Papers

Generative AI (GenAI) tools entered higher education classrooms faster than the field was able to study their effects on learning. One concern is that GenAI may displace the critical thinking and AI literacy that students will need after graduation. This paper reports a Design-Based Research pilot of a GenAI-assisted critical thinking framework, in which ChatGPT was used as a thinking partner in an undergraduate research methods and statistics course during Spring 2025 (N = 14). The mixed-methods design combined pre- and post-intervention measures of statistical learning (AASCDM), AI literacy (MAILS), and critical thinking (WGCTA) with instructor field notes, student artifacts, and student-AI interaction logs. Pre-post tests showed gains on every AASCDM dimension and on eight of nine MAILS dimensions, while WGCTA percentiles did not change. Qualitative analysis identified four themes: the ways students positioned the LLM (as answer generator, validator, or co-thinker); the depth of student engagement (procedural vs. conceptual); occasional humanizing of the tool; and the role of curriculum design in shaping each of the prior three. Read together, the findings indicate that one semester of GenAI-assisted instruction can move domain learning and self-reported AI literacy but does not move standardized critical thinking, and that the modal student-LLM relationship is one of validation instead of dialogue. We end with design principles for the next iteration of the framework and implications for research on adaptive and personalized learning.

Autonomous vehicles interacting with passengers through natural language must reason beyond immediate commands. Passenger intent may span multiple stages of behavior, depend on future events, refer to surrounding agents or landmarks, and remain relevant as driving conditions evolve. Existing language-enabled driving datasets largely focus on short, localized interactions, leaving these longer-horizon forms of passenger intent comparatively underexplored. We introduce doPlan, to our knowledge the first publicly available, human-annotated real-world dataset designed to study passenger language as persistent task context. Built on nuPlan, doPlan contains 5,154 human-written passenger instructions spanning 169.1 hours of cumulative instruction-aligned context over 50.9 hours of unique driving, with annotation windows ranging from 30.0 to 508.8 s. The annotations capture immediate, deferred, event-conditioned, persistent, and multi-stage passenger intent. The dataset, annotation interface, and supporting resources are publicly available at https://github.com/Mi3-Lab/doPlan. We evaluate four language-conditioned driving models and find that sensitivity to passenger language does not reliably translate into behavior consistent with the requested direction. More broadly, among 2,161 examples with a matched future maneuver, the first associated maneuver occurs a median of 24.6 s after the evaluation point, and only 9.8% occur within the models' common 5 s prediction horizon. These findings highlight the need to connect persistent passenger intent with successive planning decisions. doPlan provides a setting for studying how unresolved goals can be retained, grounded in evolving scenes, and tracked across multiple stages, including how a planner determines when a future goal becomes relevant to the current plan.

For operation at the High-Luminosity LHC, the CMS collaboration is developing a significantly enhanced silicon detector system with improved radiation hardness, higher granularity, and extended coverage up to a pseudorapidity of $|\eta|\approx$ 4. The Outer Tracker will be equipped with two types of modules (2S and PS) that will provide tracking information to the L1 trigger of the HL-LHC CMS detector, by rejecting signals from particles below a certain transverse momentum threshold. Hence, these modules must be assembled with high precision and have to pass strict requirements. In this note, an overview of 2S module production is presented, as well as representative examples of problems and experience acquired during module assembly and performance tests.

On-policy distillation (OPD) trains a student on its own generated responses using dense, token-level supervision from a stronger teacher. Vanilla OPD treats all teacher signals equally, assuming that the teacher's supervision is equally important for every token. However, teacher signals at different tokens may have very different effects on the student's performance: some correct important reasoning errors, while others have little effect on the final answer. Motivated by this observation, we introduce Dr. OPD (OPD Done Right), which defines the optimal weighted OPD to maximize the student's performance. We formulate Dr. OPD as a bilevel optimization problem in which the student learns from weighted teacher supervision, while the weights are selected to maximize the expected reward of the resulting student. To solve Dr. OPD, we develop an efficient iterative solver that updates the token weights and student policy alternatively. At each round, it updates weights in closed form and then takes one gradient step on the resulting weighted OPD objective. Under regularity conditions, we show that this weighted update achieves a higher expected reward than a vanilla OPD update. Empirically, across strong-to-weak and same-size distillation on math and code, Dr. OPD consistently outperforms all evaluated baselines. In particular, in the strong-to-weak distillation setting, Dr. OPD improves average math performance by $9.7$ points over vanilla OPD, and enables the smaller student to surpass its larger teacher.

An agent skill is a reusable, actionable natural-language artifact that guides an agent to perform a task effectively under a given harness. Recent studies have explored the optimization of agent skills, contributing to a growing collection of publicly available skills spanning diverse tasks, domains, and harnesses. Despite millions of publicly shared skills, existing skill optimization methods largely overlook this accumulated knowledge, instead relying solely on expensive agent rollouts to iteratively refine skills for a target task. To address this, we propose Retrieval-Augmented Skill Optimization (RASO), a framework that leverages an external skill corpus as prior knowledge throughout skill optimization. RASO retrieves relevant knowledge from existing skills and adapts it to the target task and harness via Cross-Harness Adaptation, accounting for mismatches in both domain and harness. RASO comprises two complementary stages: Retrieval-Augmented Skill Initialization (RASI) constructs a knowledge-grounded initial skill without requiring agent rollouts, while Retrieval-Augmented Skill Update (RASU) iteratively refines the skill by retrieving external knowledge guided by execution feedback. Across four agent benchmarks and two models, extensive experiments show that RASO consistently outperforms baselines without retrieval-augmented skill initialization and updating.

Physics-Informed Neural Networks (PINNs) embed governing equations into deep learning, but enforce them only through loss residuals, leaving highly oscillatory wave behavior to be discovered by optimization. As a result, methods that achieve relative $L_2$ errors below $10^{-3}$ on standard manufactured Helmholtz benchmarks can fail on practical radiation problems involving singular excitations, absorbing boundaries, and wave fields spanning tens of wavelengths. Architectural physics embedding addresses this limitation by factorizing the field into analytically derived oscillatory kernels and learnable envelopes. However, the kernel dictionary must be manually constructed and scales with the number of elementary units, growing exponentially with the depth of hierarchically structured systems such as antenna arrays and metasurfaces. We propose PE-EK-PINN (Physics Embedded with Evolving Kernels), which treats physics kernels as reusable learned representations rather than fixed analytical inputs. A converged subsystem field is frozen and promoted to an evolved kernel, whose transformed copies are reused to represent higher-level configurations without deriving new governing equations. The resulting hierarchy makes the peak number of active kernels independent of system size and reduces cumulative training cost from $O(N)$ to $O(\log N)$. Experiments on dipole arrays, composite line-source geometries, and cross arrays demonstrate the dramatic training cost reduction, while achieving a reduced or comparable relative $L_2$ error. One notable example is PE-EK-PINN solves a $256$-dipole array more than 30 times faster than direct PE-PINN.

The LHCb Collaboration has observed a narrow state in the $\bar{B}_s^0\pi^0$ spectrum, about 74 MeV below the $B^-K^+$ threshold, which is the natural candidate for the bottom partner of the $D_{s0}^*(2317)$. Such a state had been predicted as a $\bar{B}K$ bound state by unitarized heavy-meson chiral perturbation theory with lattice-QCD input. We analyze the LHCb spectrum with these coupled-channel amplitudes, including isospin breaking through the physical meson masses, $\pi^0$-$\eta$ mixing, and $m_u\neq m_d$ in the chiral amplitudes, thereby allowing the $\bar{B}_{s0}^*$ to decay through its only open strong channel, $\bar{B}_s^0\pi^0$. With the parameters fixed by lattice QCD and no experimental input from the LHCb measurement, the predicted state lies within 1.1-1.7 standard deviations of the observed peak. Adjusting a single parameter, the spectrum is described as well as with the zero-width limit of the LHCb analysis. The new state thus emerges naturally from these amplitudes, with a large $\bar{B}K$ component and an isospin-violating width of tens of keV, and its $J^P=1^+$ heavy-quark spin partner is predicted with a precision comparable to that of the measured mass.

This report audits evaluation of a long-term-memory retrieval chain on the 500 LongMemEval-S development questions. Its strongest historical reader lane scores 479 and 475 under an adapted GPT-4o rubric; re-judging the same pass-1 answers changes three labels and yields 478. Fixed-answer knowledge-update re-scoring gives 70/72 under the upstream template and 69/72 under the modified template. Reader lanes span 93 to 479 on fixed packets; paired tests between the two strongest historical lanes establish neither superiority nor equivalence. A different-family reader, configured without client tools or operator files, scores 474, 1.0 percentage point below the headline pass (paired 95% interval [-3.0,+1.0]). Live reader request bodies were not retained. With the same requested reader label, route and judge snapshot, the full package scores 474 versus 454 for baseline sessions, a difference of +4.0 percentage points [95% interval +2.2,+6.0]. Eighteen of the 23 gains, and no losses, occur where baseline packets lacked listed evidence; this post-hoc split does not identify a component effect. In recovered LoCoMo data, token-F1 gains do not survive answer-line extraction. A negative control rejects a verifier that repairs three wrong drafts but breaks eleven correct ones. All questions were used to develop the components; no untouched holdout was evaluated. These findings do not establish a new leaderboard leader or transferable memory advantage. The A/D comparison has one pass per arm, including six reused identical-prompt outcomes, with no pinned reader snapshot; B/C and repeats remain unrun. Original headline requests cannot be reconstructed and stages 1–4 remain closed. Released artifacts support packet inspection and saved-verdict recounting and re-scoring; they do not reconstruct the method.

The rapid growth of artificial intelligence (AI) data centers has introduced new challenges to power system operation. As their power demand becomes larger and more variable, quantitatively characterizing their demand flexibility is increasingly important for effective power system coordination. However, heterogeneous workload characteristics and resource requirements make this flexibility difficult to characterize directly. This paper proposes a framework for assessing the grid-compatible demand flexibility of AI data centers via batch workload temporal shifting. An averaging-based resource usage processing method is developed to map fine-resolution CPU, GPU and memory usage into unified time intervals compatible with power system operation. A workload temporal scheduling model is then formulated to shift batch workloads while preserving execution continuity, delay constraints, and server resource capacities, and is coupled with a utilization-dependent server power model to translate workload scheduling decisions into server power demand. Two complementary flexibility metrics are evaluated: short-term peak demand shaving and the maximum duration of sustained power reduction. Numerical results based on real GPU cluster traces demonstrate that workload temporal shifting can provide quantifiable and grid-compatible demand flexibility for AI data centers with limited disruption to computing workloads.

We present RGOR (Reference-Grounded Oral Refinement), an audio-driven lip-sync framework that renders the mouth of the specific person being dubbed rather than a generic one. Existing lip-sync systems follow the audio closely and keep the face recognizable, yet the mouth they render is an average mouth: the shape and texture of the lips, the arrangement of the teeth, and how much of them shows as the mouth opens are not that person's. The problem persists because nothing in current training or evaluation asks for the person's own mouth: perceptual losses accept any plausible mouth, face identity is carried mostly by the skin around it, and the released inference code of inpainting systems uses the unmasked target frame as the reference, which hides the gap. To address this, RGOR conditions every generated frame on frames from separate enrollment recordings of the same person and on HD patches of the mouth that bypass the VAE, and trains the generator against a paired judge that compares each rendered mouth with the person's reference and learns to reject a realistic mouth of someone else. We further build an evaluation protocol and use it to compare open-source and commercial lip-sync systems on held-out identities. Experiments show that RGOR achieves the best or second-best result on most metrics, and preserves the person's own lip and dental detail while keeping synchronization and the rest of the face intact.

Group relative policy optimization (GRPO) learns only from prompts whose sampled responses disagree: a group that is entirely correct or entirely incorrect has zero reward variance, contributes no gradient, and still consumes its rollouts. Prompt-selection methods reduce this waste by steering sampling toward intermediate pass rates, but they choose the target, its width, and the uncertainty model heuristically, in raw pass-rate or logit coordinates. We show that GRPO comes with a natural coordinate for pass rates: the arc length $\psi=\arcsin\sqrt{p}$ on the Bernoulli Fisher–Rao manifold. In arc length, the expected GRPO update is uniform up to two boundary ramps; the probability of a zero-variance group is bounded by two Gaussian boundary layers of width $1/\sqrt{2G}$; pass-rate evidence has constant noise; and the gradients of the pass@$k$ and pass$^k$ objectives are Gaussians whose center and width follow from $k$ in closed form. A prompt curriculum for GRPO is therefore a Gaussian in arc length, and choosing its center amounts to choosing the objective. We turn this observation into ARCUS, a drop-in sampler that tracks every prompt with a Kalman filter in arc length, scores prompts by an objective-matched Gaussian kernel times the predicted probability of an informative group, keeps only informative groups for the unchanged GRPO update, and paces the target toward the hardest objective whose predicted yield stays within a small slack of the best. Across six mathematical reasoning benchmarks and three backbones, ARCUS improves the average accuracy of GRPO by 2.8–2.9 points and that of dynamic sampling by 1.1–1.2 points, while generating 48–57% fewer rollouts than dynamic sampling.

For Bernoulli bond percolation on every regular tree of degree at least three, we prove that infinite clusters appear strictly before absolutely continuous spectrum. Throughout an explicit supercritical interval, the adjacency operator of every open cluster almost surely has no absolutely continuous component. Together with Bordenave's theorem, this gives $p_c<p_{ac}<1$.

Constrained Reinforcement Learning has recently gained increasing attention in the field of Safe Autonomous Driving, where the general mechanism is to maximize the expected reward while keeping the overall action risk bounded. In this way, the safety issues arising in AD can be mitigated through constrained actions. However, existing Constrained RL methods still lack dynamics on the imposed constraints. For instance, the action cost adopted by the existing Primal-Dual/soft-constrained methods is often defined as static state-to-cost mapping, and the safe-action projection in hard-constrained methods relies on the static projection with the fixed feasible region boundary estimated from offline demonstrations. The above drawback tightly couples the imposed constraints to the training scenarios, leaving the AD policy hard to handle different interaction scenarios, due to the improper state-level action-cost and the static projection boundary. Consequently, in this paper, we propose Brain-SAD, a brain-inspired safe autonomous driving control framework with dynamic fear-oriented constraints. By perceiving the current vehicle-interaction scene, Brain-SAD generates dynamic fear signal as fear reaction to online decide long-term policy for regular interaction or short-term policy for urgent-collision defense. In such two policy, the above fear-reaction will be constructed as the dynamic fear constraints, respectively reflecting the overall fear cost directly coupled with action-impact, and the dynamic fear boundary of the feasible region derived from different risky neighbors, both of which will in turn serve for the online policy optimization. Experimental results show that Brain-SAD outperforms existing methods, achieving higher success rate in shorter task-completion and collision-recovery time, and exhibits stronger reliability across continuous intersections of fluctuating complexity.

In these lecture notes, we introduce the variational Monte Carlo method for the Hubbard model. First, we briefly introduce the single-orbital Hubbard model and its multi-orbital extension, focusing only on electronic models. Then, we introduce the general scheme under which variational Monte Carlo works and we present the variational wave functions that can properly describe superconductive and magnetic states in single-orbital and multi-orbital Hubbard models. We also detail how to perform Monte Carlo calculations for the given wave functions. We also describe how to optimize a variational wave function with the stochastic reconfiguration method. Finally, we present a few results for correlation functions that can be computed in variational Monte Carlo, to determine, for instance, the presence of charge or spin order and superconductivity.

We propose a single-level reformulation (SLR) for the pessimistic bilevel optimization problem that does not rely on complementarity conditions or optimal value functions. For this reason, we refer to it as a true single-level reformulation (tSLR). A remarkable consequence is that this reformulation can satisfy the classical linear independence constraint qualification, despite the fact that even the weaker Mangasarian-Fromovitz constraint qualification is known to systematically fail for standard single-level reformulations of both optimistic and pessimistic bilevel programs. We leverage on these constraint qualifications to construct new necessary optimality conditions for pessimistic bilevel optimization. The reformulation also has a striking limitation: under the assumptions of our analysis, the classical second-order sufficient condition fails at every Karush-Kuhn-Tucker point of the problem. Nevertheless, preliminary numerical experiments demonstrate that algorithms based on the proposed tSLR can outperform existing approaches for pessimistic bilevel optimization. Overall, the proposed framework suggests that pessimistic bilevel programs may be considerably more tractable than previously believed and need not be inherently more difficult to solve than their optimistic counterparts.

We construct an automorphism of a smooth projective K3 surface with positive topological entropy and a non-empty Fatou set. The Fatou set contains invariant domains biholomorphic to products of two annuli. On each domain, the automorphism is holomorphically conjugate to a fixed-point-free irrational rotation. These domains are two-dimensional analogues of Herman rings. The construction answers a question posed by McMullen in 2002 and several questions posed by Cantat in his 2018 ICM address.

JavaScript and TypeScript are widely used in modern web development, making their security critical; however, automated vulnerability detection is often constrained by the availability of high-quality training data. Here we present JsVul, a dataset curated from seven major sources. Unlike generic multi-language datasets that may retain noise – such as minified code and cosmetic edits – JsVul utilizes a language-specific pipeline. We collected pre-fix and post-fix versions of files around security fixes and, by filtering irrelevant artifacts and applying automated syntax normalization, isolated security-related changes. We ensured data integrity through multi-stage deduplication and heuristic-based labeling. Provided in a time-ordered JSONL format, JsVul supports robust model training in the JavaScript and TypeScript ecosystem and demonstrates the importance of language-aware preprocessing in building vulnerability datasets.

Modern machine learning systems are trained on mixtures of data from different domains, and choosing the right mixture can substantially improve downstream performance. Despite an extensive literature on data mixing and reweighting, existing work is largely empirical and it remains unclear when auxiliary data genuinely improves scaling laws rather than merely providing more samples. To gain insight into this question, we study a high-dimensional mixed-data regression model with a shared regression function, heterogeneous covariances and noise levels, and dataset sizes that may grow at different rates. We establish the minimax risk under an ellipsoidal parameter constraint for the general covariance structure and derive deterministic equivalents for the test error of ridge regression under commutative covariances. We then specialize to a target domain and an auxiliary domain with aligned power-law covariance spectra, where the theory yields explicit scaling laws in terms of spectral decay, target regularity, and the relative growth of the two datasets. These laws identify regimes in which combining data mixtures provably yields a faster scaling rate than using either dataset alone. In particular, improving the scaling law requires a specific interplay between spectra and relative sample sizes of the domains. Our numerical experiments on language models exhibit the same qualitative phenomenon: appropriate data mixtures yield a faster decrease in target-domain test loss than training on either domain alone.

Modern computer vision models achieve high accuracy when trained on large-scale annotated datasets. In critical domains such as construction safety monitoring, data collection is costly, hazardous, and ethically constrained. This paper presents a systematic study comparing two complementary data generation paradigms, (1) Unity Simulation-based rendering and (2) Controllable Diffusion-based generation (CIA), for object detection under real data-scarce conditions. A unified experimental framework enables controlled dataset mixing across real, simulated, and generative sources, while maintaining identical model and training settings. Quantitative evaluation using Precision, Recall, mAP, and custom $\Delta$-metrics, reveals that neither simulation nor generative augmentation alone achieves optimal transferability. Unity-only training yields an mAP@0.5 drop of $-50\%$ relative to real data, while CIA-only training shows a milder $-16.5\%$ degradation. Hybrid compositions significantly improve performance, with the 90% real + 10% Unity configuration achieving the best overall mAP@0.5 of $62.68\%$ ($+7.64\%$ over baseline), and the 90% real + 10% CIA configuration maximizing precision at $74.45\%$. Results demonstrate that limited synthetic inclusion enhances generalization, while excessive substitution induces domain drift.

Pulsed light is represented by density matrices whose correlation functions have different properties of stationarity and coherence. We show that these density matrices yield equivalent measurement results after accounting for finite detection times. This in turn reveals a physical meaning to the length of a quantization cavity and explains the origin of the physical spectrum for pulsed light. By generalizing the density matrices for pulsed Gaussian states of light, we can describe pulsed light of varying statistics in the same basis and relate the photocount rate to the pulse spectrum.

Computer use agents (CUAs) have demonstrated strong capabilities in completing digital tasks. However, existing CUAs either rely solely on graphical user interface (GUI) interactions, which are often inefficient and error prone, or augment GUI interactions with application specific APIs or tools, which require substantial engineering effort and are difficult to scale across applications. We argue that the next generation of CUAs should combine GUI interactions with the command line interface (CLI), leveraging the generality of the GUI and the efficiency of shell commands. A critical challenge, however, is that current models do not know when or how to use the CLI during task execution. To address this challenge, we develop a data construction pipeline that produces three types of trajectories: GUI only, CLI only, and interleaved GUI and CLI trajectories. This pipeline results in HybridCUA-8K, containing 5K hybrid trajectories and 3K verified RLVR tasks. Building on these data, we propose a training framework with two stages: supervised fine tuning on the constructed trajectories, followed by reinforcement learning with our CLI aware rewards that encourages agents to use the CLI selectively and reliably. Experiments show that HybridCUA-9B achieves 53.6% accuracy on OSWorld, improving over the base model by 14.8 percentage points, and improves performance on WindowsAgentArena by 4.0 percentage points. These results demonstrate the effectiveness and cross platform generalizability of the hybrid GUI and CLI paradigm for computer use agents.

We establish improved locally Markovian properties of quantum Gibbs states at every fixed positive temperature. For finite-range interactions of bounded interaction degree, the conditional mutual information (CMI) decays exponentially with separation, with a prefactor exponential in the boundary size of only one of the two regions and independent of the size of the other region and the total system. This improves the exponential volume dependence of the local Markov bounds established by Chen and Rouzé \cite{Chen2025}. For exponentially and stretched-exponentially decaying interactions, we obtain stretched-exponential CMI decay with a boundary-dependent prefactor. For power-law interactions, we prove inverse-logarithmic or algebraic decay with polynomial region-size dependence. Our proof is static, in contrast to the dynamical approach of Chen and Rouzé. We compare the Gibbs state with a product state obtained by removing boundary interactions, reducing the CMI bound to a relative-entropy loss that we control using locality estimates.

Running a language model locally offers advantages in privacy, latency, and cost, but local hardware fits only small models, which are less capable than frontier models. The usual remedy for a hard query, escalating it to a cloud model, gives up the privacy and cost advantages of running locally. A deployment that stays local faces two decisions for hard queries instead. First, it can spend more computation on a query, e.g., reasoning before answering, which raises accuracy at a cost in latency, so it must decide which queries are worth the extra computation (efficiency). Second, some queries are beyond the local model, and delivering a wrong answer is worse than deferring the query to a human in the loop, so it must decide which answers are safe to deliver (safety). We show that both decisions can be made from the model's own hidden states. The prefill state, computed before any token is generated, predicts whether the model will answer correctly, and the answer state, at the end of the generated answer, predicts whether that answer is correct. HARISSA fine-tunes the model so that both states predict correctness, then makes both decisions with one policy that cascades through the ways of answering from cheapest to most expensive, skipping a way the prefill state predicts will fail and deferring the query when the answer it stops with is predicted wrong. On a device running a single model, HARISSA is within one accuracy point of chain-of-thought at 2.7 times lower latency. On a server holding four sizes of one model, HARISSA is more accurate than the FrugalGPT and Self-REF cascades at the same latency, and at the same deferral rate the answer state leaves fewer wrong answers than the standard confidence signals in five of six task and setting pairs.

Large language models (LLMs) are increasingly proposed as decision assistants who must reason probabilistically from available evidence under explicit decision costs. We propose a decision-theoretic framework that decomposes LLMs' decision loss into two components: forming accurate beliefs from provided evidence and translating those beliefs into actions that optimize a provided utility function. Using a synthetic benchmark with known ground truth, we apply the decomposition to characterize probabilistic reasoning in frontier and open-sourced models. We further evaluate whether RL interventions targeting beliefs, decisions, or both improve these components across three domains, whether improvements transfer across components and elicitation formats, and whether decision performance can improve without improvement in belief formation. We find that targeting one component of probabilistic reasoning redistributes decision loss, improving the target without necessarily transferring to others, and that jointly targeting belief formation and decision-making improves both but hinges on matched formats between training and evaluation.

Time series anomaly detection (TSAD) plays a crucial role in healthcare, finance, industrial monitoring, and other sectors. Within and between these settings, anomalies span vastly different temporal scales, from sub-second point spikes to multi-hour drift patterns. However, most existing TSAD methods commit to a single temporal granularity, and multi-scale designs either analyze different scales in isolation or are constrained to a predefined coarse-to-fine hierarchy, both failing to sufficiently capture multi-scale interactions. To resolve this limitation, we propose Multi-Scale Autoencoder with Cross-Scale Attention for TSAD (MSCAD), a simple yet powerful semi-supervised TSAD framework founded on parallel autoencoder branches corresponding to different patch sizes. A stack of symmetric bidirectional cross-scale attention blocks enables every pair of scales to exchange information before reconstruction without allowing any single scale to be privileged. On the comprehensive TSB-AD benchmark (40 datasets, 530 series), MSCAD achieves large performance gains against 50 baselines across multiple metrics, with VUS-PR of 0.57(+9.6%) on the univariate split and 0.47(+9.3%) on the multivariate split compared to the state-of-the-art.

The semiconductor diode, which allows current to flow preferentially in one direction, is a fundamental building block of numerous modern electronic circuits. Its superconducting analogue—the superconducting diode effect—enables directional dissipationless current flow and may provide similar functionality in future superconducting quantum circuits. Realization of such a nonreciprocal supercurrent requires broken time-reversal symmetry. At zero applied field, this has been typically associated either with intrinsic unconventional superconductivity or extrinsic spin-split electronic states induced by magnetic proximity. Here we demonstrate a field-free superconducting diode effect in conventional superconducting Al proximitized by collinear antiferromagnetic CuMnAs, whose electronic structure breaks time-reversal symmetry without generating spin splitting. By tuning the proximity effect through an insulating AlAs interlayer and correlating the reversal of the diode polarity with the reversal of the remanent Néel state, we establish that the antiferromagnet controls the superconducting diode effect. Our results show that neither spin-split bands nor net magnetization is required for a magnetically controlled field-free superconducting diode effect, extending superconducting nonreciprocity to a broader class of collinear compensated magnets.

Lazard asked which $N$-series of a group arise from multiplicative filtrations of the integral group ring. Given an $N$-series $H_\bullet$ of a group $G$, let $A_\bullet$ be the filtration of the augmentation ideal spanned by products of elements $x-1$, $x\in H_a$, weighted by $a$. Every realizing filtration contains $A_\bullet$, so $H_\bullet$ is realizable exactly when the maps $H_n/H_{n+1}\to A_n/A_{n+1}$, $h\mapsto h-1$, are injective. We show that each kernel is the cokernel of an explicit map of normalized bar groups; for a finite group this makes realizability decidable by integer linear algebra. For a profinite group with a cofinal tower of finite quotients and an $N$-series induced from it, suppose $h\in H_n$ and $h-1\in A_{n+1}$ in every finite quotient. Then $h-1\in A_{n+1}$ in the discrete group ring exactly when the widths of its finite-level Losey expressions are bounded. This descent can fail. Using Tahara's class-three $2$-groups and the Hartl-Mikhailov-Passi description of the fourth dimension subgroup, we construct a countable product $P$ of finite $2$-groups. It contains an element of $\gamma_3(P)$ that lies in the fourth dimension subgroup of every finite subproduct of $P$ but not in $D_4(P)$.

We prove a geometric version of Poonen's "Mordell–Lang plus Bogomolov" theorem for semi-abelian varieties in characteristic zero. This is a generalization of the Mordell–Lang conjecture and the geometric Bogomolov conjecture.

Time delay estimation (TDE) is a fundamental component of microphone-array sound source localization. Generalized cross-correlation (GCC) is widely used because it is efficient and interpretable, but its handcrafted spectral matching and predefined frequency weighting are vulnerable to noise and reverberation. Existing neural GCC variants mainly improve robustness by enhancing input signals or modeling GCC responses, while the cross-channel spectral matching step itself remains handcrafted. We propose QK-GCC, a learnable GCC-like framework that replaces handcrafted weighted spectral matching in GCC with Query-Key matching between two microphone signals. The two microphone signals are encoded as magnitude-phase frequency tokens and mapped to Query and Key representations, respectively, enabling frequency reliability learning and local spectral evidence aggregation for delay estimation. Experiments in simulated reverberant rooms across diverse SNR and reverberation conditions show that QK-GCC improves TDE accuracy over GCC-PHAT and learning-based GCC variants, while remaining lightweight and generalizing to unseen source types. The code is available at https://github.com/zhangjinkai33-ui/QK-GCC.

Kac's walk is a classical local random walk whose action on a single real unit vector in dimension $d$ mixes in total variation in $\Theta(d\log d)$ sequential steps~\cite{PS17}. Lu, Qin, Song, Yao, and Zhao introduced a parallel version of Kac's walk that mixes a single quantum state in $O(\log d)$ rounds~\cite{LQSY+26}. After discretizing the randomness and replacing it by suitable pseudorandom primitives, this parallel walk gives rise to pseudorandom state scramblers, and was subsequently shown to yield pseudorandom unitaries~\cite{LQSY+25}. We study what happens when the parallel Kac's walk acts simultaneously on $k$ orthonormal quantum states. We prove that, for any $1\leq k < d$, after $O\!\left((k+\log d)\log(d/\varepsilon)\right)$ steps, the joint distribution of the $k$ output states is $\varepsilon$-close, in both Wasserstein and total variation distance, to that obtained by applying a common Haar-random unitary to the same inputs. This generalizes the dispersing property of the parallel Kac's walk from a single quantum state to multiple orthonormal quantum states. Equivalently, viewing an ordered collection of $k$ orthonormal states as a point on the complex Stiefel manifold $V_{d,k}=\{X\in\mathbb C^{d\times k}:X^\dagger X=I_k\}$, we show that the parallel Kac's walk mixes rapidly on $V_{d,k}$, with both Wasserstein and total variation mixing times bounded by $O\!\left((k+\log d)\log(d/\varepsilon)\right)$. This extends the Wasserstein mixing result of Pillai, Smith, and Vaikuntanathan for the standard Kac's walk on real Stiefel manifolds~\cite{PSV26}.