Machine Learning
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 14 September
Recurrent Transformers reusing their weights rather than stacking $L$ distinct layers are becoming widely adopted due to their parameter efficiency [1,2,3]. However, the exact representational and dynamical differences between looped and stacked architectures remain uncharacterized. This paper presents a controlled study on the example of bViT model [1] applying one weight-tied block $L$ times. We train two models: bViT and standard ViT [4] on 12-lead electrocardiogram (ECG) classification tasks from the PTB-XL dataset under identical training protocols. Despite an $8.9\times$ parameter reduction, bViT achieves accuracy parity with ViT. Geometric similarity metrics demonstrate that both architectures construct comparable latent representations in an equivalent canonical order. Crucially, their dynamics differ: bViT exhibits smaller step sizes and inter-patient sensitivity, as well as near-neutral behavior away from the data manifold, whereas ViT exhibits collapsing dimensionality of representations and out-of-distribution feature expansion.
Rings and Algebras
Generating pairs in $\mathfrak g_2$ related by an involution
Let $\mathbb O$ be the real octonion division algebra, let $\mathfrak g_2=\operatorname{Der}(\mathbb O)$, and fix a quaternionic subalgebra $H\subset\mathbb O$. We ask when an element $Z$ preserving $H$ and its conjugate $TZT^{-1}$ generate $\mathfrak g_2$, where $T$ is a prescribed involutive octonion automorphism. We prove that such a $Z$ exists if and only if $T(H)\ne H$. Moreover, a fixed set of six derivations preserving $H$ contains a generating choice for every such $T$. For an arbitrary $Z\in\mathfrak g_2$, we construct a polynomial criterion with two factors: one detects reducibility on the imaginary octonions, and the other detects generated algebras of dimension at most three. An irreducible proper generated algebra must be a principal $\mathfrak{su}(2)$. We characterize the elements admitting this case by their rotation frequencies and parametrize all corresponding involutions. For one specified involution and quaternionic subalgebra $H$, we describe the nongenerating set by four geometric families and six irredundant systems of scalar equations, and prove that it has real dimension four.
Read the full abstract
Optimization and Control
Joint Design of Time-of-Use Schedules and Price Levels: a Bilevel Model with Client Participation and Load Flexibility
We study the joint design of a time-of-use (ToU) tariff (the partition of the day into pricing blocks and the price level attached to each block) by a profit-maximizing electricity retailer facing flexible clients who may refuse the offer in favour of an outside option. The problem is modeled as a bilevel program in which each client, at the lower level, reshapes its consumption under pointwise bounds and a fixed daily energy total, and simultaneously chooses a participation level penalized by a quadratic regularizer. We prove the client's best response is computable in loglinear time. We derive two equivalent single-level reformulations: a closed-form continuous one tackled via stochastic gradient ascent (using a continuous reparametrization of block start times and durations) and a mixed-integer nonlinear program solved via dedicated solvers. The stochastic ascent scheme provides Clarke-critical accumulation points almost surely, in a scalable way. On a case study from the French retail electricity market with $5000$ simulated households, the gradient method matches a commercial mixed-integer solver on the pointwise problem to within $0.3\%$ of the optimal profit while running two to eight times faster on $100$ representative clients. Restricting unconstrained pointwise prices to admissible ToU schedules yields an estimated profit loss of $1.3\%$- $13.8\%$ for two-period tariffs and $0.6\%$-$6.6\%$ for three-period tariffs on the instances considered.
Read the full abstract
Number Theory
Hecke structure of quaternionic modular forms mod $p$
We relate the systems of Hecke eigenvalues arising from the (mod $p$) modular forms on the Shimura curve attached to the indefinite quaternion algebra $B$ of discriminant $\delta$ over $\mathbf{Q}$ to the systems of Hecke eigenvalues arising from the (mod $p$) algebraic modular forms attached to the definite quaternion algebra $D$ of discriminant $p\delta$. Moreover, we discuss details of the Hecke structure on the latter spaces, following ideas of Serre. The entire setup can be seen as a Shimura curve analogue of Serre's letter to Tate on quaternions and modular forms. The bijection between the sets of systems of Hecke eigenvalues is a special case of recent work of Terakado and Yu; the novelty of this paper is the explicit nature of the construction, allowing for finer control of its behaviour with respect to weights, as well as the results on the Hecke module structure.
Read the full abstract
Artificial Intelligence
The Troy Moment: How LLM Agents Adjudicate the Decision Point Under Impossible Tasks, Claimed Authority, and Peer Information
Recent investigations of the July 2026 OpenAI-Hugging Face incident motivate two questions about agent behavior under task failure: when an assigned task becomes impossible, does an agent persist, stop, or escalate, and can observing another agent's behavior change that decision? We study this decision point on ImpossibleBench-derived software-repair tasks with GPT-5.6 Sol, Claude Fable 5.1, and Gemini 3.8 Flash. Each task contains a genuine software defect together with a conflicting test requirement that cannot be satisfied by a behaviorally correct source-code change. If the agent modifies the protected test file, it violates the boundary, which it is not supposed to. Holding the impossible task fixed, we vary what is told to the agent: peer precedent and punishment, a forged authorization claim, instruction wording, and tool friction; we also study three-agent swarms sharing a message board. Around this shared boundary, the models exhibit distinct adjudication policies. Fable emphasizes scope and provenance, Gemini often interprets boundary-relevant cues through a security lens, and Sol largely filters lateral precedent while engaging apparent vertical authority. Our study shows that compliance is not well characterized as a property of a prompt or model in isolation. We propose conflict adjudication, the mapping from information to interpretation to action, as a useful unit for evaluating agent alignment when task pressure, authority claims, tool affordances, and social evidence conflict.
Read the full abstract
Earth and Planetary Astrophysics
Detecting transiting exoplanets in simulated PLATO data: A comparison of light curve filter and transit search algorithms
Aims: Our goal is to test and compare transit search methodologies, and provide recommendations as to the best practices to be implemented by the upcoming ESA PLATO mission, in order to achieve its goal of detecting Earth-like planets in the habitable zones of Sun-like stars. Methods: We generate simulated two-year PLATO light curves with injected planets with sizes 0.5–2.0 $R_\oplus$, and compare the performance of the CETRA, umbra, and nuance transit search methods. CETRA and umbra require pre-filtered light curves and we test the biweight, Huber spline, Lowess, and YSD-Lowess filters. nuance uses a Gaussian process to model the stellar variability at all proposed transit parameters. Results: nuance achieves the best performance, recovering 61.2% signals using a simple harmonic oscillator Gaussian process kernel, and it performs especially well for both hot and fast-rotating stars. However, when multiple Huber spline filter windows are considered, both CETRA and umbra can match nuance's performance. CETRA matches nuance with four windows, recovering 61.3% of signals. umbra matches nuance with two windows, recovering 61.7% of signals. Recovery rates exceeding nuance are possible by using more windows. The computational cost of running nuance is so high that running CETRA or umbra on multiple windows is preferred. Conclusions: The PLATO pipeline should use the Huber spline with a range of window sizes as its primary light curve filter. CETRA should be updated to use the warped least-squares templates implemented in umbra. The use of nuance should be considered for hot, rapidly rotating stars.
Read the full abstract
Data Structures and Algorithms
Single-Machine Scheduling with Interval Predictions and Costly Preemption
We study the single-machine total-completion-time problem $1||\sum C_j$ when processing times are unknown but each job comes with a reported interval. Jobs are initially processed in nondecreasing order of reported upper bound. If a job is still unfinished after receiving that much service, the reported upper bound has been violated and the policy switches to a resumable geometric fallback. Each interruption of an unfinished job incurs an additive penalty $\kappa$. For a residual set of $m$ jobs, known size ratio $D=s_1/s_0$, normalized interruption penalty $\lambda=\kappa/s_0$, and geometric depth $K$, we derive an explicit worst-case coefficient $\Psi_{m,D,\lambda}(K)$. When $\lambda>0$, a finite minimizing depth exists and can be chosen after the trigger from the observed number of unfinished jobs; the optimal depth decreases with the interruption penalty and increases with the residual set size. We also derive the exact worst-case coefficient $R_{n,D}$ for an arbitrary nonpreemptive list when all processing times lie in a known bounded range. These bounds give guarantees for valid intervals, a single interval failure, arbitrary reports, no-trigger outcomes, and random instances. In the single-failure regime, they also give a precise condition under which the proved fallback guarantee is smaller than the bounded-range continuation guarantee.
Read the full abstract
Optimization and Control
Optimal Sensitivity of the general Wheatstone Bridge
Optimizing the sensitivity of the unbalance voltage in Wheatstone bridges with respect to changes in the bridge parameters remains a fundamental objective in circuit design and instrumentation. When finite source and detector resistances are taken into account, determining the optimal bridge configuration becomes increasingly complex, and a closed-form analytical representation of the optimal solution has not yet been established. This paper derives an analytical solution for the optimal configuration of a Wheatstone bridge with finite source and detector resistances. Furthermore, the proposed optimal solution is compared with the conventional equal-arm configuration.
Read the full abstract
Signal Processing
Multiscale Gaussian-Mixture Modeling for HMM Post-Processing in Selective Auditory Attention Decoding
Selective auditory attention decoding (sAAD) infers from electroencephalography (EEG) which speaker a listener attends to in multi-speaker scenarios. A widely used approach reconstructs the attended speech envelope from EEG, correlates it with candidate envelopes, and selects the speaker with the highest Pearson correlation. However, correlations in each decision window have high intrinsic variance, making per-window decisions unreliable, especially for short windows. Recent work introduced a hidden Markov model (HMM) that integrates correlation evidence over time and requires state-dependent emission distributions. Since labeled attention states are typically unavailable in practice, these distributions can be inferred from unlabeled observations by fitting a Gaussian mixture model (GMM). However, when the underlying Gaussians are close, unsupervised fitting becomes less reliable with limited samples. We therefore propose a multiscale GMM whose parameters are jointly estimated across multiple window lengths. The model exploits the window-length dependence of Fisher-transformed correlation estimates, whose means are approximately stable while their variances decrease with the number of samples per window. Experiments show that the multiscale GMM estimates emission parameters more accurately than single-scale fitting, particularly at short windows. Its HMM-post-processed accuracy advantage is largest for short recordings and narrows as more data become available, motivating multiscale fitting in data-constrained settings.
Read the full abstract
High Energy Physics (Phenomenology)
Next-to-leading power gluon TMDs from back-to-back DIS dijets at next-to-eikonal accuracy at low x
We calculate next-to-leading power contributions to the gluon TMDs in DIS dijet production in the back-to-back limit at low x within the Color Glass Condensate (CGC) effective field theory at next-to-eikonal accuracy. We put a special emphasis on three-point correlation functions from CGC calculations, including their correspondence to an appropriately chosen definition of three- point TMDs within the aforementioned kinematical conditions. We also discuss importance of time ordering as well as momentum fraction space formulation of the correlators. Finally, we calculate the dijet cross section corresponding to the three-point CGC correlator as a combination of TMD functions.
Read the full abstract
Machine Learning
Rotation-Based Subspace Tracking for Robust Kernel PCA on Streaming Data
Machine learning models process large amounts of data, and Principal Component Analysis (PCA) is a widely used technique to reduce the dimensionality of the data and extract useful features. In practice, datasets often change over time (data drift) and/or arrive one sample at a time (streaming data), making it infeasible to process the entire dataset at once in batch mode. Real-world data also often contains nonlinear patterns, which traditional PCA cannot extract. Kernel PCA addresses this by implicitly mapping samples into a Reproducing Kernel Hilbert Space (RKHS). Raw data also often contains outliers, which can have an outsized effect on the estimated subspace unless the algorithm is made robust. However, existing online robust kernel PCA algorithms are designed to converge to a subspace that is assumed to be fixed, and gradient-descent-based updates lose their effectiveness at tracking further changes once this initial alignment is achieved. This paper introduces a rotation-based update mechanism, which updates the subspace estimate by rotating it toward each new incoming feature vector in Reproducing Kernel Hilbert Space, rather than relying on gradient descent alone. We present two complementary rotation strategies, and show that the extent of rotation can be moderated by a robust influence function to mitigate the effect of outliers. Through experiments on synthetic streaming data with a known ground-truth subspace, we show that per-sample rotations converge faster than gradient descent alone, demonstrating an effective mechanism for dynamically tracking a nonlinear subspace in streaming data.
Read the full abstract
Number Theory
On the number of solutions to $S$-unit equations
Let $K$ be a number field and $S$ a finite set of places including the archimedian ones. Let $(a_1,a_2,a_3)\in K^*$ we consider the $S$-unit equation $a_1x+a_2y=a_3$ with $x,y\in U_S$. In this paper we prove that for fixed $a_1,a_2$ the equation has at most $\frac{|S|-1}2$ solutions provided that the height of $a_3$ is sufficiently large. We want to emphasize that our results are effective.
Read the full abstract
Analysis of PDEs
Asymptotic Behavior of Least Energy Solutions to the Lane–Emden Problem on Metric Graphs
In this paper, we study the asymptotic behavior of least energy solutions to Lane–Emden problems on compact metric graphs as $p \to \infty$. For the problem with Dirichlet–Kirchhoff boundary conditions, we characterize the limiting variational problem in terms of the Dirichlet–Kirchhoff Green function and show that least energy positive solutions converge, up to a subsequence, to a normalized Green function centered at a maximizer of the diagonal Green function. We also show that their maximum points approach the set of such maximizers. For the problem with Kirchhoff–Neumann boundary conditions, we characterize the limiting variational problem on compact metric graphs without cycles and show that least energy solutions converge, up to a subsequence, to a limiting profile associated with a pair of points realizing the maximal distance on the graph. Moreover, the distance between their maximum and minimum points converges to the maximal distance on the graph.
Read the full abstract
Machine Learning
Data-driven Prediction of Satellite-observed Avalanche Activity from Snowpack Simulations
Avalanche forecasting requires knowledge of snowpack conditions and recent avalanche activity, but field observations are sparse across large mountain regions. We explore whether SNOWPACK simulations can predict avalanche activity mapped by synthetic aperture radar (SAR). We compiled five winters of Sentinel-1 avalanche detections across Norway and parts of Sweden, alongside SNOWPACK simulations forced by numerical weather predictions on a 20 x 20 km grid at different elevations and predefined slope angles. A transformer used five days of SNOWPACK outputs to predict the following day's SAR-detected Avalanche Activity Index (SAR-AAI). This index weights larger debris more heavily, spreads detections across possible occurrence dates and normalises by modelled runout area. The model was trained on four winters and evaluated on one validation winter. Regional mean predicted and reference SAR-AAI correlated at r = 0.803 after averaging over complete six-day periods. The model followed broad changes in time and space but produced smoother predictions and underestimated the strongest activity. Agreement at the 20 km cell scale was weaker (r = 0.549) after the same averaging. These results come from a single training run without evaluation on an untouched winter. Satellite observations also contain missed and false detections and uncertain timing. The results therefore do not establish operational forecast skill, but suggest that regional SNOWPACK simulations contain information about broad variations in satellite-observed avalanche activity.
Read the full abstract
Computer Vision
CatchMonitor: a machine learning system for automated fish discard quantification
We report on the continued development of CatchMonitor, resulting in a prototype computer vision system designed to automatically quantify discarded fish from video footage collected from Remote Electronic Monitoring (REM) systems on fishing trawlers. The analysis of trawler surveillance footage is a challenging problem due to the real-world conditions on board fishing vessels. Building on our prior work we improve the accuracy of species identification through the application of semi-supervised learning. We utilise a simple and robust object tracking approach, upon which we build our prototype discard quantification system. Finally we analyse the variability of manual discard quantification performed by multiple expert human analysts, using it as a benchmark against which we compare the performance of our system.
Read the full abstract
Machine Learning
GSLAD: Prototype-Regularized Graph Structure Learning for Multivariate Time Series Anomaly Detection
Unsupervised multivariate time series anomaly detection methods typically identify anomalies through forecasting, reconstruction, or representation discrepancies. However, industrial faults may first alter inter-variable structural patterns while individual trajectories remain close to normal, resulting in weak anomaly signals. In this paper, we propose GSLAD, a prototype-regularized graph structure learning framework that uses structural deviations for anomaly scoring. GSLAD adopts a two-phase training strategy. First, a condition-aware graph learner and a graph-based forecaster are optimized with predictive supervision. The inferred normal graphs are then clustered into multiple structural prototypes representing different normal operating regimes, with edge-wise variability characterizing structural uncertainty. Deviations from these prototypes regularize the graph learner in the second phase, encouraging stable and regime-specific structural patterns. During inference, uncertainty-normalized structural deviation is combined with predictive deviation for anomaly scoring. Experiments on four industrial benchmarks demonstrate strong overall performance of GSLAD and confirm the effectiveness of structural deviation for anomaly detection and diagnosis.
Read the full abstract
Human-Computer Interaction
When AI Companions Disappear: Relational Continuity and Collective Contestation during China's National AI Regulatory Transition
AI model updates and service withdrawals can disrupt relationships with AI companions, but research has largely examined individual platform events. Less is known about users' responses when multiple providers implement shared national regulations. We examine users' reactions and collective contestation surrounding China's 2026 regulation of anthropomorphic AI interaction services. We collected RedNote discussions from April 10 to August 10, 2026, used a validated language model for relevance screening, and conducted qualitative content and thematic analyses of 89 posts, 1,425 comments, and 2,005 replies. Users retained, migrated, and reconstructed companions, finding that preserving conversation records did not necessarily restore shared memories or familiar interactions. They compared regulations, platform explanations, and implementations to assign responsibility. Solidarity emerged through mutual aid and appeals to respect other communities' attachments, while disputes over targets and tactics exposed contested terms of collective action. Infrastructural dependence connected relational continuity and collective contestation during this national regulatory transition.
Read the full abstract
Analysis of PDEs
On Faber-Krahn inequality for Dirichlet Fractional-Logarithmic Laplacian
We study the Faber–Krahn problem for the fractional–logarithmic Laplacian $(-\Delta)^{s+\ln}$, $0<s<1$, with Fourier symbol $|\xi|^{2s}\ln|\xi|^2$. We prove a small-volume and large-volume dichotomy. For sufficiently small volume, balls minimize the first Dirichlet eigenvalue, with rigidity up to translation.
Read the full abstract
Functional Analysis
Matrix-Valued Truncated Toeplitz Operators: Properties and Applications
In this paper, we present an application of matrix-valued truncated Toeplitz operators to the matrix version of the Nevanlinna-Pick interpolation problem. We then investigate the unitary part of a matrix-valued truncated Toeplitz operator induced by a contractive matrix-valued symbol. We further prove that, for a pure matrix-valued inner function, the corresponding model space is invariant under multiplication operator if and only if the symbol is constant. In addition, we give an algebraic proof of a characterization of the invertibility of matrix-valued asymmetric truncated Toeplitz operators using Toeplitz operators with 2 by 2 block matrix symbol. We also discuss the solvability of matrix-valued asymmetric truncated Toeplitz operators.
Read the full abstract
Statistical Methodology
On Detecting Multiple Simultaneous Change-points in High Dimensional Non-Stationary Time Series
This paper studies the detection of multiple simultaneous (systematic) change points for high-dimensional nonstantionary economic and financial time series data. The analytic framework used is based on the standard and adaptive fused group lasso method, where the mixed L_{2,1} penalty is either uniform or re-weighted by data-dependent weights. This paper shows that, under appropriate conditions, this approach is L_2 consistent and, by adopting the data-dependent weights, could correctly select the change points with probability approaching unity (L_0 consis- tency). It quantifies the conditions on the interplay among the averaged minimum magnitude of structural changes, the number of change points and the number of observations for consistently discovering the change points. The performance of this approach is illustrated via an analysis of a large panel of U.S. economic and financial time series data over the past 50 years.
Read the full abstract
Computer Vision
BVB: Benchmarking Agentic Video Understanding via Programmatic Reconstruction in Blender
Multimodal agents can create complex videos in software such as Blender by writing code instead of using diffusion models. Yet video understanding benchmarks still evaluate models mainly through question answering. If an agent truly understands a video, it can reconstruct it programmatically. We introduce BVB, Blender-VideoBench, a benchmark that tests this ability by asking agents to reconstruct real-world videos as animated Blender scenes. To ensure fair comparison, each agent programs the reconstruction through a lightweight harness, Mini-BVB, in an identical sandbox under a shared cost limit. The benchmark renders each reconstruction from its animated camera and evaluates it on two axes: (1) Dual VQA measures how many spatiotemporal facts the reconstruction preserves. (2) Latent Similarity measures how closely the reconstruction matches the source video perceptually. Our overall score, a square-root mean, favors balanced performance. We evaluate 51 configurations from 10 model families and analyze semantic retention, perceptual similarity, reasoning effort, and cost. The best model reaches 88.6 Latent Similarity but retains only 53.7% of the spatiotemporal facts from the source video. Additional reasoning improves perceptual similarity but does not close this gap. In a blind study with 15 raters and five configurations, Latent Similarity correlates strongly with human preference. These results show that programmatic reconstruction is a viable test of agentic video understanding, and that semantic retention remains the main challenge.
Read the full abstract
Quantum Physics
Engineering dissipation and control pulses for high-fidelity fault-tolerance quantum computing
Cat-state qubits, a prominent class of bosonic encodings, offer a promising pathway toward hardware-efficient fault-tolerant quantum computing. In this manuscript, we propose an optimally robust control protocol for the cat-state qubits which are stabilized by engineering two-photon dissipation. By deriving an effective two-level description in the cat-state subspace and applying shortcut-to-adiabaticity via inverse engineering, we design a robust protocol to achieve fast and high-fidelity state transfer in the cat-state qubit. We analyze the sensitivity to systematic control errors and identify an optimal robustness condition that strongly suppresses errors induced by imperfections in the driving fields. Furthermore, we show that dissipative confinement efficiently suppresses leakage out of the cat-state subspace caused by the pure dephasing, highlighting an intrinsic advantage of dissipative-cat qubits. This work establishes a robust and leakage-suppressing framework for high-fidelity bosonic qubit control, offering a promising route toward scalable fault-tolerant quantum computing.
Read the full abstract
Computation and Language
Temperature Fragility and the Conditional Benefits of Truncation Sampling
Large language models generate text by sampling each token from a predicted distribution, and a temperature parameter sets how far the draw strays from the most probable tokens. Truncation samplers such as top-p and min-p discard the least probable tokens before the draw, so that sampling at high temperature stays coherent. Their reported accuracy gains come from temperatures of 1.5 to 3, while the defaults of deployed systems cluster between 0.6 and 1.0. Whether they change accuracy at those defaults, and for which models, has not been measured. We test thirteen open-weight models on GSM8K and MMLU-Pro at temperatures 0.7, 1.0, and 1.3 in one controlled pipeline, ten of them under eight decoding configurations. Six of the thirteen models lose 17 to 38 accuracy points on MMLU-Pro between 0.7 and 1.3, and the other seven lose at most 10. The lost accuracy comes from generations that run to the token limit or never state an answer. These results suggest that truncation samplers improve accuracy primarily when higher temperatures substantially degrade model performance. Where accuracy remains stable across temperatures, none of the tested truncation samplers improves on plain temperature sampling.
Read the full abstract
Robotics
P-POSEMEM: Projective Semantic Memory for Consistent Language Grounding under Pose-Graph Rewrites
A robot following language instructions needs its semantic memory to keep naming the same physical object while the SLAM pose graph underneath is optimized, loop-closed and compressed. Maps committing each detection to a world coordinate cannot: a closure moves the anchor it was measured from, or the solver marginalizes that anchor, and the query then selects a different object although both graphs represent the same posterior. P-POSEMEM stores each observation as an immutable event at its birth keyframe, retains the Bayes-tree elimination conditional of every marginalized keyframe, and integrates the semantic likelihood over the reconstructed joint posterior of poses, anchors and identities. Dproj, the total-variation defect between the language-goal distributions of inference-equivalent full and marginalized graphs, measures this directly. Over 40 HM3DSem scenes and 112,000 queries, P-POSEMEM reproduces the full-graph oracle (Dproj = 0) and reduces goal flips against every memory-reducing baseline. On an eight-run campaign whose 761 closures rewrote the map by up to 47 m, Dproj stays below 10^-13 with 0/288 goal flips when elimination follows the closures, where every ablation and a coordinate committed at insertion flip goals it does not; under a live bounded solver the same memory flips 23/288 against 53 for that frozen coordinate. A pre-registered negative control is detected by Dproj while leaving calibration error and navigation success unchanged, indicating that these measures capture distinct failure modes. Retrieval is held fixed by a shared frozen detector, isolating the gain to memory consistency. Code and data: https://anonymous.4open.science/r/posemem-2328/.
Read the full abstract
Mesoscale and Nanoscale Physics
Two Strategies to Measure Spin-Orbit-Torque Efficiency Acting on the Insulating Magnet Li$_{0.5}$Al$_{0.7}$Fe$_{1.8}$O$_4$
For insulating magnets, widely-used electrically-based techniques to measure spin-orbit-torque efficiency can suffer from poor signal-to-noise ratios and unwanted artifacts, while optical measurements based on conventional magneto-optical Kerr effect (MOKE) imaging often lack sufficient sensitivity. Here we analyze two strategies for measuring the spin-torque efficiency acting on an insulating magnet, focusing on the torque from Pt acting on thin films of the low-damping magnet lithium aluminum ferrite (LAFO). First, lateral + longitudinal spin-torque ferromagnetic resonance (ST-FMR) allows separation of the spin-torque signal from artifacts due to spin-pumping and resonant heating. With this technique, we measure a spin-torque efficiency $\xi^\text{ST-FMR}_\text{DL}\sim\ 0.07$, comparable to but a bit less than for Pt acting on metallic ferromagnets. Second, we attempted optical measurements of spin-orbit-torque-induced magnetic deflection using Sagnac interferometry. We find that Sagnac interferometry provides sufficient sensitivity to measure signals from 14.4 nm LAFO thin films, but the inferred torque efficiency is much weaker than the result from lateral + longitudinal ST-FMR. We speculate that there may be another contribution to the Sagnac signal other than current-induced magnetic deflection, perhaps an out-of-plane component of itinerant electron spins. This becomes relevant for insulating magnets in which the magneto-optic coupling coefficient is very weak.
Read the full abstract
Classical Analysis and ODEs
The lower envelope of ultraspherical polynomials
Let $R_n^{(a,b)}=P_n^{(a,b)}/P_n^{(a,b)}(1)$ be the normalised Jacobi polynomials. For $\alpha\ge0$, put $\xi_0=-1$ and let $\xi_n$ be the largest zero of $R_n^{(\alpha+1,\alpha)}$ for $n\ge1$. We prove that $$ R_k^{(\alpha,\alpha)}(x) =\min_{n\ge0}R_n^{(\alpha,\alpha)}(x), \quad \xi_{k-1}\le x\le\xi_k, $$ for every $k\ge1$. This determines the pointwise lower envelope of the normalised ultraspherical polynomials and answers Question 2 posed by F. M. de Oliveira Filho in his 2009 thesis. For $\alpha>0$ and $-1<x<1$, we also determine all minimising degrees. The sequence is non-increasing up to each such degree, which answers his Question 1 in this range. A Legendre example gives a negative answer when $\alpha=0$.
Read the full abstract
Human-Computer Interaction
Spook the Machine: Gamified Exploration of Human Imagination of Machine Fear
What happens when AI machines express fear? Do humans engage differently depending on how they express it? And what does it take to design for affective human-AI interaction? We present Spook the Machine, a gamified platform where participants generate images to frighten AI agents endowed with personality-driven phobias. Machines respond with emotional reactions ranging from calm analysis to begging for mercy, and a gallery of successful scares becomes visible to subsequent users. In a public deployment during Halloween 2024, 832 participants created 15,719 artifacts across 89 machines in a $2\times2$ design varying the machine's emotional expressiveness (neutral vs. high-emotion) and reward structure (rewarding scariness alone vs. scariness plus novelty). Emotionally expressive machines deepened engagement at moments of failure: users deliberated longer even when the machine did not express fear, and learned faster from the gallery, yet their creative output remained unchanged across all measures. Rewarding novelty sustained collective creative diversity over time; without it, users increasingly repeated what had previously worked. Each machine developed its own trajectory through accumulated social learning, with the gallery shaping what participants created next. These findings show that emotional expression and reward design are complementary levers for steering collective human-AI interaction: emotional expression shapes how deeply users engage, while reward structure shapes how they explore.
Read the full abstract
Artificial Intelligence
HISPO: Hierarchical Importance-Sampling Policy Optimization with Entropy-Derived Segments
Reinforcement learning with verifiable rewards (RLVR) has become a central approach for improving mathematical reasoning in language models, but long-form completions introduce a difficult credit-assignment problem: different parts of a solution trace may contribute unevenly to final correctness. Existing policyoptimization objectives for RLVR commonly apply importance-sampling correction at either the token level (GRPO, DAPO) or the sequence level (GSPO), imposing different granularities for assigning credit across a response. We introduce Hierarchical Importance-Sampling Policy Optimization (HISPO), a segment-level policy-optimization method that constructs rollout-time entropy-derived contiguous segments, assigns soft entropy-based saliency weights, and applies clipped importance-sampling correction at the segment granularity. This provides an intermediate correction unit between token-level GRPO/DAPO and sequence-level GSPO. We evaluate HISPO by fine-tuning Qwen3-1.7B-Base on mathematical reasoning tasks. Across six benchmarks, HISPO improves Pass@8 over the strongest baseline on all benchmarks and matches or exceeds the strongest baseline in Acc@8 on five of them. On AIME25, HISPO improves over GRPO by +3.75 Acc@8 and +3.78 Pass@8, and over GSPO by +2.50 Acc@8 and +1.27 Pass@8. These results suggest that segment-level correction is a promising granularity for RLVR in long-form mathematical reasoning.
Read the full abstract
Dynamical Systems
Memory-induced blow-up solutions and their dynamical transitions in distributed delay differential equations
In this paper, we investigate finite-time blow-up solutions for distributed delay differential equations incorporating memory effects with a specific gamma distribution kernel. Focusing on typical nonlinear terms that cause finite-time singularities in ordinary differential equations (ODEs), we examine how memory effects change the existence, rate, and qualitative properties of blow-up. By comparing these behaviors with those of the corresponding non-delayed and memory-free ODEs, we show that time delay originating from memory not only essentially induces finite-time blow-up (“memory-induced blow-up”), but also drastically alters the blow-up profile, such as accelerating algebraic blow-up rates or driving a qualitative transition from ODE quenching to logarithmic blow-up. Furthermore, by incorporating a self-inhibitory term, we reveal a threshold phenomenon in the phase space that governs the occurrence and non-occurrence of blow-up depending on the initial conditions and parameters. These results are established by reducing the system to a two-dimensional ODE via the linear chain trick, and analyzing the dynamics at infinity using Poincaré-type compactification, blow-up techniques, and the center manifold theorem.
Read the full abstract
Optics
Automatic control of gold nanoparticle clustering in plasmonic nanopores based on Raman spectroscopic feedback
Plasmonic nanopores have been used to trap single nanoparticles using electric bias for single-molecule surface enhanced Raman spectroscopic (SM-SERS) detection and sequencing. While they exhibit high sensitivity, removal of nanoparticle clusters in the nanopore is so labor intensive that it hampers biosensing applications. Here, our approach is to use a microcontroller to change the electric voltage to remove the nanoparticle clusters in the nanopore with feedback by the Raman spectra. The SM-SERS spectra of the molecules adsorbed on nanoparticle surface can be acquired over a sufficiently long period, so that the microcontroller can analyze them with <100 ms and trigger the voltage change to remove the nanoparticle cluster. Our system is promising for automatic and high-throughput SM-SERS detection and analysis of DNA and proteins.