility weighting, and find on the
benchmark EHRSQL that Sentinel raises this score from +0.08 to +0.24 when mistakes have a low utility weighting, well above the +0.03 earned by refusing every question, with delivered-answer accuracy rising from 54% to 69% as coverage falls from 84% to 55%. At higher stakes the
agents we test rarely answer reliably enough to deliver, and Sentinel detects this on its own, abstaining to that same +0.03 where the unmonitored
agent scores -3.02. The same stopping decisions cut computation: at low stakes, where the system still answers, Sentinel eliminates 13-28% of
agent steps at little or no reliability cost.
Read the full abstract ▾
All authors Mincheol Daniel Song, Joshua Ward, Jake Jung, Guang Cheng