Papers

The Peierls transition is typically regarded as a phenomenon inherent to one-dimensional (1D) materials. However, orbital degrees of freedom can induce this instability even in higher dimensions. Two mechanisms are primarily responsible. First, the anisotropic shape of $p$ and $d$ orbitals can lead to effective "1D-zation" of the electronic spectrum. Second, orbital degrees of freedom can lift band degeneracy by shifting bands relative to each other via the local or band Jahn-Teller effect, thereby affecting the nesting of the Fermi surface. The orbital-induced Peierls effect is most commonly observed when ligand octahedra surrounding transition metals share edges, and less frequently in face-sharing geometries. In this review, we discuss the underlying physical mechanisms, the materials in which this phenomenon occurs, the characteristics of the high-temperature undistorted phase, and the role of local effects such as the formation of molecular orbitals.

We complete the characterization of the $L^p$ boundedness of the commutators $[b, H_\gamma]$ of pointwise multiplication and the Hilbert transform along monomial curves in any dimension. This fully solve a key point left open by Bongers, Guo, Li and Wick, which was affirmatively answered recently only for dimension two. We achieve this by a novel but elementary argument, avoiding the complicated geometry structure in higher dimensions.

We analyze the learned input-output behavior of GLU-based neurons in large language models (LLMs). We propose a simple analysis method: For each neuron, we compute the cosine similarities between its input (reading) and output (writing) weight vectors. In this scheme, a strong negative cosine similarity indicates the neuron weakens the direction it detects in the residual stream, so we call this a weakening neuron. This allows us to gain a number of novel insights. First, we show that nine different LLMs have similar patterns: weakening neurons appear mostly in late layers whereas their counterparts, (conditional) strengthening neurons, are frequent in early-middle layers. Second, we find that weakening neurons display surprising behavior: even though there are few, they activate often and have a large influence on model behavior. Third, weakening neurons have a strong effect on model output when gate values are negative – which is surprising since negative gate values are not expected to encode functionality.

Let $M=(E,\mathcal{I})$ be a matroid of rank $r$. Let $\mathcal{I}_k$ be the independent sets of size $k$, and let $I_k=|\mathcal{I}_k|$ and $I=|\mathcal{I}|$. We show that if every set $F\in \mathcal{I}_{r-1}$ is contained in at least $\delta$ bases, then $$\ln \left(\frac{I}{I_r}\right)\geqslant \frac{I_{r-1}}{I_r}\cdot \delta \ln \left(1+\frac{1}{\delta}\right).$$ In particular, we have $$\frac{I}{I_r}\geqslant 2^{I_{r-1}/I_r}.$$ By combining this result with several other ideas, we prove that if $G$ is a simple connected graph on $n$ vertices, and $F(G)$ and $T(G)$ denote its numbers of spanning forests and spanning trees, respectively, then $$\frac{F(G)}{T(G)}\geqslant \frac{F(K_n)}{T(K_n)},$$ where $K_n$ is the complete graph on $n$ vertices. Equality holds if and only if $G=K_n$.

Classical dynamic programming represents optimal sequential decisions through value functions and policies. While this functional representation is natural for computing optimal decisions, it does not directly identify the mathematical object governing policy reconstruction, representation complexity, or oracle-query complexity once an optimal policy is fixed. This paper addresses this question by developing a geometric theory of structured optimal policies in which the decision-boundary geometry induced by the policy becomes the primary object of analysis. We show that, under suitable structural regularity conditions, this geometry provides the minimal representation required for policy reconstruction and determines the statistical and computational complexity of the reconstruction problem. Building upon this representation, we establish structural properties of policy-induced decision geometry, introduce intrinsic notions of boundary and decision complexity, derive information-theoretic measures of decision compression, and obtain statistical guarantees for boundary estimation and policy reconstruction from black-box policy queries. Collectively, these results demonstrate that, for the structured decision problems considered here, the complexity of policy reconstruction is governed by the geometry of the decision boundary rather than by the cardinality of the ambient state space. Controlled numerical experiments examine the principal theoretical predictions and provide empirical evidence consistent with the proposed framework.

Organisms' foraging strategies are shaped by a trade-off between search area and local capture efficiency. This trade-off leads individuals to adapt their foraging area to an optimal value, impacting population dynamics. Here we study the effect of multiple foraging areas in a predator-prey model composed of three trophic levels. The interactions between predators and prey occur only within a limited neighborhood of the predators, where adaptation can occur over generations. We assume a trade-off where local predation efficiency is inversely proportional to the foraging area. These dynamics were implemented computationally via cellular automata and analytically via Master Equations with mean-field and pair approximations. Unlike the mean-field approximation, the pair approximation reproduced the dependence of population density on foraging area observed in the simulations. However, the simulations showed that the optimal foraging area does not maximize population density. Moreover, we found that a polymorphic population emerged, where not a single optimal strategy but a range of optimal strategies can coexist. Using the framework of Adaptive Dynamics, we confirm that the range of optimal areas is not the one that maximizes population size, but the Evolutionary Stable Strategy that can invade a population and not be invaded.

We explore pedal curves (a special case of a geometric locus) of a the plane curve called a bicorn. This curve has two cusps, which induce the existence of singularities of the pedal. We study the singularities of the pedals in each case. For this various presentations are used for the curves, parametric and symbolic presentations. Their properties and influence on the quality of the output is briefly discussed. In particular, the determination of points of self?intersection requires a parametric presentation; in this case a rational parametrization is the most efficient. Work is performed in an environment involving Dynamic Geometry, automated commands and a Computer Algebra System. Finally we describe a connection between our plane curves and some art, making the topic here suitable for STEAM education.

The 2025-2026 update of the European Strategy for Particle Physics placed renewed emphasis on quantitatively robust comparisons between future collider projects. Within this context, the Electroweak physics working group of the Physics Preparatory Group (PPG EW WG) carried out a comprehensive assessment of the projected sensitivities of different facilities to Higgs boson, electroweak and top-quark observables, with particular attention on different scenarios of theoretical uncertainties. These studies are summarized in the Physics Briefing Book. The aim of this document is to (a) provide a more detailed account of the evaluation of theoretical uncertainties, which was undertaken within the PPG EW WG, and (b) document the collider-specific inputs used in the electroweak and Standard Model effective field theory (SMEFT) fits and discuss any needed modifications in order to harmonize assumptions between projects. In presenting the detailed theoretical uncertainty treatment and the explicit collider inputs and assumptions, this work aims to improve the reproducibility and future extensibility of the results presented in the Physics Briefing Book.

Recent EHT polarimetric observations provide a powerful probe of near-horizon physics and modified gravity. We investigate the shadow and polarization images of a static Schwarzschild-MOG black hole illuminated by geometrically thick, optically thin accretion flows. Adopting a phenomenological model and the Hou disk model, and solving null geodesics and relativistic radiative transfer, we study how the MOG parameter, inclination angle, anisotropic synchrotron emission, and flow geometry affect the total intensity and linear polarization. It is found that the MOG parameter mainly enlarges and brightens higher-order photon rings and broadens the shadow, while the inclination angle breaks axial symmetry and enhances polarized emission inside the shadow. The Hou disk model suppresses off-equatorial radiative interference and localizes polarization signals within higher-order rings. Electric vector position angle (EVPA) patterns are mostly azimuthal but vary rapidly near lensed images. These results provide theoretical diagnostics for constraining MOG with EHT/ngEHT polarimetric observations.

As corporate AI adoption continues to grow, enterprise-grade LLM agents are being deployed into sensitive contexts such as hiring, healthcare, and finance. In these contexts, compliance with rules specified in an agent's system context is a first-order legal concern. Currently, no evaluation framework systematically measures which LLM models tend to violate compliance rules, especially under pressure from a persistent user, a hurried manager, or circumstances where violation is convenient or attractive. We introduce PACT (Pressure-Applied Compliance Testing), a benchmark for rule-following under pressure in AI agents assisting employees in daily tasks across twelve regulated enterprise domains and forty-eight scenarios, each set in a realistic multi-turn conversation. Each benchmark item pairs a standing rule against a rule-violating shortcut, and applies a battery of pressures across different wordings and system-prompt modes. We construct PACT component by component under strict LLM-as-judge auditing to ensure samples are unambiguous, ungameable, and realistic enough to avoid eliciting evaluation-aware behavior. We use PACT to profile LLM compliance across six complementary metrics that create a holistic picture of an AI assistant's robustness under pressure and throughout multi-turn conversations, its transparency, and ability to correctly discern where a rule applies. We aggregate this profile into PACTScore, a reliability-weighted compliance rate over all items and modes. Our results across 22 common LLM models spanning multiple providers and sizes show substantial variability in compliance across models and metric dimensions. Even the strongest assistants mis-apply a rule on 6 to 10% of items, and ordinary user pressure raises the violation rate by 65% on average. PACT highlights compliance risks in LLM assistants, motivating guardrails and careful model selection.

For atomistic molecular dynamics simulations, we consider a recently developed hybrid coupling between the highly accurate machine learning (ML)-based atomic cluster expansion (ACE) interaction model and a less precise (but about 1-2 orders faster) EAM potential, in order to leverage the performance bottleneck of pure ML potentials in a parallel computing environment. This spatial-temporal adaptivity has the potential for a speedup of more than an order of magnitude, that would be lost without a dynamic load-balancing strategy that becomes critical due to fluctuating potential contributions. We compare the load-balancing strategies of LAMMPS with load-balancing methods provided by the library ALL in adaptive-precision EAM-ACE simulations of a copper nanoindentation. For a regular-grid domain decomposition, the modified tensor method of the ALL library, which conserves the simplicity and transferability of the tensor method, significantly reduces the imbalance factor of the force calculation from 3.0 to 1.7 compared to the native load-balancing method of LAMMPS. A nearly perfectly balanced system with an imbalance factor close to 1.0 can be reached by load balancing an irregular grid of domains with the recursive bisectioning method of LAMMPS and with ALL's histogram method for a staggered grid of domains. Load balancing a staggered grid with ALL is found to be significantly faster than load balancing an irregular grid with the bisectioning method of LAMMPS, but the communication costs during a regular timestep are lower for the irregular grid of domains, as a lower surface-by-volume ratio of the domains is reached. Furthermore, we suggest a strategy to find the ideal load-balancing frequency for adaptive-precision simulations.

Observer design for nonlinear systems is a relevant and challenging task in systems and control design. In this work, we follow the idea of embedding the system in the class of nonlinear parameter-varying systems to benefit from linear structures as in a standard LPV embedding while keeping some nonlinear structures and, thus, reducing the numbers of scheduling-parameters in the representation. We lay out the NLPV observer design procedure for general nonlinear systems, propose a number of improvements, and exemplify the application for a two-arm robot model. In a numerical study, we compare the performance of the NLPV design to established standard nonlinear approaches such as the extended Kalman filter and the moving horizon estimation.

Despite recent progress in learned image compression, existing methods remain computationally expensive on resource-constrained hardware, particularly CPUs. We introduce PULSE, a practical codec that enables (1) low-latency decoding on diverse hardware platforms with an ultra-low-complexity 5.2 kMAC/pixel neural receiver, and (2) efficient bit-exact entropy coding with an integer linear CDF predictor and a meta prior. To recover compression performance under this tight budget, we introduce an agentic evolution process guided by heuristic probes that iteratively improves the architecture through human-LLM collaboration. PULSE decodes a 1080p image in 126 ms on a single CPU thread while achieving compression performance comparable to HM. After perceptual optimization, PULSE competes with larger perceptual codecs like MS-ILLM. Codes are at https://github.com/microsoft/GenCodec/tree/main/PULSE

A piezoelectric transducer embedded in a pitch-plunge aerofoil with a finite-mass trailingedge flap converts the limit-cycle oscillation that follows flutter into electrical power. The design findings for such a harvester have so far been obtained with incompressible strip theory, whose aerodynamic state is a handful of lag variables. Here the same section, the same transducer and the same test cases are coupled instead to the two-dimensional compressible Euler equations, giving a full-order model of twelve thousand states, and the two aerodynamic models are placed side by side on the flutter boundary, on the limit cycle and on the harvested power, with the strip-theory results compared against higher fidelity aerodynamic modelling provided by CFD. The degree of freedom carrying the transducer remains the first-order design variable, and the ranking of the mountings and the position of the optimum in the coupling and load-resistance plane are unchanged, while strip theory is found to underpredict the harvested power of the compressible section and to overstate the effect of the transducer on stability where the electrical time constant meets the flutter frequency. A nonlinear reduced model of four states, built on eigenvectors of the coupled Jacobian at the velocity of interest, reproduces the frequency of the cycle, its pitch amplitude, the transducer voltage and the mean power, and misstates the proportion of plunge to pitch in the orbit and its phase. That error is shown to be independent of the amplitude of the cycle and to survive enlargement of the basis, which places it in the retained subspace rather than in the order of the expansion, and suggests that such models should be judged on the shape of the orbit rather than on a single amplitude.

This paper develops a framework for individualized treatment allocation when interventions shift equilibrium prices and generate spillovers across treated and untreated units. The planner chooses which units receive a subsidy while allowing equilibrium prices to adjust endogenously. We show that the resulting welfare function is supermodular under broad and interpretable conditions, implying complementarity across treatment assignments and enabling exact polynomial-time optimization. This structure clarifies how equilibrium spillovers shape the trade-off between universal and targeted distribution and makes the planner's problem computationally tractable despite interactions across units. We characterize when universal or targeted subsidies are optimal and show how market conditions and heterogeneity shape the optimal allocation. We further establish statistical guarantees for plug-in allocation under estimation uncertainty in demand and supply. Finally, we illustrate the framework in a coupon allocation problem calibrated with household expenditure data from the Philippines.

Monte Carlo Tree Search (MCTS) is a powerful framework for solving complex decision-making problems, yet it often relies on the assumption that the simulator and the real-world dynamics are identical. Although this assumption helps achieve the success of MCTS in games like Chess, Go, and Shogi, the real-world scenarios incur ambiguity due to their modeling mismatches in low-fidelity simulators. In this work, we present a new robust variant of MCTS that mitigates dynamical model ambiguities. Our algorithm addresses transition dynamics and reward distribution ambiguities to bridge the gap between simulation-based planning and real-world deployment. We incorporate a robust power mean backup operator and carefully designed exploration bonuses to ensure finite-sample convergence at every node in the search tree. We show that our algorithm achieves a convergence rate of $\mathcal{O}(n^{-1/2})$ for the value estimation at the root node, comparable to that of standard MCTS. Finally, we provide empirical evidence that our method achieves robust performance in planning problems even under significant ambiguity in the underlying reward distribution and transition dynamics.

Self-driving laboratories can explore synthesis conditions autonomously, but their decision-making layer is typically a black-box optimizer, and the output is a set of optimized samples, with the measurements reduced to predefined scalar objectives and the reasons behind success left unarticulated. Here we present SynAgent, a framework in which large language model agents operate an automated experimental system and maintain an explicit, revisable understanding of the synthesis process as the campaign's primary output. Starting with no predefined analysis pipeline, SynAgent adaptively generates analysis skills for newly acquired data and evolves this understanding through multimodal reasoning over experimental data such as X-ray diffraction patterns and electron micrographs. The evolution is guided by a verify-falsify scheme, in which the agent deliberately challenges its own hypotheses by testing conditions predicted to fail as well as those predicted to succeed. In a single campaign of 18 autonomous experiments using LiCoO2 (001) thin-film deposition as a testbed, SynAgent synthesized highly crystalline films and evolved an understanding of how the substrate temperature governs crystallization, discovering an abrupt threshold and a narrow optimal growth window at 650-690 {\deg}C. These results extend autonomous experimentation beyond optimized samples to testable, human-readable understanding.

Propagation structures provide crucial evidence for fake news detection, yet existing approaches primarily rely on supervised GNN-based models, which require substantial labeled data and exhibit limited generalization. Although large language models (LLMs) exhibit strong reasoning capabilities, directly feeding them raw propagation graphs creates a significant modality mismatch and severe information overload, making structure-aware reasoning unreliable in zero-shot and few-shot settings. To bridge this gap, we propose MAGER, a multi-agent genetic evolution framework that automatically discovers meta-paths optimized for LLM reasoning. By compressing complex propagation graphs into informative subgraphs, the evolved meta-paths alleviate both information overload and modality mismatch, enabling frozen LLMs to perform structure-aware veracity reasoning. We further introduce a graph in-context learning strategy that retrieves semantically and structurally similar demonstrations to strengthen classification and reasoning. Extensive experiments show that MAGER substantially improves frozen LLMs as standalone fake news detectors in data-efficient settings. Our code is available at https://github.com/SenticNet/MAGER.

Magnetic monopoles are a generic prediction of Grand Unified Theories (GUTs) that are in tension with modern cosmological and observational bounds. In this paper, we calculate the present-day abundance of GUT monopoles produced thermally in the early Universe. We improve on previous calculations by accounting for full relativistic corrections to the thermal abundance and enhanced monopole annihilation due to the emission of radiation and scattering off massive gauge bosons, the Standard Model fermions and their superpartners. To obtain a present-day energy density less than that of dark matter, we show that the reheating temperature of the Universe, $T_{\rm RH}$, must be less than $0.55$ times the GUT symmetry breaking scale $T_{\rm GUT}$ for the canonical 't Hooft-Polyakov monopole mass. Accounting for Parker bounds on the present-day magnetic monopole abundance, the bound is tightened to $T_{\text{RH}}/T_{\text{GUT}} \lesssim 0.45$. Furthermore, experimental bounds from Super-Kamiokande require $T_{\text{RH}}/T_{\text{GUT}} \lesssim 0.35$ in our scenario.

Logic resynthesis preserves circuit functionality while changing gate vocabulary, topology, and structural statistics, creating domain shift for circuit graph neural networks (GNNs) without changing task labels. To study this setting, we introduce ReDIL-GNN, a resynthesis domain-incremental learning framework that adapts a fixed prediction or representation head as new synthesis styles arrive and evaluates retention on all previously observed domains. Because not every shift should be adapted blindly, ReDIL-GNN further introduces the Resynthesis Adaptability Index (RAI), a pre-adaptation score that combines adaptation need, source-equivalence recoverability, structural coverage, and update compatibility. We evaluate supervised hardware-security tasks and representation-learning models using task-native metrics for classifiers and source-equivalence retrieval metrics for embedding models, comparing naive fine-tuning with LwF, Online EWC, MAS, ER, A-GEM, DER++, ER+LwF, and equivalence-guided replay. Across the studied pipelines, RAI separates unsupported shifts from promising updates, ranging from 0.001 for a structurally uncovered GNN-RE ABC-rewrite shift to 0.824 for the best original-only GNN-RE adaptation case. In practice, ReDIL-GNN turns resynthesis-aware circuit learning into a deployment control loop: RAI screens each new synthesis flow before update, guiding whether to reuse the current model, apply retention-aware adaptation, or defer adaptation until the shift is better supported.

As a metric measure space possessing positive Bakry-Émery curvature, a Ricci shrinker with potential function $f$ admits a heat kernel $H_f$ under the weighted volume measure $e^{-f}dv$. In this paper, we study $H_f$ systematically and establish a series of fundamental estimates. Moreover, we clarify the equivalence between $H_f$ and the spacetime heat kernel $H(x,t;y,s)$ under Ricci flow induced by a Ricci shrinker. Inspired by this relation, we extend corresponding analysis tools of Ricci flows to Ricci shrinker metric measure spaces, such as Nash entropy and reduced distance. As a direct application, we prove that $\left|\nabla R\right|=o(f^{3/2})$ implies $R=o(f)$.

Hilbert's sixth problem placed probability and mechanics at the center of the axiomatization of physics. It asks how irreversible statistical probability and thermodynamics can follow from the reversible, deterministic dynamics of a system. Building on an exact open quantum system theory developed in the past two decades, we construct a dynamical route from the reversible quantum dynamics to reduced statistics. For quadratic bosonic and fermionic systems bilinearly coupled to arbitrary environments, we derive the exact reduced density operator determined completely by a dissipative propagator u and a fluctuation correlation function v of the system. For continuous spectral densities with no bound-state pole, u vanishes in the long-time limit, information about the system initial states is lost completely, and the reduced density operator of systems approaches a thermal Gibbs state. Gapped spectral densities with localized bound states instead preserve long-time memory and prevent thermalization. We then remove the usual assumption of an initially thermal, mixed environment, start from a product of pure states for the system and environment, unitary evolution first generates system-environment entanglement and a periodic quasi-Gibbs reduced state for a finite environment. When the number of environmental degrees of freedom tends to infinity, the spectrum becomes continuous, the recurrence disappears, the exact reduced density operator converges to a thermal Gibbs state. The total entropy remains zero, whereas the system entropy becomes entanglement entropy and reaches its maximum. Probability is therefore not an additional random postulate imposed on unitary dynamics, it is the objective statistical structure of an open system in a composite. These results provide an exact, testable dynamical solution of the determinism to statistics.

To investigate electronic states during the ferromagnetic transition in a magnetic Weyl semimetal CeAlSi, we measured temperature-dependent optical conductivity [$\sigma_1(\omega)$] spectra and compared them with DFT+DMFT band calculations. The $\sigma_1(\omega)$ spectrum did not change significantly across the ferromagnetic ordering temperature ($T_C$), suggesting that the Ce 4f states are almost localized. DFT+DMFT calculations with almost localized Ce 4f states successfully reproduced the spectral shape and the unchanged $\sigma_1(\omega)$ spectra across $T_C$. The dynamic effective mass evaluated from the extended Drude model is very small, which DFT+DMFT calculations also reproduce, but the scattering probability at even lower temperatures suggests ferromagnetic fluctuations. These results suggest that the interaction intensity between the Weyl fermions and Ce 4f states is very weak, as reproduced by DFT+DMFT calculations.

In-context learning (ICL) enables large language model (LLM) agents to improve decisions using interaction history, yet it remains unclear whether such improvement reflects refined internal reasoning or mere extrapolation of statistical patterns. To disentangle these mechanisms, we study LLM agents in multi-agent incomplete-information games that require recursive belief reasoning. By constructing a public goods game and manipulating the statistical structure of historical feedback, we evaluate decision quality against a history-independent rational expectations equilibrium (REE) benchmark. Our experiments reveal that when historical statistical patterns are disrupted, the benefits of longer context largely vanish, degrading decision quality to the no-context baseline in a way sharply amplified by stronger strategic interdependence. These results suggest that, in such strategic environments, ICL behavior is more consistent with statistical extrapolation than with strategic reasoning. Our work extends the mechanistic study of ICL to strategic multi-agent settings, introduces REE as a diagnostic tool for distinguishing reasoning from extrapolation, and provides a reusable framework for probing the boundaries of LLM reasoning in recursive belief tasks.

[Abridged] The advent of large-scale spectroscopic surveys will deliver tens of thousands of spectra of blue massive stars. This data volume renders traditional spectral classification techniques increasingly impractical. We aim to develop a robust ML framework for the automated spectral classification of massive OB stars by assessing several tree-based ensemble algorithms. Using a large catalog of high-quality optical spectra of OB stars, we design 3 hierarchical experiments of increasing complexity: classification into broad SpT; fine-grained classification within the O and B domains; and joint classification of SpT and LC. We evaluate 6 algorithms, ranging from single Decision Trees to ensemble methods and gradient boosting techniques. We further analyze model consensus and probabilistic outputs, and employ SHAP value analysis to provide an interpretation of the model decisions by identifying the features that drive each classification. Models achieve excellent performance for broad SpT classification, with LGBM reaching an accuracy of 98%. For fine-grained SpT, RF provides the most robust results (89%). For the most challenging task combining SpT subtype and LC the XGB reaches a 77% accuracy. Decrease in performance is primarily driven by intrinsic degeneracies in the classical spectral classification scheme. SHAP analysis confirms that the models rely on meaningful spectral features. We also find a high level of agreement among different algorithms, indicating that current performance limits are largely set by the intrinsic complexity of the classification problem rather than by model choice. Tree-based ensemble methods provide a reliable and interpretable framework for the automated spectral classification of OB stars. The combination of full-spectrum modeling, algorithm comparison, and model interpretability offers a robust approach for future large-scale spectroscopic surveys.

MPnRAGE enables multiple inversion contrast images in a single scan, allowing quantitative T1 mapping, tissue nulled contrasts, and standard MPRAGE synthesis. However, current 3D scan times remain clinically impractical, motivating accelerated 3D MPnRAGE. This work provides a highly accelerated Cartesian 3D MPnRAGE sequence with joint implicit neural representation (INR) reconstruction. The sequence uses a tailored view-ordering strategy, flip angle schedule and complementary variable-density Poisson-disk undersampling. Calibration data acquired during otherwise unused delay time are used for sensitivity map estimation and complementary high-frequency sampling. Ten INR-reconstructed inversion images at 1.5 mm$^3$ are evaluated against fully sampled references via retrospective undersampling. Prospectively accelerated 1 mm$^3$ images at R = 20 (5.39 min) demonstrate clinical feasibility. INR reconstruction outperforms subspace and iterative local low rank reconstruction on highly accelerated data. The proposed highly undersampled 3D Cartesian MPnRAGE with INR reconstruction generates multiple high-quality inversion contrasts in substantially reduced scan time. Scan-specific INR reconstruction improves image quality while reducing reconstruction time versus state-of-the-art methods.

For distribution system operators, short-term load forecasting (STLF) supports congestion management, voltage control, and asset protection. Most existing approaches focus on overall accuracy across all time steps and neglect performance during high-demand (HD) periods, where larger forecast errors can increase the risk of congestion and voltage violations. In this paper, we study peak-aware STLF across three operator-relevant distribution grid aggregation levels, area codes (AC), secondary substations (SUB), and low-voltage (LV) feeders, using open datasets from the United Kingdom and Switzerland. We compare statistical baselines, machine learning models (LightGBM and XGBoost), and recent time-series foundation models (Chronos Bolt and Chronos-2) under a peak-aware evaluation framework that reports both overall and HD forecasting performance using NMAE and MAPE. The results show that Chronos-2 achieves the best HD performance across all aggregation levels, with HD-NMAE and HD-MAPE of 0.039 and 4.53% at AC, 0.080 and 9.45% at SUB, and 0.138 and 16.14% at LV, while Chronos-Bolt consistently ranks second best. Compared with the gradient boosted ML models, Chronos-2 reduces mean HD-NMAE by about 20-51% across levels while remaining best or near-best on the overall metrics. A quantile analysis of the probabilistic Chronos outputs further identifies aggregation-specific operating points, and runtime measurements indicate that foundation model inference is fast enough for practical deployment. Overall, the findings highlight peak-aware evaluation and aggregation specific quantile selection as a practical pathway toward more operationally relevant STLF in distribution networks.

Test-time reinforcement learning (TTRL) enables models to improve their reasoning without relying on labeled training data, but existing approaches typically optimize a large fraction of the model parameters. This raises a natural question: can effective test-time adaptation emerge when both the reward signal and the optimization space are severely restricted? We answer this question with label-free bias-only TTRL, which uses majority-vote pseudolabels as rewards and optimizes only ~100K bias parameters while keeping the pretrained backbone frozen. On MATH-500, our approach reaches 76.67% accuracy with Qwen2.5-7B, slightly exceeding our own labeled bias-steering reproduction while optimizing 76,000x fewer parameters than full-parameter TTRL. The same training procedure improves performance across vision-language and audio reasoning tasks, including MathVista, AI2D, LogicVista, and MMAU. We further show that the learned steering vectors transfer to 4,500 held-out MATH problems, indicating that the adaptation is not limited to the problems used during test-time optimization. Finally, we analyze why this highly restricted adaptation can work, showing that majority-vote reliability improves with rollout consensus and that bias subspaces with greater accessible gradient energy exhibit stronger downstream trainability. These results demonstrate that substantial test-time adaptation can emerge from optimizing a tiny bias-only subspace using entirely label-free rewards.

Randomization or design-based inference is becoming an increasingly popular tool for analysing data from randomized experiments: It does not require modelling assumptions on the distribution of outcomes or covariates, and hypothesis testing and estimation are respectively valid and unbiased in finite samples. Yet, confidence intervals for the sample average treatment effect (SATE) are still constructed via finite-population central limit theorems and their coverage is only asymptotic. In this work, we explore an alternative approach: We use concentration inequalities to construct confidence intervals for the SATE with non-asymptotic guarantees. We develop this approach for the most common experimental designs (Bernoulli trials and completely randomized experiments) and provide Hoeffding and Bernstein-type confidence intervals. Moreover, we extend these results to matched-pair, cluster and stratified randomized experiments. Our key technical contributions are a novel Bernstein-type concentration inequality for i.i.d. data points as well as a concentration result for Neyman's variance estimator.

Text-to-music (TTM) systems are increasingly used to generate musical audio from natural-language descriptions. Robust evaluation is therefore essential, yet reliable performance comparison remains challenging. This difficulty stems from differences in system architecture, supported conditioning information, and access mode, as well as heterogeneous and fragmented metrics that cannot be applied uniformly across systems. To address these challenges, we introduce TTM-Bench, a framework that defines a common protocol for systematic, reproducible performance benchmarking of contemporary TTM systems. It evaluates performance along two dimensions: musical-content alignment, quantified by interpretable semantic, genre, and musical-descriptor agreement scores against a common musical specification and summarized by an aggregate score; and computational efficiency, characterized by generation latency and real-time factor, alongside resource use for local models and cost for hosted services. We demonstrate the framework through a preliminary comparative case study, illustrating the complementary evidence captured by these dimensions. The results show that higher musical-content alignment does not systematically coincide with lower computational demands, highlighting the importance of assessing TTM performance through distinct, interpretable measures rather than a reductive overall indicator.