Papers

Resolving how ensemble growth kinetics emerge from particle- and grain-level dynamics is essential for predictive control of nanocrystal synthesis. Here, time-resolved four-dimensional scanning transmission electron microscopy (4DSTEM), extended into five dimensions (5DSTEM) through continuous acquisition, is used to simultaneously resolve morphology, crystallographic orientation, and lattice strain during Au nanocrystal growth in aqueous HAuCl4. Although ensemble growth follows surface-reaction-limited kinetics, individual nanoparticles exhibit distinct pathways, including continuous growth and discrete coalescence through oriented attachment. These pathways display characteristic orientation dynamics, with stable or gradually selected orientations during continuous growth and abrupt reconfiguration followed by alignment during coalescence. At the ensemble level, the out-of-plane orientation distribution is established early and remains stable, whereas in-plane orientations remain broadly distributed without global alignment. Tensile strain develops progressively from particle surfaces and interfaces, directly linking lattice distortion with evolving morphology and crystallographic structure. By connecting ensemble kinetics with particle- and grain-resolved structural evolution, 5DSTEM provides a general framework for uncovering heterogeneous growth mechanisms and for relating synthesis pathways to structure-dependent properties in functional nanomaterials.

Extracting Gaussian information from data is well understood, but characterizing non-Gaussianity is challenging. We introduce HAWCS, a healpy-based Python package for efficiently computing wavelet scattering transform coefficients from full-sky maps. These coefficients provide compact summary statistics that are sensitive to higher-order structure and interactions across angular scales. We test the implementation on controlled Gaussian and non-Gaussian fields and demonstrate its computational efficiency. We then apply HAWCS to thermal Sunyaev-Zeldovich maps derived from Planck observations and four simulations, quantifying and comparing their higher-order statistical properties. We also use the method to expose the signatures of gravitational lensing in cosmic microwave background (CMB) temperature maps. These results demonstrate that HAWCS is a practical tool for validating component-separation pipelines and cosmological simulations, and a promising summary statistic for characterizing non-Gaussian structures beyond the power spectrum. These coefficients could also be used to constrain or generate simulated maps that reproduce the same statistical features of real data.

The monopulse technique determines the angle of a target by comparing signals from two narrow beams, yielding a precise angular estimate with low complexity. However, it struggles to resolve multiple closely spaced scatterers within the same resolution cell. Existing methods for estimating multiple scatterer angles involve complex signal processing and system modifications. We propose an effective method to estimate the angular centroid of scatterers using the mode of monopulse angle estimates. A semi-analytic expression for the angle estimate distribution is derived, confirming that its mode aligns with the centroid. To enhance estimation accuracy, we employ frequency diversity to reduce sample correlation. Numerical results validate the advantages of the proposed method, demonstrating superior performance over conventional techniques with low complexity.

Music creation often involves iterative refinement, changing selected musical details while retaining the rest. To support such refinement, we introduce SpanSynth-Edit, a flow-matching model for MIDI-guided synthesis and editing of multi-instrument audio mixtures using low-frame-rate scalar-quantised latents. MIDI Span encodes instrument-labelled note lifecycles as unordered event sets with continuous-valued attributes and pools each set into one conditioning vector per audio-latent frame. The model uses contextual audio for instrument-specific timbre guidance and supports editing by resynthesising the target region from revised MIDI. Experiments on single- and multi-instrument benchmarks show competitive performance and demonstrate within-frame onset control. We also discuss limitations of transcription-based note-adherence evaluation.

An unusual object in the sky raises two related questions: what kind of object it is and what physical process drives its motion. Departures from a gravity-only trajectory can provide evidence, but the same apparent acceleration may be generated by distinct dynamical mechanisms. We formulate this as an identification problem and establish conditions for identifiable uniqueness among three reference classes: inert dynamics, impulsive forcing, and active stabilization such as a controlled spacecraft. The distinction rests on how each class responds to disturbances. Although different mechanisms can produce similar average trajectories, the specified reference models imply different stochastic responses around those trajectories. We show when these differences uniquely determine the dynamical class from the observed trajectory and when several classes remain observationally compatible.

In this paper, we study local minimizers of the energy functional$J(\mathbf{u}) = \int_D (|\nabla \mathbf{u}|^2 + Q^2(x)\chi_{\Omega_{\mathbf{u}}})\,dx$ which give rise to a singular cooperative system. Here \(\mathbf{u}=(u_1,\dots,u_m): D\to\R_+^m\) is a vector-valued unknown function, \(\Omega_{\mathbf{u}}:=\{|\mathbf{u}|>0\}\) is the positive set, \(\chi_{\Omega_{\mathbf{u}}}\) is the characteristic function of the positive set \(\Omega_{\mathbf{u}}\), and \(Q(x)\) is the Bernoulli weight function. This problem was first introduced by Caffarelli, Shahgholian, and Yeressian in the poineer work (Duke Math. J. 167(10), 2018), where the regularity theory for minimizers and for the free boundary \(\partial\Omega_{\mathbf{u}}\) was established for nondegenerate Bernoulli weights, i.e., \(Q(x)\ge Q_{\rm min}>0\). The present paper studies the same free boundary problem but for a degenerate Bernoulli weight. The main difficulty lies in the coupling between the degeneracy of \(Q(x)\) and the vector-valued nature of the problem. We show that the the free boundary set \(\partial\Omega_{\mathbf{u}}\cap\{Q(x)=0\}\) decomposes into a nondegenerate part and a degenerate part. Nondegenerate points admit nontrivial blow-up limits; degenerate points have zero weighted density. We classify nondegenerate points into single-phase, multi-phase non-branching, and multi-phase branching points, analyzing their geometric structure. The branching-point singular structure for multi-phase points is purely vectorial and new in this setting. Finally, following Naber–Valtorta (Ann. Math. 185, 2017) and Edelen–Engelstein (Trans. Amer. Math. Soc. 371, 2019), we prove the nondegenerate set has locally finite $\mathcal{H}^{n-2}$-measure and is countably $(n-2)$-rectifiable.

Motivated by Eliahou's 64-modular Hadamard construction at the smallest open Hadamard order n=668, we introduce an explicit half-flip family of 32-modular Hadamard matrices at orders n=4L for L = 3 (mod 8). A master identity reduces the four-sequence Golay-quadruple condition under the half-flip ansatz (s, s*, sq, (sq)*) to a single-sequence type-restricted autocorrelation c_k^tau(s). The construction yields a closed-form expression for c_k^tau, a true Hadamard matrix at L=11, and 32-modular matrices at every L = 3 (mod 8), including the open orders n=716 and n=1132. Existence of 32-modular at L = 3 (mod 4) is due to Eliahou-Kervaire (2001); Eliahou's 2026 follow-up in J. Algebraic Combin. gives 64-modular matrices at L = 3 (mod 16) and L = 7 (mod 32) via the same ansatz and correlation identity we call the master identity, subsuming our construction on that residue subclass. Our contribution is therefore primarily structural. Applying Barrera Acevedo-O Cathain-Dietrich (2019) and Alvarez et al. (2020), the family is non-cocyclic over any group at every prime L in {11, 19, 59} in the YES set, yet pseudococyclic over the Goethals-Seidel Moufang loop GS_{4L} at every L. A half-flip H-set decomposition theorem parameterizes the symmetric difference of any two family elements by a single sequence flip set, giving the family the structure of a length-L Hamming cube. A symbolic mod-tower verifier (mod-8 is F_2-linear) classifies true Hadamards in the family through k=14 (L <= 115): the YES set is empirically bounded by k=7, refuting four H4 predictions and excluding L in {179, 283} within the ansatz. A Grobner basis at L=11 exhibits a previously unrecorded even-T0-block linear identity. All code and JSON certificates: github.com/michelkulhandjian/hadamard-halfflip-structural

Corporate default prediction is a core problem in financial risk management, yet traditional credit models rely heavily on financial statements that are often sparse or unavailable for many firms. Corporate transaction networks offer a complementary view of real economic activity, but how risk propagates through buyer-seller relationships remains underexplored. We conduct a large-scale empirical study using real-world electronic tax-invoice data spanning six years that links transaction histories with default events, revealing that transaction-driven risk is both role-dependent (buyer or seller) and scale-dependent. Based on these findings, we construct multiplex buyer-view and seller-view transaction networks and propose DefaultGNN, a dual-perspective graph neural network-based framework for corporate default prediction. DefaultGNN integrates both views to model how risk flows through transactional relationships, achieving strong improvements over both attribute-based and graph-based baselines, especially for firms with limited intrinsic risk signals. We further provide interpretable network-based explanations by visualizing how distressed trading partners contribute to default risk. In collaboration with a licensed credit rating agency, we validate that DefaultGNN's predictions complement existing credit scoring models, improving approval rates by 7-11%p without increasing default risk among approved firms. The source code can be found at https://github.com/jhkim611/DefaultGNN

Tabular foundation models learn to predict cell values in context, whereas world-model self-supervision asks for prediction in representation space (LeCun, 2022; Assran et al., 2023). On a tabular foundation-model prior, the latent term of a joint-embedding predictive architecture (JEPA) collapsed in our earlier runs and took the encoder with it to a constant map. We report a recipe under which the latent term survives to convergence beside the value objective: the value head reads the encoder field rather than the predictor, and the target is an exponential moving average (EMA) difference. To bound its cost against the value-only arm, both arms train until a plateau rule stops them, with no fixed step budget. A fixed horizon had confounded a slowdown with a ceiling, since the value-only arm was still improving well past the usual budget. At convergence, in one run per arm, the JEPA arm trails the value-only arm across 147 real datasets, 32:70 wins to losses on classification (29:63 with one entry per dataset name) and 8:24 on regression, the margin small on classification and wider on regression, and the count leans the same way in each stratum and each benchmark. The JEPA arm (jepa) needs 1.42 times as many steps as the value-only arm (ds), and 1.66 times its wall-clock, to reach its plateau.

We present a variety of possible observational sequences from planet detection to characterization (i.e., survey strategies) for the initial exoEarth detection and characterization survey phase of the Habitable Worlds Observatory. An array of survey strategies were collated by the HWO Exoplanet Survey Strategies Task Group; by investigating and comparing expected resultant exoEarth yields for multiple survey strategies, we can better understand the most efficient strategy to characterize multiple planets in unconventional ways. We present eight varied survey strategies, presented as steps of Detection, Photometry, Orbit Measurement, and/or Characterization, including discussion of possible shortcomings and benefits of each strategy. We present exoEarth yield calculations from the Altruistic Yield Optimizer (AYO) and the Exoplanet Open-Source Imaging Mission Simulator (EXOSIMS) for select strategies, with initial considerations of thermal emission and detector noise. Herein, a yield is the expectation value given the astrophysical assumptions of the number of planets that meet the observational criteria during the survey. We find that AYO and EXOSIMS find the highest yields for H2O at 0.9 microns, although the calculated yields vary. When allowing for wavelength optimization within AYO, we find that H2O is most optimal to observe at 0.9 microns when given the full range of VIS and NIR values, with an expected exoEarth yield of 29.1. H2O and O2 dual characterization is possible at the expense of yield planet loss, with a calculated possible yield of 20.9 in the VIS. The characterization of CH4 and CO2 in the NIR result in the lowest exoEarth yields at 19.9 and 6.4 possible expected planets, respectively. Future work updating AYO and EXOSIMS is necessary to explore further survey strategies that follow different observation procedures, some of which is currently underway.

We study $n$-torsionfree objects in abelian categories with enough projectives. Frobenius functors preserve $n$-torsionfreeness, and faithful ones reflect it. We prove that stabilization of the torsionfree filtration implies weak Gorensteinness. For Frobenius extensions satisfying a generator condition, we compare the terms of minimal injective resolutions and obtain transfer of Auslander-type conditions and of the Auslander–Gorenstein conjecture. We also compute a family of non-Gorenstein algebras whose torsionfree filtrations stabilize at level two and contain explicit nonprojective Gorenstein projective modules.

While single-view 3D reconstruction has seen significant progress, extrapolating complex 3D structures from inherently ambiguous 2D observations remains fundamentally ill-posed, particularly in the critically underexplored data-scarce regime. To address this challenge, we propose Point Diffusion Mamba (PDM), a method that integrates the generative power of diffusion models with the efficiency of state-space model for single-view 3D reconstruction under data-scarce conditions. Specifically, PDM employs a lightweight reconstruction module tailored to handle unordered point-cloud inputs effectively. By combining a Local Geometric Aggregation module with Mamba blocks, our approach jointly models global geometric structures and local details. In 3D reconstruction, each point in the initial noisy input requires a precise prediction, yet the high-level features extracted by the Mamba module capture only abstract semantic information from sparse points. To bridge this gap, we introduce the Hierarchical Feature Integration Network, which fuses high-level semantic and local geometric features for each point, overcoming the limitations of token-based point-cloud reconstruction. Furthermore, we propose a Dynamic Weighted Sampling strategy that adaptively unifies 3D generation with single-view reconstruction by leveraging generative priors to enhance reconstruction quality. Experimental results on the ShapeNet and Pix3D benchmarks demonstrate that PDM outperforms state-of-the-art methods, providing an effective solution for 3D reconstruction under data-scarce settings. Code is available at: https://github.com/NWUzhouwei/PDM.

Large language model (LLM) inference is constrained by the quadratic scaling of self-attention and the linear scaling of the KV cache, increasing latency, energy consumption, and GPU memory demand as context length scales. Existing soft-compression methods either lack query-guided memory selection at inference time, train without answer-targeted supervision, or couple compression tightly to a specific decoder architecture. We propose a Context-to-Answer-Aligned Memory Compression (CMC) framework, which compresses long input contexts into compact Context Memory Embeddings (CMEs) aligned to any frozen decoder's embedding space, reducing inference costs without modifying decoder weights. CMC introduces a two-tier KV cache that combines question-guided CME selection with a local context window, and trains the compressor with answer-targeted distillation from a frozen LLM. Experiments across nine encoder-decoder combinations and four QA benchmarks show that CMC consistently outperforms the baseline, achieving up to 7.3 EM and 4.0 F1 point gains on SQuAD, while reducing inference time and energy consumption by up to 20% and peak reserved GPU memory by up to 50% at 3,000 generation tokens. Ablation studies confirm that each architectural component and training objective contributes to the performance.

I connect two literatures developed independently: the path-measure formulation of non-equilibrium statistical mechanics, where a trajectory action decomposes into a time-antisymmetric (entropic) and time-symmetric (frenetic) sector, and the stochastic modelling of ribosomal traffic on messenger RNA. The biological target is a proposed intervention – antisense oligonucleotide (ASO) interference with wobble-uridine (U34) modification of transfer RNA – whose intended effect is not to abolish translation but to perturb its timing, driving ribosome collisions and collision-triggered transcript decay preferentially on high-flux, codon-biased transcripts. Dynamical-activity and large-deviation analyses of generic lattice exclusion models – notably the symmetric and totally asymmetric simple exclusion processes – are well established. To my knowledge their formalization specifically for ribosomal traffic queues, translation elongation, and collision-triggered no-go decay remains unoccupied; this paper addresses that narrower gap, not the general one. Two claims here are load-bearing and untested. First, selectivity: transcripts whose loss is therapeutically desirable are separable, by vulnerable-codon-pair burden, from transcripts whose loss is toxic. Second, non-redundancy: the frenetic decomposition yields predictions, specific to ribosomal queueing and collision-triggered decay, not already obtainable from rate-level exclusion-process models or existing activity/large-deviation analyses of exclusion processes. This paper establishes neither. It specifies both as falsifiable tests with pre-registered decision rules, including outcomes under which the framework should be abandoned or narrowed. It is a research programme proposal, not a result. No new experimental, computational, or bioinformatic results are reported.

Zero-Knowledge Proofs (ZKPs) enable a prover to cryptographically convince a verifier of the validity of a statement without revealing any underlying secrets, forming a foundational primitive for verifiable computation. The ZKP landscape is undergoing a fundamental shift from classic zk-SNARKs such as Groth16, which rely on trusted setup and are vulnerable to quantum adversaries, toward transparent, post-quantum constructions such as zk-STARK. These systems achieve post-quantum security by relying solely on collision-resistant hash functions, however, at the cost of substantial computational overhead. In particular, the Fast Reed–Solomon Interactive Oracle Proof of Proximity (FRI) protocol dominates prover complexity, generating massive data volumes, repeated Merkle-tree commitments, and irregular memory access patterns that limit performance and energy efficiency on general-purpose processors. To address these challenges, this work proposes FASTAR, a novel FPGA-based accelerator for the FRI protocol. Unlike accelerators that pursue fixed high-performance kernels on expensive ASIC process nodes, FASTAR adopts a constraint-driven design methodology. Our framework is implemented with High-Level Synthesis (HLS) and composed of fully parameterizable building blocks for the major stages of FRI, including polynomial evaluation, recursive split-and-fold, and Merkle-tree construction. From user-provided board specifications, FASTAR automatically generates hardware implementations tailored to the resource and memory constraints of the target FPGA, enabling deployment across a wide range of platforms without manual redesign.

Liquid scintillator detectors instrumented with distributed silicon photomultiplier (SiPM) arrays can be used in compact, topology-sensitive, and low-background experiments, but the temperature dependence of SiPMs immersed directly in the scintillation medium has not been widely characterized. We report the operation of a 125-liter linear-alkylbenzene-based liquid scintillator detector read out by 125 SiPM channels immersed in the active volume, over the range from room temperature to $-30\,^{\circ}\mathrm{C}$. The detector response was measured with cosmic-ray muons, including stopping muons followed by their Michel-electron decay. Cooling from $+15\,^{\circ}\mathrm{C}$ to $-30\,^{\circ}\mathrm{C}$ reduced the SiPM dark-count rate by a factor of 13.5, increased the single-photoelectron response by 49.1%, and increased the cosmic-ray muon light yield by 16.7%. The improved photoelectron separation and baseline stability at low temperature enabled a selection of stopping-muon events, from which the effective muon lifetime was measured to be $1959\pm132\,\mathrm{ns}$, consistent with the value expected for a hydrocarbon scintillator once $\mu^{-}$ capture on carbon is taken into account.

To improve understanding of how precipitation frequency and intensity change with warming, we combine the coupled land–atmosphere water balance with a stochastic hydrological model that retains explicit dependence on precipitation frequency and event depth. We find a linear relationship between co-variations of rainfall frequency and intensity, where the slope is a hydrological sensitivity determined by the dryness and storage indices and the intercept is a forcing term associated with changes in moisture convergence, potential evapotranspiration, and effective water-storage capacity. While a given relative change in either rainfall frequency or intensity produces an identical relative change in the mean rate, rainfall intensity additionally influences the runoff coefficient by altering terrestrial water storage relative to rainfall depth. This means frequency and intensity have different impacts on the water cycle, allowing the frequency–intensity slope to depart from -1. Analysis of long-term MOPEX records shows that the forcing term is dominated by moisture convergence and that the frequency–intensity relation becomes steeper from wet to dry regimes, consistent with the slope predicted by the theory. The framework provides a physical interpretation of the observed negative covariation between precipitation frequency and intensity and clarifies the roles of hydrological partitioning and atmospheric moisture supply in their responses to warming.

Liquid scintillator detectors instrumented with photosensors inside the scintillation volume preserve local optical information that is largely lost in conventional boundary-readout geometries. We demonstrate that this information is sufficient for charged-particle topology reconstruction using a sparse three-dimensional lattice of silicon photomultipliers. After validating the Geant4 detector response against measured photon-count distributions, a simulation-trained, time-informed convolutional neural network reconstructs the entry and exit points of through-going muons with median residuals of 1.91~cm and 2.39~cm, respectively. The reconstructed endpoints are geometrically consistent with acceptance regions defined by external trigger counters in cosmic-ray muon data. The same framework also reconstructs the production vertices of simulated positron starting-track events with a median residual of about 4.5~cm. These results establish the feasibility of topology-sensitive reconstruction using sparse in-liquid photosensor arrays in homogeneous liquid scintillator detectors.

While Python is designed as a cross-platform language, real-world applications encounter portability failures when deployed across different operating systems. We present the first large-scale empirical study of cross-OS portability issues in Python, analyzing 2,042 open-source repositories using two complementary approaches: systematic cross-OS test reexecution and manual analysis of GitHub issues. Our cross-platform testing of 500 projects reveals that 11.2% exhibit OS-dependent test failures. Through systematic analysis of 240 GitHub issues, we confirm 102 genuine portability problems spanning 95 additional projects. We develop a comprehensive taxonomy identifying 7 primary failure categories - with file/directory operations, process management, and library dependencies being most prevalent - along with 24 distinct sub-categories, 15 diagnostic signatures, and 4 systematic repair patterns. Our evaluation reveals that existing static analysis tools provide minimal support for portability detection, while large language models achieve 40-79% accuracy in identifying issues and 50-77% success in generating fixes when provided with structured guidance. Through 33 contributed pull requests, we demonstrate practical applicability and developer acceptance (17 merged, zero rejected) of our findings. This work establishes the first comprehensive baseline for understanding and addressing cross-OS portability issues in Python, providing actionable insights for developers, tool designers, and the broader research community.

We study the holographic description of the RN-like black hole in bumblebee gravity, which is a vector-tensor theory with spontaneous Lorentz violation. By combining the first laws of the outer and inner horizons, the thermodynamics method yields the right- and left-moving entropies, temperatures, and chemical potentials of the dual two-dimensional conformal field theory (CFT), determines the central charges from the mass-independent entropy product, and reproduces the outer and inner horizon entropies through the Cardy formula. An independent dynamical test based on the hidden conformal symmetry of a charged scalar probe confirms these results. A central outcome is the equality of the right- and left-moving central charges, $c_R=c_L$, which we trace to the spontaneous character of Lorentz breaking. Because the action of bumblebee gravity preserves diffeomorphism and local Lorentz invariance, and the symmetry is broken only by a vector vacuum expectation value, no chirally asymmetric structure arises in the dual theory, in sharp contrast to topologically massive gravity in $(2+1)$ dimensions. All CFT data depend smoothly on the Lorentz-violating parameter and reduce to the known RN results when Lorentz violation is switched off.

Let $G$ be a connected graph. The $k$-token graph of $G$ is the graph $F_k(G)$ whose vertex set consists of all subsets of $k$ vertices of $G$, where two of them are adjacent whenever their symmetric difference is an edge of $G$. Every automorphism of $G$ induces one of $F_k(G)$, as does complementation when $k=|G|/2$; automorphisms of this form are called induced. Fabila-Monroy et al.\ (Graphs and Combinatorics 42, 2026) show that token graphs can have many non-induced automorphisms, arising from twin cuts. These are cut sets $\{x,y\}$ whose two vertices have the same neighbours (other than themselves) in $G$. These non-induced automorphisms send configurations with a prescribed number of tokens on each component of $G\setminus \{x,y\}$, totalling $k-1$, and exactly one token on one vertex of $\{x,y\}$, to the configuration obtained by moving (flipping) the token at $\{x,y\}$ to the other vertex of $\{x,y\}$. These automorphisms send an induced $4$-cycle generated by moving two tokens on two disjoint edges to one generated by moving two tokens on a $4$-cycle; thus introducing what we call a twist. We prove a partial converse: if an isomorphism $\varphi\colon F_k(G)\to F_{k'}(G')$ sends some $4$-cycle generated by moving two tokens on two disjoint edges to one generated by moving two tokens on a $4$-cycle, then $G$ and $G'$ have twin cuts $\{x,y\}$ and $\{x',y'\}$, respectively. We also show that any twist can be undone by composing $\varphi$ with twin-cut flips.

Flaky tests pass or fail non-deterministically on unchanged code, eroding trust in test suites and inflating the cost of every failure. Shaker detects them by injecting resource contention (CPU, memory, and I/O stress) to amplify non-determinism caused by concurrent execution, and was reported to detect 95% of the flaky tests in a Java and Android benchmark against 37.5% for plain re-execution (ReRun). We present the first empirical evaluation of Shaker for Python. Drawing non-order-dependent flaky tests from the ground-truth dataset of Gruber et al., we compare Shaker against a budget-matched ReRun baseline in a paired design, giving both techniques the same number of test executions: Each of 137 tests is run 100 times under each. As configured for Java and Android, Shaker provides no statistically significant detection advantage over plain re-execution (37.2% vs. 35.8%; McNemar exact p = 0.84). Two findings explain why. First, fewer than half of the ground-truth flaky tests reproduce as flaky at all on independent hardware under either technique, and most of the tests that fail to reproduce never diverge once across 100 runs. Second, the tests that do reproduce are dominated by flakiness from network interactions and randomness rather than the concurrency Shaker targets. Beyond the tool, this exposes a broader hazard for the field: reusing a flaky-test ground truth across execution environments silently converts genuine flaky tests into apparent true negatives, deflating any tool's measured recall.

Current robot-assisted minimally-invasive surgery (RMIS) platforms provide a fixed console for the surgeon to view stereo endoscopic images and teleoperate instruments inside the patient. Several researchers have proposed the use of a head-mounted display (HMD) as a portable console, with video pass-through rendering of the endoscope images which, like the fixed console, restricts the operator to a single endoscopic viewpoint and limits depth perception. We present a digital twin-driven virtual reality (VR) teleoperation platform, where the digital twin is created from markerless perception of the surgical environment and streamed for display on the HMD. This overcomes the limitations of video pass-through by providing multi-view rendering and natural motion-parallax cues, enabling decoupling of the user's hand posture from strict instrument alignment. The system utilizes VR hand controllers to increase the teleoperation workspace and to improve the robustness and stability of instrument control compared to the hand tracking approach adopted by most prior systems. A 15-participant user study on the da Vinci Research Kit (dVRK) shows that our VR platform significantly outperforms a state-of-the-art HoloLens 2 mixed reality baseline, reducing path length by 86% and jerk by 95%, while achieving depth perception confidence comparable to or exceeding the traditional console across all conditions.

Binary interaction fundamentally alters the evolution of massive stars, yet the physical consequences of mass transfer remain poorly constrained. Here we present high-resolution spectroscopic observations of Plaskett's Star (HD 47129) that reveal small radial-velocity variations of its magnetic secondary, coherent with the system's 14.4 d orbital period. These measurements support an evolved, post-mass-transfer configuration in which the narrow-lined primary is a partially Roche-lobe-filling stripped star with a mass of $5.9^{+1.8}_{-1.4},M_\odot$, while the secondary is a $40.8^{+9.3}_{-6.7},M_\odot$, rapidly rotating magnetic accretor. The system provides a rare opportunity to observe the immediate aftermath of mass transfer in a massive binary and offers compelling new evidence linking binary interaction to the emergence of strong stellar magnetism. Because magnetic fields regulate angular-momentum loss and transport, their formation through mass transfer could alter the subsequent evolution of massive binaries, with consequences for stellar populations, supernova progenitors and yields, and the formation of gravitational-wave sources.

The Automated Ball-Strike challenge system that Major League Baseball adopted in 2026 offers a distinctive setting for studying human AI interaction in which umpires make every ball and strike call, while players can selectively ask an automated system to publicly overturn those decisions. We analyze 4,114,256 called pitches from 2015 through 2026 and 8,447 challenges from the 2026 season to examine how algorithmic review reshapes umpire judgment and player behavior. We study where umpires placed the effective strike zone boundary, how consistently they applied that boundary, how they responded to overturned calls, and which calls players chose to challenge. In 2026, the effective called boundary shifted toward the automated strike zone beyond the trajectory observed in prior seasons, while the consistency of that boundary largely continued its existing trend. Following an overturned call, umpires temporarily adjusted subsequent decisions near the corrected boundary, although these effects did not consistently persist into the next game. Count dependent variation in calling remained, while differences associated with player status narrowed. Players, meanwhile, left many overturnable calls unchallenged and appeared to base challenge decisions more strongly on immediately observable evidence than on the precise geometry of the automated zone. Together, these findings show that selective AI review does more than correct individual errors. It reshapes human judgment, adaptation, and strategic behavior around an algorithmic authority.

We prove that every Deligne–Mumford superstack over $\mathbb{Z}[1/2]$ with finite inertia admits a coarse moduli superspace. This generalizes the Keel–Mori theorem to the super setting. As part of our argument, we introduce a notion of integral morphisms in algebraic supergeometry and use it to establish an effective descent result for étale morphisms of algebraic superspaces. We also include an example of a non-Deligne–Mumford superstack with finite inertia that does not admit a coarse moduli superspace.

Transition states are defined by reaction pathways, yet most machine-learning methods predict them as isolated geometries. We introduce MARC-TS, a two-stage framework that learns a continuous, endpoint-conditioned path, queries it at any resolution and uses local path context to refine a transition-state candidate. We construct T1x-IRC-8K, a dataset of 8,209 reactions and 1,088,725 path-resolved geometries. On held-out reactions, the path model reduced complete-path error by 48.4% relative to endpoint interpolation, and the localizer achieved a mean aligned structural error of 0.127 Å. Quantum-chemical optimization and vibrational analysis yielded 405 frequency-confirmed first-order saddle-point candidates from 410 predictions. In a 100-reaction nudged elastic band comparison, learned-path initialization reached a joint geometry-and-force target for 66% of reactions, compared with 12% for geometric interpolation after 100 optimizer steps. By treating the path as a reusable representation rather than an auxiliary output, MARC-TS connects transition-state prediction, mechanistic interpretation and quantum-chemical refinement.

Second-order black-hole perturbation theory will be an important component of precision gravitational-wave modeling and tests of strong-field gravity with next-generation detectors. Beyond quasinormal modes (QNMs), whose quadratic interactions have been the focus of recent work in black-hole spectroscopy, a generic retarded solution consists of a continuum of real-frequency scattering states whose nonlinear interactions in Kerr remain comparatively less explored. We use numerical relativity to study the quadratic response of Kerr black holes to nearly monochromatic $(\ell,m)=(2,\pm2)$ incident gravitational waves, measuring the complex self-coupling to the outgoing $(\ell,m)=(4,4)$ daughter at twice the parent frequency for black-hole spins up to $a=0.95$. At low frequencies, the coupling is strongly suppressed and exhibits opposite spin dependence for prograde and retrograde scattering. At higher prograde frequencies, we identify a daughter-mode resonance whose frequency and width extracted from the nonlinear response track the fundamental $(\ell,m,n)=(4,4,0)$ QNM. Near the fundamental $(2,2,0)$ QNM frequency, the in-mode coupling grows with spin, in contrast to the decreasing quadratic QNM coupling, demonstrating that the nonlinear response depends on the full parent scattering state rather than on its frequency alone. A complementary semi-analytic second-order Teukolsky calculation reproduces the nonlinear response from our numerical relativity simulations. Our numerical-relativity results and their semi-analytic extension provide a basis for generic homogeneous radiative perturbations of Kerr at second order, with applications to dynamical tides, near-extremal dynamics, and second-order gravitational self-force theory.

Personalized AI memories, defined as persistent decision-relevant states distilled by platforms from long-run human-AI interaction, are increasingly portable, provenance-certifiable, and tradable. Provenance answers the question of origin, but it does not settle who holds which rights over a memory, nor how wider deployment affects its future supply. We develop a unified model in which memory use is non-rival, while refresh, the maintenance of continued validity, is relationally co-produced by the represented individual and the platform. Even assuming perfect provenance and complete information, the model still yields a set of inherent limitations. First, there is an extraction ceiling: decision value is bounded by lived experience. Second, a rights-separation theorem shows that no budget-balanced ownership assignment can make both co-producers residual claimants, and that control rights and cash-flow rights operate on distinct margins. Third, a depletion result indicates that reach-expanding policies such as portability, interoperability, and competition raise current access but shrink the long-run memory stock. Fourth, we offer a formal comparison between memory ownership as it ought to be (subject-held) and as it is (platform-held); the welfare ranking reduces to two measurable quantities, both dominated by interior renewable-claim bundles. Finally, endogenous-quality and market-formation results show that uncompensated markets pollute their own raw material, that transparency and compensation are complements, and that the de facto platform-ownership regime is self-perpetuating. On the design side, the analysis points to separated provenance, performance, and permission credentials; compensation indexed to quality-deployment; and portability and salience, rather than nominal ownership, as the binding policy margins.

WebGPU is a low-level graphics and compute API that exposes modern GPU functionality to web applications. While the official WebGPU Conformance Test Suite (CTS) focuses on well-formed usage under the WebGPU specification, it is not designed to stress implementations with semantic edge cases or adversarial inputs. General-purpose fuzzers, in contrast, struggle with WebGPU because of its complex graphics stack and multi-process architecture. We introduce LANTERN, a specification-guided dynamic conformance testing framework that mutates CTS tests using constraints extracted from the WebGPU specification. LANTERN extracts explicit syntactic API rules from WebIDL definitions and recovers semantic constraints, such as command ordering and object lifetimes, from natural-language specification text. Selected rules guide AST-located textual transformations that generate both valid and intentionally invalid CTS variants. We execute the resulting tests at scale on AddressSanitizer-instrumented Chromium to discover bugs. Our evaluation discovers three reproducible bugs, including a heap corruption, in Chromium versions current at the time of study. These results demonstrate that syntactically seeded, semantics-aware mutation of conformance tests provides a way to uncover browser bugs during WebGPU testing.