Papers

A promising route to superconducting artificial neural networks is a hybrid digital-Analog architecture that combines digital single-flux-quantum (SFQ) communication with compact Analog nonlinear processing. The study focused on the dynamic conversion of a discrete signal passing through a digital-to-Analog-to-digital (DAD) converter, in which the role of the Analog cell was performed by a $\Sigma$-neuron with a nonlinear transfer function – the basic cell of perceptron-like neural networks. Furthermore, the DAD converter, the elementary functional block of the hybrid architecture, combines a digital-to-analog converter (DAC) and an Analog-to-digital converter (ADC), and re-encodes the Analog $\Sigma$-neuron waveforms as an SFQ pulse sequence. Circuit-level simulations demonstrate how input values encoded by SFQ pulse trains are converted into analog signal levels, transformed by the $\Sigma$-neuron, and mapped back to pulse-based outputs. As a key experimental step, we fabricated and characterised a redesigned $\Sigma$-neuron and measured a sigmoid-like transfer characteristic suitable for activation-function implementation. The extracted response was incorporated into system-level simulations to assess the influence of realistic device parameters on the conversion process. We delineate the operating-range matching requirements for the DAC, neuron, and ADC blocks, supporting the feasibility of the proposed interface as a building block for perceptron-like superconducting neural networks with digital inputs and outputs. Finally, we developed two perceptron networks, one using a mathematical sigmoid activation and the other the measured $\Sigma$-neuron transfer characteristic, which reached classification accuracies of $97.0\%$ and $91.9\%$, respectively, on the MNIST handwritten digit dataset.

Kannan's algorithm, as analyzed by Hanrot and Stehlé in 2007, solves the exact Euclidean shortest vector problem in polynomial space and $n^{\frac{n}{2e}+o(n)}$ time. In the classical setting with polynomial space, we obtain the first improvement on this bound via a randomized algorithm that runs in $n^{\frac{n}{4e}+o(n)}$ time. The main idea is to represent a fixed shortest vector in many ways as a difference of samples, thereby enabling the low-space collision search of Lyu and Zhu (SODA 2023) to replace exhaustive enumeration in the original analysis.

We prove that no family of two continuous self-maps witnesses that the circle is a topological fractal, answering a question of Karasová and the present author. Since three maps are known to suffice, this bound is optimal. In contrast, for every $\varepsilon>0$ there are two continuous self-maps of the circle, depending on $\varepsilon$, whose images cover the circle and an integer $N$ such that every composition of $N$ of them has image of diameter less than $\varepsilon$. Thus two maps suffice at any prescribed scale, but no fixed pair works at all scales.

Tabla Stroke Transcription (TST) converts tabla audio into symbolic stroke sequences, but the scarcity of annotated recordings makes fully supervised training challenging. We propose a semi-supervised framework that uses sequence-level labelled and unlabelled tabla recordings. A teacher model generates pseudo-label sequences, while a Stroke-Level Confidence Estimation Model (S-CEM) estimates confidence for each predicted stroke. To improve training with uncertain pseudo-labels, we propose Confidence-Guided Markov Weighted Alternative Temporal Classification (CMW-ATC), which weights candidate sequences within uncertain spans using learned stroke transitions and reliable neighbouring strokes. Experiments across three TST evaluation settings show consistent improvements over conventional teacher–student training and Alternative Temporal Classification in its replacement form (ATC-R). Ablation studies further show the contributions of S-CEM, Markov transition weighting, and the use of reliable strokes on both sides of uncertain spans.

Model-free reinforcement learning can acquire contact-rich robotic manipulation skills through trial-and-error interaction, but it often requires the policy to learn both task strategy and low-level motion generation. In this setting, the action representation is critical because it determines how policy outputs are converted into robot motion, shaping both exploration and physical execution. Direct Cartesian command interfaces require the policy to generate motion at every decision step, coupling task-level adaptation with continuous low-level control and increasing the learning burden. We propose PA-RL, a reinforcement-learning framework that uses artificial potential fields as the action representation. Instead of commanding motion directly, the policy adapts the parameters of an energy-like potential field, which generates a state-dependent guidance direction executed through a Cartesian impedance controller. We evaluate PA-RL on peg-in-hole insertion, a representative contact-rich task with nonlinear dynamics and discontinuous contact transitions. In simulation, PA-RL is compared with Cartesian velocity, Cartesian pose, and variable-impedance action spaces using the same RL algorithm. PA-RL is the only method to reach a 100% evaluation success rate within the allotted training time, while the best baseline reaches 92.6%. It also reduces joint-torque variation by 55.4% and Cartesian acceleration variation by 70.8% relative to the best baseline, without explicit motion-quality penalties in the reward. The simulation-trained policy further completes 9/9 real-robot insertions without fine-tuning, demonstrating the deployment feasibility of the learned potential-field interface.

Despite a growing demand for causal evidence about social media, independent researchers remain severely constrained in their ability to conduct experiments directly on online platforms. To cope, multiple methodological workarounds have emerged - from controlled surveys and simulations to client-side overlays and platform partnerships - each requiring distinct trade-offs between desirable experimental properties. The recent emergence of open social media platforms offers a qualitatively different methodological opportunity. Here we propose a design space of social media experimentation and discuss Open Platform Field Experiments (OPFEs). OPFEs represent a distinct class of experimental approaches that enable independent researchers to directly intervene on functional platform components - such as clients, recommendation systems, and moderation services - within live social media environments. Through a comparative analysis of experimental archetypes, we show that OPFEs occupy a previously unexplored region of the design space. We then bridge theory and practice by characterizing the architectural and governance elements that enable OPFEs, mapping them onto Bluesky and the AT Protocol, and illustrating the end-to-end lifecycle of a complete OPFE design. Overall, this work establishes OPFEs as a practical methodological paradigm for independent, transparent, and ecologically grounded experimentation on open social media.

In this paper, we construct a Gelfand model for the Okada algebra $O_n(X,Y)$ with generic parameters $X$ and $Y$, on the space of symmetric Okada arc diagrams using a conjugation-type action. The model is constructed inductively by identifying the Okada algebra as a diagram algebra and using the Jones basic construction to obtain a tower of algebras that are themselves Okada algebras at lower levels. We use the model to obtain all the irreducible representations of $O_n(X,Y)$, indexed by the elements of rank $n$ of the Young–Fibonacci lattice, and identify them with the cell modules of $O_n(X,Y)$.

The pHz to sub-nHz GW regime remains largely unexplored but is crucial for mapping the early inspiral stage of SMBHBs and probing early-Universe physics. Astrometry and pulsar timing offer orthogonal and deeply complementary secular observables to investigate this frequency band. We aim to present the first joint data analysis combining real astrometric proper motions with binary pulsar timing, to search for and constrain continuous gravitational waves (CWs) sourced by SMBHBs in the ultra-low-frequency regime ($10^{-12} \le f_{GW} \le10^{-9} Hz$). We employ a Bayesian model selection and upper limit estimation framework to combine apparent proper motion displacements of $\sim 1.5 \times 10^6$ quasars from the Gaia CRF3 catalog with the line-of-sight orbital period derivatives ($\dot{P}_b$) of 11 high-precision binary pulsars. To prevent spurious detections, we heavily model instrumental and astrophysical systematics: we propagate the Galactic potential uncertainty for pulsars via Monte Carlo simulations and perform a Vector Spherical Harmonics (VSH) decomposition up to the octupole order (l=3) for quasars. We find no statistically significant evidence for a CWs signal in the joint analysis ($\ln B_{joint} = -0.42 \pm 0.03$). In the absence of detection, we set the tightest constraints to date on CW strain in the pHz band, yielding a 95% upper limit of $h_0\le6.4x10^{-11}$ at a reference frequency of $f_{ref} = 4 \times 10^{-10}$ Hz. The combined dataset achieves full sky coverage and improves single-dataset upper limits by 20%-30%. Combining orthogonal observables successfully breaks spatial degeneracies intrinsic to isolated searches. Furthermore, forecasts from inj.-rec. indicate that with the extended temporal baseline and reduced uncertainties of the upcoming Gaia DR4, this joint framework is poised to break the $h_0 < 10^{-11}$ upper limit barrier for sub-nHz CWs.

Recurrent Transformers increase computational depth through temporal recurrence, feeding each token's high-level hidden state into the computation of the next. This raises a natural question: is additional computation better spent on more temporal steps or greater physical depth? We investigate this question using Latent Recurrent Transformers (LRT), which retain one backbone forward pass per vocabulary token during decoding and provide a controlled setting for comparing these two ways of adding computation. We insert a latent thought token between consecutive vocabulary tokens, adding a hidden-state refinement step through the shared backbone before predicting the next token. On 16- and 20-layer NanoChat MoE backbones, we compare this model with double-depth vanilla Transformers and LRT, two-loop Transformers, and PonderLM-2 with one thought token. All configurations execute 2L blocks per vocabulary token during decoding. Under matched training FLOPs, LRT with one thought token outperforms PonderLM-2 and overtakes the two-loop Transformer as compute increases. Its gap to vanilla double-depth narrows from 0.014 to 0.001 bits per byte, despite approximately 46% fewer parameters and approximately half as many unique training tokens. At the training budget used for our main comparison, one thought token recovers an average of 89% of the gains from doubling LRT depth across the two backbone sizes, with approximately 48% fewer parameters. These results highlight temporal thinking as a parameter-efficient way to extend computation in recurrent Transformers.

The baryon cycle governs the exchange of gas, metals, and dust between galaxies and their environments, but simultaneous constraints on these constituents remain scarce beyond z~1. The ALMA Chemical Evolution (ACE) survey is a Cycle 11 ALMA Large Program designed to address this by studying the molecular gas, dust, and metal content of sub-solar metallicity at cosmic noon. ACE consists of CO(3-2) and dust continuum observations of 25 galaxies at z=2.0-2.5 with robust gas-phase metallicity measurements spanning ~35% to 83% of solar metallicity. The survey approximately doubles the number of unlensed main-sequence galaxies at z>1 with CO, dust-continuum and metallicity measurements and extends such studies to almost an order of magnitude lower stellar masses and metallicities than previous surveys. The ACE observations yield 17 CO detections and 17 Band 7 continuum (rest-frame ~270um) detections. CO detectability correlates most strongly with metallicity and stellar mass, while dust continuum detectability is more closely linked to star formation rate (SFR) and infrared luminosity. Using the ACE measurements, we derive a new empirical scaling relation linking CO(3-2) luminosity to stellar mass, SFR, and metallicity, providing a practical benchmark for estimating molecular gas content in low-mass, low-metallicity galaxies. The survey reveals several particularly intriguing systems, including some of the lowest-metallicity CO and dust detections currently known at z>1, galaxies with extreme gas and dust fractions, and systems exhibiting offsets between stellar, dust, and molecular gas emission. ACE provides the first comprehensive view of the interplay between molecular gas, dust, metals, and star formation in sub-solar metallicity galaxies at cosmic noon, enabling direct tests of models for baryon cycling, chemical enrichment, and dust evolution during the peak epoch of galaxy assembly.

Let $\mathcal A$ be a complete deterministic finite automaton on a state set $Q$ of size $n$ with $k$ letters, and for a proper nonempty subset $S$ of $Q$ let $\mathrm{minext}(S)$ be the length of a shortest word $u$ with $|Su^{-1}|>|S|$, where $Su^{-1}=\{q: q\cdot u\in S\}$. To each state $q$ attach the integer $\beta^{\ast}_q=\sum_{t=1}^{n-1}k^{\,n-1-t}(\mathrm{indeg}_t(q)-k^{t})$, where $\mathrm{indeg}_t(q)$ counts the pairs $(p,u)$ with $|u|=t$ and $p\cdot u=q$, and let $B(S)=\sum_{q\in S}\beta^{\ast}_q$. On every synchronizing automaton, $B(S)\ge0$ implies $\mathrm{minext}(S)\le n-1$, so, as $B(Q)=0$, one of $S$ and $Q\setminus S$ extends within $n-1$; when $B(S)>0$ no hypothesis is needed. Kari's Eulerian extension lemma is the case $\beta^{\ast}=0$, and $\beta^{\ast}$, like every member of the family $\sum_{t=1}^{n-1}c_t\sigma_t$, $c_t>0$, vanishes identically if and only if the automaton is Eulerian, where $\sigma_t(S)=\sum_{q\in S}(\mathrm{indeg}_t(q)-k^{t})$. On strongly connected automata $\sigma_t(S)/k^{t}$ has Cesàro limit $n\,e(S)/e(Q)-|S|$ for Friedman's weight $e$; that limit certifies singletons but no larger subset in general. The hypothesis $B(S)\ge0$ cannot be relaxed by one integer unit, nor can the constant $n-1$ be improved. A second-moment test on the sizes $|Su^{-1}|$ certifies 60 to 95 percent of the subsets with $B(S)<0$ at $n\le7$. Along non-Eulerian automata whose words of length $n-1$ merge a fraction of the state pairs bounded below, with $\max_q\mathrm{indeg}_{n-1}(q)=o(nk^{n-1})$, it certifies all but a vanishing share of them. The functional $B$ certifies half of the subsets outside $\{B=0\}$. At each subset size coprime to $n$ ($n\ge4$) some synchronizing Eulerian binary automaton attains the constant $n-1$; whether only there is open. No reset bound follows: Černý's automata have subsets not extending within $n-1$.

Cooperative multi-monostatic sensing enables sub-meter passive localization in integrated sensing and communication~(ISAC) networks by fusing base station~(BS) observations at a central unit~(CU). Existing studies, however, treat the fused data as ideally available at the CU, overlooking how the per-BS 3GPP functional split jointly constrains fronthaul bitrate, computational load, and whether coherent or non-coherent fusion is feasible at the CU. We propose the Split-Aware Joint Resource Allocation for Multi-Monostatic Sensing~(SARAMS) algorithm, which minimizes the worst-case multi-target squared position error bound~(SPEB) by jointly optimizing per-BS power, bandwidth, and observation time under fronthaul, computational, and power constraints. A split-dependent coefficient embeds the feasible fusion type into the Fisher information matrix~(FIM), casting SPEB minimization as a mixed-integer nonlinear program (MINLP) solved via a semidefinite program for power allocation and block coordinate descent (BCD) over a dominance-pruned configuration set. Simulation results demonstrate that SARAMS reduces the 90th-percentile worst-case SPEB by $65.7\%$ over equal power allocation while attaining an $8.7\%$ optimality gap relative to exhaustive search.

Given a cyclic code $C_g = \langle g(x) \rangle$ of odd length $m$ over $\mathbb{Z}_4$, one common way to build a quasi-cyclic (QC) code is to pick $f_1, \dots, f_\ell \in \mathbb{Z}_4[x]$ and let $C = \langle (f_1 g, \dots, f_\ell g) \rangle$. Because $\mathbb{Z}_4$ is not a field, the type of $C_g$ ($4^{k_1}2^{k_2}$) is not necessarily inherited by $C$. Determining conditions under which the type of $C_g$ is inherited by $C$ was posed as an open problem recently in \cite{AydinLuOnta2023}. In this paper, we settle this problem. We first give two sufficient conditions for the type to be preserved: one requires a single $f_i$ to be coprime to $x^m-1$ over $\mathbb{F}_2$, the other only requires the $f_i$ to be jointly coprime to it. Neither condition is necessary in general. Using the fact that $x^m-1$ is squarefree for odd $m$, we decompose $\mathbb{Z}_4[x]/\langle x^m-1\rangle$ into a product of finite chain rings using Chinese remainder theorem (CRT) and derive a condition on $f_1, \dots, f_\ell$ that is both necessary and sufficient for $C$ to match the type of $C_g$. This condition depends only on the irreducible factors of $x^m-1$ where $g$ does not already vanish. This also yields a simple test for when $C$ is a free $\mathbb{Z}_4$-module. Finally, we report many new QC codes over $\mathbb{Z}_4$, found by computer searches using Magma software~\cite{Magma1997} guided by this criterion, with Lee distances greater than previously known codes of the same type.

Access to academic support is a key determinant of student success, yet students experience it unequally: some readily seek help from lecturers or tutors, while others hesitate due to anxiety, fear of judgement, uncertainty about expectations, or low confidence in their understanding. This may be especially evident in computing education, where programming tasks are cumulative and cognitively demanding. Although students increasingly turn to general-purpose generative AI tools, these can produce responses that are inaccurate, insufficiently contextualised, or misaligned with module expectations. This study presents and evaluates Beacon, a course-specific Retrieval-Augmented Generation (RAG) system providing private, immediate, module-aligned academic support. Grounding responses in approved teaching materials, Beacon was designed to lower barriers to help-seeking while encouraging independent learning. Using a design-based research approach, Beacon was developed iteratively and evaluated via mixed methods, combining questionnaires and semi-structured interviews with students and staff at a Higher Education institution. Students described Beacon's responses as closely aligned with module content and more trustworthy than unrestricted generative AI tools, valuing its use of pseudocode and scaffolded explanations over direct solutions. Although participants remained cautious about trusting AI-generated responses without verification, they viewed the system as a valuable first point of support before consulting lecturers or official resources. The findings suggest that carefully designed course-specific AI systems may reduce barriers to academic support by occupying an intermediary space between independent study and formal support. Rather than replacing educators, educational AI may be most valuable when it broadens access to guidance while preserving the pedagogical role of lecturers.

Fraudulent job posting detection aims to identify job advertisements that are corrupted either through fake content, misleading information, or negative intent, disrupting the online eco-system of job-seekers and employers. Existing studies in this domain lack effective methods to simultaneously achieve high accuracy and meaningful structure of latent-space representations that capture subtleties among fake posts. To this end, we propose Centroid-Guided Contrastive Loss (CGCL), a loss function which unifies classification with densely formulated clustering to consistently reshape latent-space through a centroid-driven top-$k$ push-and-pull mechanism. The complementary nature of CGCL enables the model to enforce accurate decision boundaries and maintain high clustering compactness, effectively capturing both class separability and latent structure. Extensive experiments demonstrate the state-of-the-art (SOTA) performance of our method on EMSCAD, a public benchmark dataset. The code associated with this work is available at: https://github.com/ali-ahmed925/CGCL_code

In D-Wave quantum annealers, various properties of the ground state have been clarified, whereas the correspondence between theory and experiment for energy relaxation rates in multiqubit excited states remains insufficiently understood. Here, we measured the energy relaxation rates of the first excited states in single-qubit systems and interacting two-qubit systems using a D-Wave quantum annealer. We analyzed the measured relaxation rates using a Gorini-Kossakowski-Sudarshan-Lindblad master equation with independent local $\sigma^{\,\,z}$-type noise channels. The calculated relaxation rates reproduced the overall trends observed experimentally, supporting the model at a qualitative level. We then used the relaxation rates measured for the uncoupled single-qubit systems to calibrate the local relaxation parameters and predict the relaxation rates of the interacting two-qubit systems. The predicted rates remained within a factor of approximately four of the measured rates. These results show that the relaxation measurements on individual qubits can provide a practical estimate of relaxation in small interacting quantum systems and may help clarify relaxation mechanisms in programmable quantum annealers.

Monocular 6-DoF pose estimation of non-cooperative targets is important for on-orbit servicing and debris removal. A single-image estimator can confuse near-symmetric spacecraft orientations, and tracking can preserve an incorrect pose. We present Hypothesis-Anchored Tracking (HAT), a causal framework that uses inter-frame motion to select among competing CAD-based pose hypotheses before alignment and fusion. Rather than independently choosing the highest-scoring hypothesis in each image, HAT retains competing orientation histories and selects a pose to anchor the relative trajectory estimated by monocular SLAM. Sparse anchors and pose fusion provide per-frame estimates after initialization without revising past outputs. The method requires only a calibrated RGB sequence, a metric CAD model, and target image regions, which can be supplied by detection or segmentation. The pretrained pose and SLAM networks require no target-specific training or fine-tuning. We evaluate two versions, Mega-HAT and Pico-HAT, using MegaPose and PicoPose, on SPARK-2024, SwissCube and SHIRT, with YCB-Video assessing performance outside the space domain. Using one temporal configuration per method, the arithmetic means of the four dataset-wise comparisons show 9.4% lower mean pose error and 3.76 times the sustained input FPS for Mega-HAT relative to independent MegaPose, and 23.9% lower mean pose error and 2.42 times the FPS for Pico-HAT relative to independent PicoPose. Mega-HAT ablations on SPARK and an offline reference examine component contributions and the effect of revising past estimates.

We present a multi-scale kinematic study of the hub-filament system in Mon R2 using JCMT $^{13}$CO(3–2) and ALMA C$^{18}$O(1–0) observations to investigate how molecular gas converges toward the central hub and connects across cloud-to-core scales. The $^{13}$CO data reveal several velocity-coherent filaments with diverse position angles converging toward the central region hosting massive stars and stellar clusters, where distinct velocity components become spatially superposed. Two dominant velocity regimes at $\sim$8.8–10.2 and $\sim$10.4–11.8 km s$^{-1}$ are identified, with their velocity separation decreasing toward the central hub. The position–position–velocity and longitude–velocity maps of the $^{13}$CO emission exhibit a butterfly/X-shaped morphology, consistent with converging motions. A phenomenological rotating-and-infalling envelope model reproduces the longitude-velocity morphology, indicating that rotation and infall may contribute to the observed kinematics, although the model does not uniquely identify the underlying mechanism. At higher resolution, the ALMA C$^{18}$O observations resolve the molecular gas into multiple velocity-coherent filaments spanning $\sim$8–13 km s$^{-1}$, with distinct kinematic domains at $\sim$9–10.5, 10.5–11.5, and 12–13 km s$^{-1}$ converging toward the hub. The ALMA and JCMT position-velocity diagrams show consistent velocity structures across their respective spatial scales, with the ALMA-resolved features embedded within the broader molecular-gas velocity field traced by JCMT, demonstrating a multi-scale kinematic connection. These results support a hierarchical picture in which converging flows and filamentary accretion contribute to hub growth and ongoing star formation in Mon R2, while other dynamical processes may also contribute to the observed kinematics.

In-context learning using Large Language Models (LLMs) offers a compelling path to training-free post-OCR correction, yet its effectiveness for Devanagari script remains entirely unexplored. We present the first systematic evaluation of LLMs (3B-32B) for post-OCR correction in Hindi and Marathi, comparing three in-context example retrieval strategies: domain-random selection, dense semantic retrieval, and our proposed CharBM25, which retrieves examples by character n-gram BM25 similarity over OCR inputs to target shared error patterns with the test sentence. Across a 20,000-sentence benchmark spanning five news domains, retrieval strategy is the decisive factor in correction quality: CharBM25 outperforms domain-random selection by 2.8-4.0pp absolute WER on Hindi and 2.9-3.8pp on Marathi, using character trigrams, which consistently outperform bigrams and unigrams. Scale dominates performance: Gemma-3-27B achieves WER reductions of 55.0% for Hindi and 33.3% for Marathi under CharBM25-5. Few-shot gains are capacity-gated: models below 8B do not reliably improve over the OCR baseline, and on Marathi the smallest models (3B) degrade more sentences than they improve. Marathi is persistently harder to correct than Hindi across all scales, reflecting its greater morphological complexity. These findings establish CharBM25 as an effective, GPU-free retrieval strategy that matches or exceeds dense retrieval at negligible computational cost, and show that combining it with a general-purpose LLM of 12B+ parameters delivers reliable, training-free Devanagari post-OCR correction without task-specific fine-tuning. Dataset: https://huggingface.co/datasets/AbhishekBhandari/Devanagari-OCR-ICL-Benchmark

Declarative SPMD programming uses tensor sharding descriptions to drive distributed execution, separating parallelization from model code. However, the evaluated PyTorch DTensor stack dispatches every operator below autograd, incurring repeated dispatch and metadata costs, while lacking an inexpensive end-to-end validation path. Existing FSDP and distributed Muon implementations also mismatch two-tier supernode topologies: FSDP relies on explicit parameter packing and unpacking, and Muon's whole-matrix orthogonalization conflicts with parameter sharding. We observe that distributed tensors need only express sharding semantics at the tensor API boundary above autograd, allowing differentiation and kernels to operate on plain tensors. Based on this insight, we present HyperParallel-FSDP, featuring: (1) dual-mode DTensor execution, using one sharding plan for both a production mode with one-time layout resolution and no steady-state dispatch overhead, and a validation mode with end-to-end metadata propagation, fail-fast checks, and gradient-equivalence testing; (2) topology-aware FSDP, with zero-copy intra-supernode collectives, fused inter-supernode reduction, and a cross-layer backward pipeline that avoids waits on slow links; and (3) layout-driven distributed Muon, with sharding-derived communication groups, deduplicated orthogonalization, and shape-fused Newton-Schulz iterations. On Atlas 900 A3 SuperPoD, HyperParallel-FSDP scales from 16 dies to 384 cards (768 ranks), sustaining 421k tokens/s for a 505B-parameter MoE while FSDP communication uses 2.9% of step time. It reduces mean step time by 29.7% versus PyTorch FSDP2 and 25.5% versus Megatron DDP, with Pearson correlation above 0.999997 over 1,000 steps. Distributed Muon improves profiler step time by 5.4-16.0% over competing systems. Source code is available at https://atomgit.com/mindspore/hyper-parallel.

Four-dimensional scanning transmission electron microscopy (4D-STEM) records a two-dimensional diffraction pattern at each electron-probe position, yielding spatially resolved reciprocal-space information but large, heterogeneous data volumes. Here we describe 4D-ImageNet, a collection of 174,000 diffraction patterns comprising 145,000 experimental patterns selected from 29 acquisitions and 29,000 multislice simulations. The experimental data cover acquisition-level labels for Ag, Au, mixed Au-Ag, CoO, Pd and ZnO specimens across multiple fields of view, scan dimensions, camera lengths and exposure times. Each acquisition contributes 5,000 quality-ranked patterns with source scan coordinates and acquisition metadata. A set-prediction detector provides model-derived pseudo-labels for the direct-beam position and Bragg-disk centres, with a confidence score for each disk. The simulation data cover 13 crystal structures and include Euler rotations, reciprocal-space sampling and approximate low-index beam directions. A grouped mixed-domain masked-reconstruction benchmark is provided to assess leakage-resistant loading and evaluation across experimental and simulated data. The dataset is intended for representation learning, disk detection, diffraction-pattern retrieval, orientation analysis and simulation-to-experiment studies.

Surface Stokes drift depends strongly on the energy and directions of short waves, which are incompletely resolved by routine wave observations. We derive surface Stokes drift vectors from wave measurements collected by compact drifting buoys during three deployments in the North-East Atlantic and the Alboran Sea. The calculation uses vertical-acceleration spectra and first directional Fourier moments, which describe the mean wave direction and directional concentration at each frequency; it accounts for the Doppler shift caused by buoy motion relative to the water and adds a calibrated high-frequency tail above an intrinsic frequency of 0.7 Hz. Across 13,139 records, the median estimated speed is 0.081 m/s at a median wind speed of 6.8 m/s. Over the measured band of 0.04-1 Hz, accounting for wave directions reduces the magnitude by a median 39% relative to the unidirectional assumption. The median ratio of the parameterised tail magnitude above 0.7 Hz to the total estimated magnitude is 0.37. Comparisons with WAVEWATCH III and Copernicus Marine MFWAM show strong covariation and similar wind-dependent differences from the buoy-derived estimates. On the station-matched sample from the two Atlantic deployments, WAVEWATCH III directional spectra indicate that these differences within the compared band arise mainly from spectral levels rather than from net directional reduction. The observations provide constraints for model evaluation; the contribution of the unresolved short waves remains sensitive to the assumed spectral tail and its directional spreading.

The kissing number in dimension $d$ is the largest number of non-overlapping congruent spheres that can simultaneously touch a central sphere of the same size. We study dimensions $25$-$31$, where the best previous constructions are based on Leech lifting from the optimal kissing configuration in dimension $24$. Our method exploits the absence of contacts between the unlifted bulk and the block consisting of lifted and auxiliary vectors. Rotating this block while keeping the bulk fixed creates room for two antipodal points in dimensions $26$, $27$, and $28$, and one point in dimension $29$. Three further modifications yield improvements in dimensions $25$, $30$ and $31$: (a) a nonorthogonal diagonal linear deformation of the lifted block admits two antipodal points in dimension $25$; (b) rotating the additional coordinates of the lifted vectors and then applying a small orthogonal transformation to the resulting lifted block as a whole admits two antipodal points in dimension $30$; (c) rotating only the additional coordinates of the lifted vectors admits four nonantipodal points in dimension $31$. Together, these constructions yield the new lower bounds $\tau_{25}\geq 197058$, $\tau_{26}\geq 198552$, $\tau_{27}\geq 200046$, $\tau_{28}\geq 204522$, $\tau_{29}\geq 209497$, $\tau_{30}\ge 220442$, and $\tau_{31}\geq 238354$.

Accurate prediction of unsteady aerodynamic loads remains a major challenge in turbomachinery design. High-fidelity Computational Fluid Dynamics (CFD) simulations are expensive, while aeroelastic Quantities of Interest (QoI) depend sensitively on the temporal evolution of the pressure field. This work introduces periodic Fourier Neural Mapping (p-FNM), a neural-operator framework for predicting unsteady pressure distributions on turbine rotor blades simulated using the chorochronic numerical hypothesis. The architecture embeds temporal periodicity into the model and learns a continuous mapping from operating conditions and time to pressure fields. Unlike sequential latent-space approaches, p-FNM predicts pressure fields independently at any time, avoiding error accumulation while preserving temporal continuity. The model is evaluated on a database of unsteady rotor-blade simulations and compared with a reduced-order baseline based on a variational autoencoder and recurrent neural network, refered as the Temporal Prediction Model (TPM). Performance is assessed for pressure fields and Generalized Aerodynamic Forces (GAFs), the primary aeroelastic QoI. Across all training datasets, p-FNM consistently outperforms TPM. On the largest dataset, p-FNM achieves a pressure-field mean absolute percentage error of 0.46% and a GAF-magnitude prediction error of 4.42%, corresponding to improvements of 60.7% and 77.6%, respectively. The minimum weighted phase error reaches 0.060 rad, demonstrating accurate preservation of the temporal characteristics of the aerodynamic response. The results show that GAF prediction is more challenging than pressure-field prediction and that temporal coherence is critical for accurately predicting spectral aerodynamic quantities. These findings demonstrate the potential of periodic neural operators for reduced-order modeling and aeroelastic analysis in turbomachinery.

We construct Riesz sequences whose Beurling density is arbitrary close to the critical density in the setting of reproducing kernel Hilbert spaces on metric measure spaces. In particular, our results apply to exponential systems on unbounded sets, nonlocalized Gabor systems and general coherent systems arising from nilpotent Lie groups. The methods used in this paper are based on our previous work on the redundancy of frames and a selector form of Weaver's conjecture.

We investigate electrons subject to a magnetic field that may tunnel along a square lattice and additionally to two electronic leads. Using non-equilibrium Greens functions we evaluate the transmission and from that particle and energy currents and zero-frequency noise of the setup. We identify sweet spots for which the transmission reaches unity over broad frequency windows – carried by topologically protected boundary modes that are quite robust against local fluctuations of on-site energies. These transmission plateaus can be exploited for the implementation robust current standards, clocks and thermoelectric devices.

This paper proposes discrete-time approximations to rough continuous-time models of realized variance (RV). The leading rough models can be viewed as autoregressive processes driven by fractional Gaussian noise. We show that the Wold representation of this noise concentrates its dependence at the first lag when the Hurst parameter is below one half. Augmenting the autoregressive (AR) and heterogeneous autoregressive (HAR) models with a first-order moving-average (MA(1)) component therefore approximates the roughness, and the MA coefficient maps almost linearly into the Hurst parameter. We refer to these extensions as the "rough" AR and "rough" HAR models. Estimating them on the log RV of ten ETFs, we find negative MA coefficients for every asset, and the implied Hurst parameters align closely with the estimates from the continuous-time models. In the HAR literature, the negative MA(1) component is a significant feature that has been largely overlooked. In out-of-sample comparisons, the "rough" models outperform their classical counterparts for nearly every asset and horizon, with the largest gains at short horizons, and their accuracy is comparable to that of the rough continuous-time models but much easier to estimate by standard off-the-shelf software.

We investigate Floquet–spin-orbit compensation in the three-band $\alpha$-$T_3$ system, which interpolates between graphene at $\alpha=0$ and the dice lattice at $\alpha=1$. By tuning the interplay between an off-resonant circularly polarized optical field and intrinsic spin–orbit coupling, we identify three distinct pairwise band-degeneracy conditions in the quasienergy spectrum: (i) valley–spin degeneracy, (ii) spin–band degeneracy, and (iii) flat–dispersive band degeneracy. The flat band is not an artifact of a two-band reduction but rather an exact property of the continuum Hamiltonian at this compensation point, where the complete three-band Hamiltonian supports a momentum-independent dark state. A winding-two low-energy Hamiltonian and a finite-momentum Berry-curvature maximum, whose radius scales as the square root of the detuning, emerge from detuning this contact. We further show that the intrinsic transverse thermoelectric response encodes these spectrum patterns. The flat- and quadratic-band degeneracy can be predicted for $0<\alpha<1/2$ because the two response scales linked to the linear contacts have a monotonic ratio independent of the common spin–orbit energy scale. This allows an inverse determination of $\alpha$ and of the common spectral scale. The analytical results are confirmed by full three-band Kubo calculations, which also define the regime in which this inverse characterization is still observable. These results connect an experimentally accessible, Berry-curvature-sensitive thermoelectric response to tunable Floquet–spin–orbit band geometry.

The DeWitt boundary condition can be implemented in the Lorentzian path integral of minisuperspace quantum cosmology by summing only over non-singular paths. In this paper, we present a general definition of the resulting DeWitt kernels, along with specific rules for their construction, and evaluate them rigorously. The lapse integral over the positive half-line yields the Dirichlet Green function of the Wheeler–DeWitt operator, whereas the integral over the whole real line yields the Dirichlet group-averaging kernel; both vanish when either endpoint lies at the singularity. Since the minisuperspace actions are at most quadratic, all lapse integrals are performed in closed form without saddle-point approximations. We obtain the kernels for the closed universe without cosmological constant, for the flat de Sitter universe, and, in terms of Airy functions, for the closed and open de Sitter universes. Throughout, the DeWitt condition is imposed as the boundary condition at the singularity, which selects a self-adjoint constraint operator on the half-line and fixes the kernels uniquely, including their normalization.

Wall-climbing robots capable of scaling vertical surfaces could help automate hazardous or labor intensive tasks such as window washing, inspection, maintenance, and construction. Active adhesion methods achieve higher payload capacities, but require power to maintain their grip. Passive adhesion devices such as suction cups are an attractive option for such robots because they do not require power to maintain their grip, but they are limited by their payload capacity. This work presents a novel high-payload wall-climbing robot that utilizes passive bistable suction cups to generate adhesion without needing to be pushed into the wall. The robot features a track-based system that automatically engages and disengages bistable suction cups to achieve locomotion on smooth surfaces. The robot is able to achieve vertical wall climbing on glass, wood, metal, and painted surfaces, sideways and upside-down climbing, and is able to tow a payload of 7.940 kg (with a payload-to-weight ratio of 2.25).