Peer Revenue Now accepting positions

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Sunday 8 November

Machine Learning

JoMoHa: Joint Model–Harness Optimization with Reinforcement Learning

The performance of language agents depends jointly on model weights and the harness. Optimizing either alone can be constrained by the other component, limiting overall performance gains. Our analysis of naive alternating training reveals progressive incompatibility between the model and harness, which increasingly interferes with task execution and may also undermine model training, leading to performance degradation. To coordinate the evolution of models and harnesses, we introduce JoMoHa, to our knowledge the first framework to jointly train an executable harness generation policy and a task execution policy through reinforcement learning. JoMoHa outperforms the evaluated weight-only and harness-only baselines and achieves the highest success rates among the evaluated methods on ALFWorld, WebShop, and DBBench, averaging 70.3% and reaching 97.0% on ALFWorld. By learning both task execution and runtime improvement from its own generated trajectories, JoMoHa performs recursive self-improvement (RSI).
Read the full abstract

All authors Xudong Wu, Guangchen Lan, Mingqi Yuan, Jiayu Chen

Cryptography and Security

EVAX: Towards a Practical, Pro-active & Adaptive Architecture for High Performance & Security

This paper provides an end-to-end solution to defend against known microarchitectural attacks such as speculative execution attacks, fault-injection attacks, covert and side channel attacks, and unknown or evasive versions of these attacks. Current defenses are attack specific and can have unacceptably high performance overhead. We propose an approach that reduces the overhead of state-of-art defenses by over 95%, by applying defenses only when attacks are detected. Many current proposed mitigations are not practical for deployment; for example, InvisiSpec has 27% overhead and Fencing has 74% overhead while protecting against only Spectre attacks. Other mitigations carry similar performance penalties. We reduce the overhead for InvisiSpec to 1.26% and for Fencing to 3.45% offering performance and security for not only spectre attacks but other known transient attacks as well, including the dangerous class of LVI and Rowhammer attacks, as well as covering a large set of future evasive and zero-day attacks. Critical to our approach is an accurate detector that is not fooled by evasive attacks and that can generalize to novel zero-day attacks. We use a novel Generative framework, Evasion Vaccination (EVAX) for training ML models and engineering new security-centric performance counters. EVAX significantly increases sensitivity to detect and classify attacks in time for mitigation to be deployed with low false positives (4 FPs in every 1M instructions in our experiments). Such performance enables efficient and timely mitigations, enabling the processor to automatically switch between performance and security as needed.
Read the full abstract

All authors Samira Mirbagher Ajorpaz, Daniel Moghimi, Jeffrey Neal Collins, Gilles Pokam, Nael Abu-Ghazaleh, Dean Tullsen

Computation and Language

TID System for Speech-COSER Challenge 2026: ASR Track

We describe the system submitted by Telefónica Innovación Digital to the Speech-COSER Challenge 2026, which features spontaneous speech from elderly speakers recorded in rural Spain. For ASR, we perform CTC-based pseudo-forced alignment scores audio-text match quality to guide a two-stage LoRA curriculum, followed by post-hoc magnitude pruning of the LoRA matrices to enhance generalization. An ensemble of seven systems combined via ROVER achieves 36.25% unnormalized and 22.17% normalized WER.
Read the full abstract

Statistical Methodology

Graph-regularized synthetic control for spatial panels with post-only treated locations

Spatial policy panels can contain treated locations that are observed only after an intervention and therefore have no pre-treatment outcome history. Without a pre-treatment trajectory to match, choosing donor weights for such post-only locations requires an additional source of information. We propose Graph-SCM, which estimates donor weights jointly across treated locations and encourages geographically neighboring locations to use similar donor combinations. GeoEco-SCM keeps the same geographic graph but adjusts the strength of its existing edges using pre-treatment economic similarity. Both methods handle complete, partial, and absent histories within one convex weight-estimation problem. Monte Carlo experiments compare them with covariate-balanced synthetic controls, equally weighted donor averages, ordinary synthetic control where estimable, and nonseparable spatio-temporal imputation and direct prediction. No method dominates. Graph-based borrowing improves on covariate balancing in several settings, but a selective-coverage design favors covariate balancing. Economic adjustment helps when geography alone misses relevant donor relationships. The results support graph regularization as a way to extend synthetic-control coverage when connected locations share similar donor representations, with its usefulness depending on the structure of the available information.
Read the full abstract

General Relativity and Quantum Cosmology

Smooth Rigidity Near Extremal Horizons

The Einstein equations constrain the transverse Taylor series at a degenerate Killing horizon, but even complete knowledge of that series need not determine a smooth spacetime: a nonzero difference can vanish to every order at the horizon. We show that this ambiguity disappears when the limiting metric on the space of Killing orbits admits a spacelike graph over a compact horizon section. In dimensions $d\geq4$, smooth vacuum Einstein germs with the same marked transverse jets are then locally isometric on a two-sided neighborhood of the whole section. The proof combines a terminal wave-map gauge with energy estimates on the orbit cylinder and yields a finite weighted matching criterion. For unwarped horizons with positive-Einstein sections, a condition on the full Lichnerowicz spectrum reduces the smooth classification to parallel first transverse data. All such data have convergent multiwarped Einstein realizations, including examples whose sections have negative sectional-curvature directions. We also obtain a finite-regularity version for round sections. The continuation argument extends to four-dimensional Einstein–Maxwell theory and gives smooth charged-Nariai rigidity. Finally, rotating Nariai data admit the required spacelike graph throughout the regular branch; an all-order tensor estimate yields smooth Kerr–de Sitter or rotating-Nariai rigidity for $a^2/r_0^2\leq1/8$, without assuming ambient axial symmetry.
Read the full abstract

Machine Learning

Regret without Optimal-Arm Feedback under Reward-Dependent Delays

We study stochastic bandits with reward-dependent delays, possibly infinite, whose reports omit the originating round. For each $c\in[0,1)$, Balanced Historical Optimism (BHO) has a delay term controlled by the largest suboptimal-arm quantile at level $1-c\Delta_i$, without requiring optimal-arm reports. Lower bounds give an inverse-square slack cost in the logarithmic statistical regime and linear minimax regret at $c=1$. A family with growing $K$ forces worst-case regret $\Omega(G\log K)$ for every learner, where $G=\max_i\Delta_i h_i$ is the largest gap-weighted suboptimal cutoff. Pending-triggered balance preserves BHO's weighted-cutoff bound for $c<1$ and attains the historical-greedy exploration coefficient with deterministic rewards, a unique optimum and fixed bounded delays. Gap-weighted comparisons and experiments show when shared waiting helps and what balanced exploration costs.
Read the full abstract

Machine Learning

Learning through Capacity Awareness to Minimize Hospital Readmissions

Predictive algorithms are increasingly used to guide hospital discharge decisions, yet these models can exhibit censoring bias for vulnerable patients due to insufficient historical data or past discharge decisions. When biased readmission risk predictors are used to inform discharges, the training data generated by discharge decisions may perpetuate similar biases when predictors are re-trained, leading to a vicious cycle. We extend a queueing-based Markov decision process for hospital discharge to incorporate periodically retrained readmission risk predictors. Our key insight is that cost-optimal, "capacity-aware" operational decision policies can naturally diversify the lengths of stay at which patient outcomes are observed, enabling consistent estimation of readmission risk without explicit fairness constraints. We prove that repeated retraining under discharge-up-to policies converges to unbiased risk estimates under well-specification. In synthetic experiments and calibrated simulations based on patient-level data from a partner hospital, we find that where commonly used threshold-on-risk policies can remain persistently biased, discharge-up-to policies reduce learning error over status quo decision policies (such as thresholding on risk) by 24%-25%. These improvements to estimation also mitigate some forms of bias, reducing readmission cost disparities (adjusted for baseline risk differences between demographic groups) by 38%-63%. Our results demonstrate how capacity-aware decision policies can improve both predictive performance and fairness by addressing censoring bias.
Read the full abstract

All authors Lydia T. Liu, Kara Schechtman, Pengyi Shi, Angela Zhou

Numerical Analysis

A Legendre-Galerkin Method for Kinetic Models of Opinion Dynamics with Quasi-Invariant Consistency

Kinetic models of opinion dynamics provide a mesoscopic framework for describing collective behavior arising from interactions among a large number of individuals. Highly accurate numerical approximation to such Boltzmann-type equations, however, remains challenging due to the nonlinear inelastic collision operators involving multiple integrals coupling opinion variables and random perturbations. To address the difficulties, we develop a Legendre-Galerkin spectral method for a general class of kinetic models with opinion-dependent interaction rules. Our method exploits an exact decoupling of the weak collision kernel to obtain a time-independent tensor representation of the Galerkin collision operator. A recursive construction reduces collision-tensor assembly to one-dimensional quadrature, allowing the coefficients to be precomputed for subsequent use. A rigorous $L^2$-error estimate, along with mass preservation, is established for the fully discrete scheme under suitable regularity and stability assumptions. In the quasi-invariant regime, we derive the limiting Fokker-Planck equation and establish consistency between the exact Legendre-Galerkin operators of the rescaled Boltzmann equation and its limit for each fixed polynomial degree. Numerical experiments support the theoretical results, illustrate quasi-invariant consistency, and capture the formation of opinion clusters under the bounded-confidence interactions.
Read the full abstract

Audio and Speech Processing

CallCleaner: Lightweight Bioacoustic Denoising from an Audio Foundation Model

Many studies of animal vocalizations require clean recordings, but animals are usually recorded in the field, with background noise from weather, vehicles, and other animals. Animal vocalization denoising is challenged by the scarcity of clean training data and the extreme diversity of animal vocalizations. Classical spectral gates leave residual noise, especially without voice activity detection (VAD), while detector-free neural denoisers designed for bioacoustics remove less noise. We show that a general-purpose source separation foundation model, SAM-Audio, removes more noise than either of these types of methods, when used without training and prompted with time spans from a simple animal VAD. To turn this heavy method into a practical tool, we distill it into CallCleaner, a 19M-parameter denoiser that needs no detector and runs ~90x faster than real time on a laptop CPU. We find that SI-SDR has limited utility in our setting as it penalizes regenerated output and conflates noise removal with call fidelity, a trade-off our separate metrics reveal. CallCleaner leaves 11 dB less noise around the call than Biodenoising, the leading neural bioacoustic denoiser with public weights, and among fast, CPU-capable systems it pushes the Pareto frontier of noise removal vs. call fidelity, without requiring a detector. We release code and weights for CallCleaner, which enables fast, private, local denoising of single-caller recordings without specialized hardware.
Read the full abstract

General Economics

Optimal power expansion planning under uncertainty in sub-Saharan Africa: A stochastic dual dynamic programming approach

Universal electricity access and decarbonization are interlinked challenges for sub-Saharan Africa. Generation expansion planning methods used in the region rarely represent sequential and irreversible investment under demand and hydrological uncertainty. This paper develops a multistage stochastic generation expansion model solved with stochastic dual dynamic programming (SDDP) to address the problem. The model represents demand and country-specific Markov hydrological uncertainty, accounts for technology-specific construction lead times, and endogenous reliability. The model is applied to the Ethiopia and Zimbabwe power systems over 2025–2050. The results show that for Ethiopia, the least-cost pathway is hydropower, with wind and solar sources providing increasing diversification. The model reduces cumulative expected unserved energy by about 78 TWh relative to a lower-cost deterministic expected value investment plan. For Zimbabwe, the least-cost pathway expands solar, hydropower, and gas while retaining residual coal as inherited coal capacity retires. Additionally, carbon pricing reduces emissions and increases resource costs by about 3% to 5%, creating a gradual cost-emissions trade-off. Combining a carbon tax policy with concessional solar financing renders the lowest average annual emissions while keeping resource costs below the baseline, which solves the model without emission constraint. At approximately matched expected emissions, emission caps have higher net resource costs than the corresponding carbon tax scenarios. The overall results imply that adaptive investment model can reduce power shortage exposure, while carbon policy design and lower clean investment costs shape the cost of decarbonization in developing power systems.
Read the full abstract

All authors Arega Getaneh Abate, Xiao-Bing Zhang, Xiufeng Liu, Ruyu Liu, Per Sieverts Nielsen

Robotics

Beyond Memory Limits: A Hierarchical Spatio-Temporal Map Representation for Large-Scale 3D Autonomous Exploration

Autonomous UAV exploration is increasingly used for large-scale surveying and inspection in inaccessible or disaster-affected environments, where onboard memory and computational resources are often severely constrained. Most existing UAV exploration systems maintain globally queryable geometric map representations for motion planning and exploration decision-making, causing memory consumption and spatial-query costs to increase with the explored area. Robot-centric local mapping bounds dense-map memory and computation but discards global task context needed for long-horizon exploration.We present FLAVOR, a hierarchical spatio-temporal framework that confines dense metric geometry to a robot-centric local map while retaining compact global observation and task context. At the observation level, lightweight LiDAR-keyframe visibility records preserve ray-wise evidence for frontier reasoning. At the task level, unresolved frontiers beyond the local-map extent are retained as compact pending tasks associated with odometry anchors. A sparse topological graph connects these anchors through retained routes for future revisits.Upon returning to a task region, FLAVOR reconstructs the frontier and exploration viewpoint from observations, while the topological route is instantiated and validated against the current local map before execution. Dense-map capacity remains fixed as exploration expands, with exploration-dependent memory growth confined to compact observation and task context.FLAVOR achieves 31.98–41.88% shorter exploration time and a 94–99% smaller map memory footprint than the best baseline that completes all three simulated scenes. In real-world deployments, it reduces this footprint by 92.67% in the forest and 62.42% at the construction site relative to an explicit point-cloud ikd-tree while preserving long-term task continuity under constrained onboard resources.
Read the full abstract

All authors Yizhe Zhang, Jianping Li, Yilun Gao, Liangliang Yin, Pengfei Wan, Bisheng Yang

Computational Complexity

A note on exact external information complexity

We study the prior-free zero-error private-coin external information complexity $B(f)$ of finite two-party functions, with correctness on every input. An elementary rectangle/entropy argument gives $B(\mathsf{AND})=\log_2 3$. A complete classification of Boolean $3\times3$ matrices leaves, up to input permutations and output complementation, only the two-overlapping-rectangles function $F$ without an exact value. An entropy potential and a finite rational protocol give $2.1212661667060486842\ldots\le B(F)<2.130998318$. We also determine its deterministic external information and zero-error communication costs. A separate product-prior certificate and internal-information additivity yield $R_0^{\mathrm{pub}}(F^{\times t})\ge2.021991187178562\ldots\,t$, strictly exceeding the fractional-cover rate $2t$.
Read the full abstract

Computation and Language

THEATRON: A Benchmark for Evaluating Persona Persistence in Long-Horizon LLM Decision-Making

Existing evaluations often test persona expression rather than whether persona information persists in extended decision-making. We introduce THEATRON, a controlled benchmark that uses financial replay to evaluate persona persistence in long-horizon sequential decision-making. In these environments, models conditioned on fixed behavioral-finance personas observe shared daily market evidence and trajectory-specific account-state feedback and make sequential decisions over multi-month trajectories. We evaluate Zero-Shot, One-Shot, and Chain-of-Thought persona injection across four LLMs, using cumulative-return and action trajectories to diagnose within-persona stability, cross-persona separability, and action collapse. Persona conditioning produces systematic behavioral shifts beyond a no-persona baseline, and persona-level separation generally exceeds the variation induced by semantically equivalent descriptions. Nevertheless, these shifts do not reliably form coherent persona-specific policies: apparently stable separation can arise from stereotyped action collapse, while diverse policies can exhibit noisy trajectory switching. These findings expose a gap between observable persona-induced behavior and persistent, context-sensitive persona-conditioned decision policies. Code and benchmark data are publicly available at: https://github.com/yuyan12138/theatron
Read the full abstract

All authors Yuyan Wang, Yuechen Jiang, Zhiwei Liu, Jimin Huang, Sophia Ananiadou

Software Engineering

MicroSim: Contrastive Graph Neural Networks for Binary Code Similarity Detection

We present MicroSim, a system for binary code similarity detection that represents each function as a graph built from IDA Pro Hex-Rays microcode and compares such graphs with a contrastive Siamese graph neural network (GNN). We evaluate on the TREX cross-optimization benchmark, a corpus of 849,072 function-level microcode graphs spanning optimization levels -O0 through -O3, under a strict class-disjoint split in which no function key is shared between train and test. Under this protocol, MicroSim reaches 0.9972 area under the receiver operating characteristic curve (ROC-AUC) and 0.9762 Matthews correlation coefficient (MCC), outperforming Gemini (0.9332 ROC-AUC), SAFE (0.9752), and jTrans (0.9740) retrained on the same split, as well as a structure-free Bag-of-Words (BoW) baseline (0.9484). We analyze the gap between pair classification and ranked retrieval in graph transformers, showing that Cross-domain Similarity Local Scaling (CSLS) under unified tie-resolved evaluation raises GraphGPS mean reciprocal rank (MRR) from 0.2542 to 0.2559 (R@1 = 0.1639, R@10 = 0.4561), strictly satisfying the theoretical MRR upper bound. Finally, over 5 seeds with equal compute, the pairwise cosine-embedding loss attains a slightly higher ROC-AUC ({\Delta} = -0.0256, 95% confidence interval [-0.3282, +0.2770]), whereas the supervised-contrastive information noise-contrastive estimation (InfoNCE) loss is significantly better on thresholded classification and ranked retrieval.
Read the full abstract

All authors Moussa Moussaoui, Tarik Houichime, Abdelghani Souhar, Abdelalim Sadiq

Probability

Sitewise Generator Decomposition and Non-Equilibrium Steady States for Generalized Zero-Range Processes

In this paper we study the generalized zero-range processes considered in [20, 11, 12] with boundary driving. In these processes mass is moved between neighboring sites with rates depending only on the departure site. For a wide class of such models, the equilibrium stationary states are known to be of product form. We show that for such models, the boundary-driven generator can be decomposed into a sum of generators acting on single sites when integrating against (possibly inhomogeneous) products of the equilibrium marginals. We exploit this structure to derive new results for the harmonic models introduced in [15]. In particular, we prove that the propagation of mixture measures in [19] is still valid on graphs with asymmetric edges. On the chain geometry with asymmetric edges, we identify a class of models for which the NESS is given precisely by the NESS of the symmetric harmonic model on the chain. In addition to these results for harmonic models, we also present a new class of asymmetric models on the chain for which the NESS can be found explicitly and exhibits shock behavior. The large deviations for the density profile corresponding to these NESS are stated and proved.
Read the full abstract

Mathematical Physics

The Lee-Yang property of the Blume-Emery-Griffiths model

The Lieb-Sokal theorem provides a tool for establishing the Lee-Yang property of the partition function of a spin model if the corresponding single-spin partition function has this property and the interaction term in its Hamiltonian is of the form $-\sum J_{ij} S_iS_j$ with $J_{ij}\geq 0$. According to the result of this paper, the spin $S=1$ model with the Hamiltonian $-\sum (J_{ij} S_iS_j + K_{ij}S_i^2S_j^2) + \sum \Delta S_i^2$ has the Lee-Yang property for all $K_{ij}\in \mathds{R}$ and $\Delta >0$, whenever each $J_{ij}$ exceeds an explicitly computed bound $C(\beta, \Delta, K_{ij})>0$, such that $C(\beta, \Delta, K_{ij})\to + \infty$ as $\beta \to 0$. The latter suggests that the Lee-Yang property of this model may hold only for sufficiently low temperatures.
Read the full abstract

Statistical Methodology

Causal Machine Learning for Multilevel Studies with Unmeasured Cluster-Level Covariates via Variational Autoencoders

Unmeasured confounders are a key obstacle to accurately estimating treatment effects in observational studies, and in multilevel data, they often arise at the cluster level. Flexible machine learning methods cannot directly resolve such unmeasured confounding. To address this, we propose the variational autoencoder-augmented R-learner (VAE-R), which leverages variational autoencoders (VAEs) to infer unmeasured cluster-level covariates from the within-cluster dependence of outcome and treatment residuals, and uses the resulting adjustments to estimate the conditional average treatment effect (CATE) and the average treatment effect (ATE). The VAE decoder is specified through an offset likelihood so that the latent variables capture only the cluster-level variation left unexplained by measured covariates. VAE-R accommodates unmeasured cluster-level confounders in observational studies and unmeasured cluster-level predictors of the outcome in multisite randomized trials. We also provide data-driven strategies for selecting the two hyperparameters of the VAE, the latent dimension size and the Kullback-Leibler weight. Our simulation study shows that VAE-R nearly eliminates the bias of the ordinary R-learner's CATE and ATE estimates in multilevel observational data and improves their precision in multisite randomized trials. Finally, we illustrate VAE-R with two datasets, one from a multilevel observational study and the other from a multisite randomized trial.
Read the full abstract

Metric Geometry

Dimension reduction by bi-Hölder maps

The Hölder dimension of a metric space is the infimum of its Hausdorff dimension over all bi-Hölder changes of the metric. Colvin asked whether it can take noninteger values. We show that every metric space of finite upper box dimension has integer Hölder dimension. For a bounded subset of a Euclidean space, the dimension reduction is carried out by a global homeomorphism of the ambient space, built from randomly translated singular functions, which brings the Hausdorff dimension of the set arbitrarily close to its integer part.
Read the full abstract

Machine Learning

One-Shot Hierarchical Federated Clustering for Distributed Non-IID Biomedical Data Analysis

In multi-institutional biomedical data analysis, comprehensive disease profiles are often fragmented across different data silos. Furthermore, local data distributions typically exhibit varying levels of granularity or even complex nested structures. Although Hierarchical Clustering (HC) is a powerful tool for exploring such data topologies, its cumbersome recursive process incurs prohibitive computational overhead and renders it highly susceptible to privacy leakage. This paper, therefore, introduces an efficient one-shot hierarchical FC framework that performs client-end (institutions) distribution exploration and server-end distribution aggregation through one-way prototype-level communication from clients to the server. Although the designed one-shot prototype transmission guarantees minimal computational overhead and mitigates privacy risks, it inevitably incurs information loss. To counteract this, a multi-granular competitive learning mechanism is proposed on the server to fuse the clusterlets, effectively maximizing the retention and mining of global distribution topologies under limited, one-way communication constraints. It turns out that the complex cluster distributions across clients can be efficiently explored, and extensive experiments comparing state-of-the-art methods on public datasets demonstrate the superiority of the proposed framework in balancing robust clustering, privacy preservation, and computational efficiency.
Read the full abstract

Information Retrieval

Beyond Ranking: Are Decision Models Calibrated Relevance Scorers?

Decision models such as Jev are a new kind of relevance scorer. Their typed score primitive takes a question, an ordered set of textual relevance levels and a state, and returns a probability for each level directly, rather than token likelihoods or generated labels. Jev's vendor promises that these probabilities are calibrated. This motivates revisiting 'Beyond Yes and No', which improved zero-shot pointwise LLM reranking using likelihood-derived scores over fine-grained relevance labels but left calibration against human assessors as future work. In this paper, we first reproduce its open-model FLAN-UL2 setting on eight BEIR collections, confirming its ranking trends, and then run Jev and two open-weight decision models with the same protocol. Jev and the 27B pplx-decider rerank ahead of FLAN-UL2 and three cross-encoders on average. We then compare Jev's native probabilities with the human assessors' labels and find them overconfident: of the DL21 pairs it scores above 0.9, only 66% are relevant. Since this is mostly an error of scale, a two-parameter calibration fitted on 2,257 DL19 pairs removes much of it and transfers to DL21. After calibration, Jev has a lower DL21 Brier score than FLAN-UL2 and the three cross-encoders, mainly because it separates relevant from non-relevant pairs better. Finally, with calibrated probabilities, choosing each query's cut-off by expected F1 needs no parameter of its own, beats a fixed top-10 cut and matches a cut-off threshold tuned on DL19.
Read the full abstract

All authors Shengyao Zhuang, Xueguang Ma, Matthias Petri, Jimmy Lin

Data Structures and Algorithms

Õptimal Facial Distance Labeling in Planar Graph

We show how to assign labels of size $\tilde O(n^{1/3})$ to the vertices of a face $F$ in an $n$-vertex unweighted undirected planar graph $G$ s.t one can deduce the distance between any two vertices of $F$ given their labels alone. This matches (up to log factors) the $\Omega(n^{1/3})$ lower bound of Gavoille, Peleg, Pérennes, Raz from 2004. Prior to our work, the smallest labels were of size $O(n^{1/2})$.
Read the full abstract

Nuclear Experiment

Precise measurements of the 3H/3He charge form factor ratio and their impact on the 3H-3He charge radius difference

Nuclear form factor measurements can be used to extract the nuclear charge radius. These can be compared to ab initio nuclear structure calculations, with the difference between 3H and 3He charge radii providing information on the spatial distribution of protons vs neutrons in these nuclei. Extractions of the 3H-3He charge radius difference are limited by the modest amount of data for tritium elastic scattering at low momentum transfer. The experiment made a precise measurement of the 3H/3He form factor ratio at low momentum transfer which is combined with the absolute 3He measurements to provide improved 3H data. This provides an improved determination of the absolute normalization corrections of world data sets, giving improved precision for the low-Q^2 measurements that are most sensitive to the radius. We extract the elastic cross sections and the 3H/3He ratio at Q^2=3.03 fm^{-2} and evaluate the impact of this point on a global fit of previous form factor extractions. We examine both the impact on the normalizations of the other data sets and the fit parameters, including the charge radius and its uncertainty. Our sensitivity study shows that the new precision data points constrain the normalization of previous data sets, reducing the uncertainties by up to 12% and yielding a 4% improvement in the radius. Including projected data based on an approved experiment yields improvements in the normalization factors of up to 20% and a 27% improvement in the 3H RMS charge radius extraction, allowing for an improved extraction of the 3H-3He radius difference from a direct fit to the form factor ratios over a range in Q^2.
Read the full abstract

All authors L. Kurbany, S. Li, S. N. Santiesteban, J. Arrington, D. Abrams, S. Alsalmi, D. Androic, K. Aniol, T. Averett, C. Ayerbe Gayoso, J. Bane, S. Barcus, J. Barrow, A. Beck, V. Bellini, H. Bhatt, D. Bhetuwal, D. Biswas, A. Camsonne, J. Castellanos, J. Chen, J-P. Chen, D. Chrisman, M. E. Christy, C. Clarke, S. Covrig, R. Cruz-Torres, D. Day, D. Dutta, E. Fuchey, C. Gal, F. Garibaldi, T. N. Gautam, T. Gogami, P. Gueye, T. J. Hague, J. O. Hansen, F. Hauenstein, W. Henry, D. W. Higinbotham, R. J. Holt, C. Hyde, K. Itabashi, M. Kaneta, A. Karki, A. T. Katramatou, C. E. Keppel, P. M. King, T. Kutz, N. Lashley-Colthirst, W. B. Li, H. Liu, N. Liyanage, E. Long, J. Mammei, P. Markowitz, R. E. McClellan, F. Meddi, D. Meekins, R. Michaels, M. Mihovilovic, A. Moyer, S. Nagao, D. Nguyen, M. Nycz, M. Olson, L. Ou, V. Owen, C. Palatchi, B. Pandey, A. Papadopoulou, S. Park, T. Petkovic, S. Premathilake, V. Punjabi, R. D. Ransome, P. E. Reimer, J. Reinhold, S. Riordan, V. M. Rodriguez, A. Schmidt, B. Schmookler, E. P. Segarra, A. Shahinyan, S. Sirca, K. Slifer, T. Su, R. Suleiman, L. Tang, Y. Tian, W. Tireman, F. Tortorici, Y. Toyama, K. Uehara, G. M. Urciuoli, D. Votaw, J. Williamson, B. Wojtsekhowski, S. Wood Z. H. Ye, J. Zhang, X. Zheng

Cryptography and Security

Robust and Private Swarm Learning via Zero-Knowledge Validator Admission and Consensus Aggregation

Swarm Learning improves resilience and removes the need for a single trusted coordinator by distributing model aggregation and validation across participating nodes. However, it also delegates the validation of model updates to peers whose private data may be biased, unrepresentative, or adversarial, and existing Swarm Learning protocols generally do not attest that data, leaving decentralized validation as an attack surface. We present a robust, privacy-preserving Swarm Learning framework that makes verifiable data quality the condition for admission to the learning task. A validator joins only after a zero-knowledge attestation proves, without revealing any raw sample, that a reference model ran correctly on its private data. The attested outputs also yield a label-free class profile, so the network assembles a collectively IID validator pool from individually skewed validators as accurately as an oracle with the true labels (total-variation distance to uniform 0.025, against 0.021 for the oracle and 0.21 for random selection). Admitted validators score client updates and, through a Tendermint-style Byzantine Fault Tolerant consensus, agree on a gradient mask and a quality-derived aggregation weight for each submission, replacing dataset-size weighting with a quality-aware signal. Adapting the cut-layer idea of Split Learning, a challenge-based audit lets peers re-verify that a mask was computed on the attested data without an expensive end-to-end zero-knowledge proof. Across four benchmarks under non-IID partitions the framework matches FedAvg on clean data; under a gradient-scaling attack that collapses FedAvg to random chance, it retains 94% of clean accuracy on CIFAR-100 and over 99% on FashionMNIST, where FedAvg achieves only 2.1% and 11%. The result is a private, robust, and auditable foundation for decentralized learning.
Read the full abstract

Computational Complexity

On the Complexity of Extending Prescribed Linear Forests to Hamiltonian Cycles in Hypercubes

Let $Q_n$ be the $n$-dimensional hypercube, and let $P\subseteq E(Q_n)$ be a set of prescribed edges. It is known that, for $n\ge 2$, every such set $P$ with $|P|\leq 2n-3$ whose edges form a linear forest is contained in a Hamiltonian cycle of $Q_n$. This bound is sharp for general linear forests. The computational complexity of deciding whether a prescribed edge set of a hypercube can be extended to a Hamiltonian cycle was posed as an open question by Dvořák [SIAM J. Discrete Math. 19 (2005) 135–144]. We prove that the problem is NP-complete even when the prescribed edges form a spanning linear forest with no isolated vertices. We give a polynomial-time reduction from a restricted 3-SAT problem, using local spanning path-forest gadgets for variables and clauses that are uniformly embedded in the hypercube. The same reduction therefore establishes NP-hardness of two versions of Hamiltonian-cycle extension problems for general prescribed edge sets.
Read the full abstract

Hardware Architecture

CU-3D: A Scalable Legality-Aware RTL-to-GDS Flow for 3D-IC Physical Design

Face-to-face (F2F) hybrid bonds shorten interdie connections, but data, clock, and power-grid (PG) connections must be realized together. CU-3D is a RTL-to-GDS physical design flow based on commercial EDA tools. It derives role-compatible hybrid-bond (HB) sites from a 17-metal ASAP7-based technology model, complete HB footprints, and constructed PG geometry. To coordinate partitioning with placement flexibility, Scaled refines an inherited 2D placement, whereas Fresh places the receiving die first and uses its pin locations and clock loads to guide placement of the other die. Data and clock HB assignments are jointly optimized while retaining alternative data sites and reserving capacity for subsequent PG assignment. The selected HB sites and stacks are preserved during 3D clock-tree synthesis, supply-aware cell insertion, and routing. Across eleven designs, including a 133-macro processor, the flow achieves zero backend DRC violations and passes signal-connectivity checks. Nine paired designs admit lower-footprint, lower-power implementations; four improve setup WNS. On JPEG, early joint planning gains 87 ps and mobility-aware partitioning gains 346 ps. Path analysis explains why cell-delay savings can outweigh net-delay increases.
Read the full abstract

All authors Mingjun Wang, Guanqi Li, Yihan Wen, Lancheng Zou, Mingjun Li, Feng Gu, Fengrui Liu, Jianan Mu, Yuxuan Zhao, Tsung-Yi Ho, Huawei Li, Sung Kyu Lim, Bei Yu

High Energy Physics (Experiment)

Machine learning assisted $\nu/\bar{\nu}$ discrimination in atmospheric neutrino events for CPV sensitivity in a 20-kton liquid scintillator detector

Atmospheric neutrinos span a broad range of energies and baselines, offering a complementary probe of leptonic CP violation. However, a JUNO-like liquid scintillator detector without a magnetic field cannot directly separate neutrino and antineutrino interactions, reducing CP violation sensitivity. This study investigates whether machine learning based statistical separation can improve CP violation sensitivity in a 20-kton detector. $\nu_{\mu}$ and $\nu_e$ charged current events are generated with GENIE (using Honda flux) in a JUNO-like detector geometry and are reweighted using OscProb three flavor oscillation probabilities incorporating Earth matter effects and NuFIT parameters. Detector response is approximated by smearing lepton and hadronic energies, neutrino direction, and neutron multiplicity. Logistic regression, XGBoost, and multilayer perceptron (MLP) classifiers are trained on inelasticity, neutron multiplicity, and charged lepton energy, with XGBoost selected for the analysis. CP violation sensitivity is evaluated for a 200 kton-year exposure using a Poisson likelihood with systematic nuisance parameters and marginalization over five oscillation parameters within their respective $3\sigma$ ranges. XGBoost achieves an area under the receiver operating characteristic curve (AUC) of approximately 0.83 for both flavor samples. Under the square root approximation, peak CP violation sensitivity improves from 1.49 $\sigma$ to 1.96 $\sigma$ for normal mass ordering and from 1.14 $\sigma$ to 1.61 $\sigma$ for inverted mass ordering. Machine learning based statistical neutrino-antineutrino separation can recover part of the CP violation sensitivity lost in charge blind detectors. These preliminary results motivate further studies incorporating validated detector response and systematic uncertainties.
Read the full abstract

All authors Dek Hao Liew, Mohammad Nizam, Shu Tian Eu, Siew Yan Hoh

Signal Processing

Radiation-Pattern-Based Channel Estimation for Reconfigurable Pixel Antennas

Reconfigurable pixel antennas provide spatial diversity through multiple radiation patterns generated by a single antenna aperture, but exploiting this diversity can require substantial pilot overhead for exhaustive channel estimation. This paper jointly considers beam-steerable pixel-antenna design and its communication-side channel estimation. A pixel antenna is designed to generate diverse radiation patterns, whose electromagnetic correlation is then exploited to interpolate unobserved channel responses from a limited number of measurements. The proposed approach is evaluated in terms of SNR gain and NMSE, demonstrating that effective pattern selection can be achieved with significantly reduced channel-observation overhead and can even outperform exhaustive scanning at low SNR.
Read the full abstract

All authors Daniel Dinis, Leo M. Lahti, Pasi Ylä-Oijala, Risto Wichman

High Energy Physics (Phenomenology)

Short-range dynamics of $X(3872)$ and $T_{cc}(3875)^+$ in a coupled-channel framework

We study the short-range dynamics in the coupled-channel description of the $S$-wave $D \bar{D}^*$ and $DD^*$ systems for $X(3872)$ and $T_{cc}(3875)^+$, respectively. Contact interactions are related through scalar- and vector-meson exchange saturation, while pion exchange is included explicitly with its local term subtracted. The coupled Lippmann-Schwinger equations incorporate the self-energy of the unstable $D^*$ mesons, thereby accounting for the three-body $D \bar{D} \pi$ ($D D\pi$) dynamics. Fixing the contact interaction from the $T_{cc}(3875)^+$ pole, we trace the predicted $X(3872)$ pole as the local magnetic dipole term from vector-meson exchange is varied. Retaining this term in full produces the shallowest poles, which remain approximately $5$–$12~\mathrm{MeV}$ below the observed $X(3872)$ mass. It is found that the pion exchange has a moderate effect on the binding energy and a stronger effect on the isospin violation for the $X(3872)$ state. The residual tension between the pure-molecular bound descriptions of $X(3872)$ and $T_{cc}(3875)^+$ is analyzed with the Feshbach projection. This analysis indicates a $DD^*$ interaction weaker than that implied by the purely molecular fits, and shows that the $T_{cc}(3875)^+$ state instead arises from the mixing of the $DD^{*}$ channel with a compact $cc\bar{u}\bar{d}$ state. We further examine the spatial structure using Born–Oppenheimer effective field theory. For separation radii of $0.5$–$1.5~\mathrm{fm}$, the combined charmonium and short-distance tetraquark weight in $X(3872)$ is $3.0$–$9.7\%$, while the short-distance weight in $T_{cc}(3875)^+$ is $1.3$–$14.4\%$. These results favor retaining the local vector-exchange contribution within the saturation framework and motivate a joint treatment of molecular and compact dynamics for the $T_{cc}(3875)^+$.
Read the full abstract

Computation and Language

Saidzoda-CA-EmbeddingGemma-300M: One Small Embedder for Tajik, Kazakh, Uzbek and Kyrgyz

The small open embedder EmbeddingGemma-300M, used as released, finds the right passage first for only 51.7% to 60.8% of Belebele questions in Tajik, Kazakh, Uzbek and Kyrgyz. We present Saidzoda-CA-EmbeddingGemma-300M, one model for all four languages. It is our Tajik fine-tune of EmbeddingGemma-300M, trained for one more epoch on 279,904 pairs: questions in Kazakh, Uzbek and Kyrgyz that our own language model wrote about source passages, a Tajik refresher, and English and Russian replay. On Belebele, recall@1 rises by 14.4 to 26.0 points in each language, to between 75.2% and 77.7%. Against single-language models trained with the same settings, the one model stays within 2.8 points at recall@1 in Kazakh, Uzbek and Tajik, with overlapping recall@10 intervals, though it also had more training. It keeps the Tajik law recall@10 of its starting point (92.2 against 92.4) but loses 4.4 points of law recall@1. A model trained on one of these languages also improves the others on Belebele, but less than training on them. The model still trails the 560M XLM-RoBERTa embedders on Kazakh, Uzbek and Kyrgyz Belebele: the better of multilingual-e5-large-instruct and BGE-M3, used as released, is 9.4 to 10.2 points ahead at recall@1. Our scores of those released models agree with the published MTEB results within 2.4 points. Russian and English recall@1 fall by 3.0 and 4.2 points.
Read the full abstract

All authors Tohir Saidzoda, Nekruz Maxudov, Bedil Karimov, Abubakr Abdullozoda, Amin Jalilov, Jonibek Kishvarov, Rahmonali Sharipov, Mariyamkhon Shodyeva

Artificial Intelligence

GOPR-CAD: Graph-of-Operations Progressive Reasoning for CAD Generation

Generating accurate CAD models requires substantial engineering expertise and careful control of geometry, topology, and mechanical constraints. Text-to-CAD aims to simplify this process, but current LLM-based methods often generate CAD programs directly. Language models remain unreliable when producing long code while coordinating coupled geometric, topological, and dimensional constraints. Consequently, executable programs may still produce structurally incorrect models, which execution rate and rendered appearance may fail to reveal. To address these limitations, we propose GOPR-CAD, an agentic Text-to-CAD pipeline based on Graph-of-Operations Progressive Reasoning. A language model acts as a decision agent over encapsulated CAD operations and progressively constructs an executable operation graph, which a deterministic compiler converts into the final CAD model. This separation requires no additional training and supports different language models. To evaluate structural fidelity, we construct MHCAD, a benchmark of 300 standard mechanical part requirements derived from Machinery's Handbook. Their explicit structures enable reference-free STEP/B-rep evaluation. GOPR-CAD achieves 82.7% Structural Yield on MHCAD, and 87.2% Structural Yield on an external CADPrompt subset.
Read the full abstract

All authors Yitian Pang, Yifan Li, Xiaowei Yue, Li Zheng