Analysis of PDEs
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Tuesday 8 September
We study the $p$-Laplacian quadrature surface free boundary problem on a compact Riemannian manifold $(\M,g)$ of dimension $N \geq 2$, for $1<p<\infty$. Given a nonnegative source $f \in L^p_g(\M)$ with compact support $K$, and $k>0$, one seeks a domain $\Omega \supset K$ on which the overdetermined problem $-\Delta_{p,g} u = f$ in $\Omega$, $u=0$ and $|\nabla_g u|_g = k$ on $\partial\Omega$ admits a solution. Following the geometric and variational framework of Djité–Seck, we prove that the integral condition \[ (NS)_{p,g} \qquad \int_C f\,dv(g) > k^{p-1}\,|\partial C|_g, \] where $C$ is the totally convex hull of $K$, is sufficient for the existence of a solution strictly containing $C$. We then show, by an explicit counterexample on the round sphere $S^n_R$, that $(NS)_{p,g}$ is not necessary in general under nonnegative curvature: the mechanism is that the perimeter is not monotone with respect to inclusion in positive curvature. Finally, we identify an additional geometric hypothesis—perimeter monotonicity on the admissible class inside a normal convex neighborhood of $C$—under which necessity is restored, and we discuss the critical equality case $\int_C f\,dv(g) = k^{p-1}|\partial C|_g$.
Game Theory
Stable Voting Rules on the Edge of Optimal Metric Distortion
We prove the existence of a randomized voting rule with metric distortion at most $2.1248$, within $0.013$ of the lower bound of $2.11264$. Our rule comes from a generalization of \emph{stable $k$-lotteries} developed in the context of committee selection. In contrast to prior work, our rule samples from a single distribution derived from a zero-sum game, without mixing between voting rules. Our result also gives sharp distortion bounds for stable $k$-lotteries. In particular we show that stable $2$-lotteries have distortion $7/3$, despite only relying on aggregate preferences over triples of candidates, and that stable $2$-lotteries are optimal in this class of voting rules.
Read the full abstract
Artificial Intelligence
Revoked but Still Authoritative: An Empirical Study of Revocation Enforcement in Agent-Memory Systems
Long-running language-model agents depend on persistent memory. Many agent-memory systems preserve history through soft revocation: a contradicted fact is marked invalid and retained rather than deleted. However, whether that mark is enforced at retrieval time is unexamined. In this paper, we measure five such systems: we load each with a revoked policy and its replacement, track whether the revoked fact is returned at retrieval and whether the agent then acts on it across nine policy scenarios and nine models, and score every trial under six defense conditions. We find that no system enforces revocation by default: the revoked fact is returned wherever the revocation label is visible to the retrieval layer, outranks its replacement, and leads agents to the unsafe action. Based on these findings, we develop a guard that sits between the agent and any memory backend and withholds records that are revoked or conflict with their replacement.
Read the full abstract
Mathematical Physics
Asymptotics of Ground States for Helium-Like and Bosonic Atoms in a Two-Cluster Region
We study the asymptotic behavior of positive ground states of $N$-particle atomic Coulomb Hamiltonians in a two-cluster region. For the $N$-particle ground state $\psi$, its one-particle density $\rho$, and $(N-1)$-particle ground state $\phi$, we show that for $0<\alpha<1$, as $|x_N|\to\infty$, $\psi(x_1,\dots,x_N)/\sqrt{\rho(x_N)}= \phi(x_1,\dots,x_{N-1})+O(|x_N|^{-2+\alpha} \ln |x_N|)$ uniformly in the region $|x_i|\le |x_N|^{\alpha}$. In the region $|x_i|\le R_0\ln |x_N|$, the convergence rate is $O(|x_N|^{-2} (\ln |x_N|)^2)$. We also establish an $O(|x_N|^{-2})$ bound on the squared $L^2$ error. These results hold both below the essential spectrum and at threshold.
Read the full abstract
Cryptography and Security
ACEA: An Adversarial Co-Evolution Arena for Head-to-Head Red-Team and Blue-Team LLM Testing
Automated red-team attacks and blue-team defenses for large language models (LLMs) are advancing quickly. However, attackers and defenders are built and tested in isolation, and the resulting scores are hard to trust. To tackle this, we present ACEA (Adversarial Co-Evolution Arena), a platform that connects a pluggable red-team adapter and a pluggable blue-team adapter to a shared target LLM and scores their attack and defense rates with an LLM judge. ACEA contributes four components. First, a pluggable, model-agnostic arena. Any red or blue project connects over a minimal HTTP protocol, which we call the ACEA Standard Adapter Protocol (ASAP). It can be written in any language, and a project that exposes nothing but the protocol is a full participant. Second, an evaluation methodology built for adversarial rounds. Seeding the target with canonical secrets gives verifiable ground truth that separates real leakage from hallucination. We also send each attack to the target even when the defense blocks it, which measures the attack's raw potency independently of whether it was stopped. Together these yield a per-round decomposition of attack strength and defense effectiveness. Third, a real-time, game-style visualization with a detailed end-of-battle report that localizes each failure. The evaluation thus becomes an actionable signal for improving a red or blue project. Fourth, an optional in-context improvement loop that turns each round's outcome into advisory hints for the next. An adapter can then adapt across rounds without keeping state, provided it reads the hints. We describe the design of ACEA and the metrics through which red and blue teams are scored head to head.
Read the full abstract
Analysis of PDEs
Existence and strong-field asymptotics of skyrmions in a fourth-order model of frustrated ferromagnets
We study a fourth-order variational model for two-dimensional frustrated ferromagnets with competing exchange interactions and an applied magnetic field of strength $H>0$. For every $H>1/4$, we prove that the energy admits minimizers in the topological classes $Q=\pm1$ and that every minimizing sequence is precompact in $H^2$ modulo translations. The main difficulty is that spectral coercivity degenerates as $H\downarrow1/4$. Using a Helmholtz circle-mean identity, we prove that the residual energy of every nonzero-degree configuration has a uniform positive lower bound, even at the degenerate endpoint. Together with a sphere-valued $H^2$-splitting construction, this provides a threshold-stable binding inequality and yields compactness throughout the coercive regime. We also identify $H=1/4$ as the sharp spectral threshold. Below it the energy is unbounded from below, whereas at the threshold nonzero-degree configurations retain a positive energy barrier and degree-zero Weyl sequences lose compactness. Finally, in the strong-field regime, rescaled minimizers approach those minimizers of the limiting functional that maximize the Dirichlet energy, while topological-charge and normalized-energy measures concentrate on the $H^{-1/4}$-scale.
Read the full abstract
Artificial Intelligence
CircuTutor: Transforming Static Circuit Problems into Intelligent and Dynamic Tutoring
Learning direct current circuit concepts requires learners to connect invisible physical quantities, such as current, voltage, resistance, and power, with observable outcomes such as bulb brightness. Conventional textbook materials and general-purpose circuit simulators provide opportunities for problem solving and exploration but offer limited support for explaining why circuit behavior changes or diagnosing the reasoning behind incorrect answers. We present CircuTutor, a circuit-state-driven intelligent tutoring system that transforms static textbook circuit problems into an interactive tutoring workflow. CircuTutor first uses multimodal problem parsing to extract the textbook question, circuit topology, component parameters, switch states, and answer options, which are converted into a structured task and validated through circuit simulation. Learners can then interactively explore the circuit (by changing parameters) and submit an answer while a SPICE-compatible solver computes physically consistent circuit states. After the learner submits an answer, CircuTutor presents a before-and-after circuit state animation corresponding to the selected operation, organizes the simulated state changes into a causal reasoning chain that explains the underlying circuit behavior, maps answer discrepancies to likely misconceptions, and generates adaptive follow-up exercises targeted at the diagnosed misconception. Our experimental results demonstrate that CircuTutor effectively improves conceptual learning and the overall learning experience. The proposed framework demonstrates how simulated circuit states can be transformed into intelligent and interactive tutoring for circuit education, with the potential to generalize to other STEM domains.
Read the full abstract
Machine Learning
Revisiting Spectral Representations in Generative Diffusion Models
Diffusion models have shown remarkable performance on diverse generation tasks. Recent work finds that imposing representation alignment on the hidden states of diffusion networks can both facilitate training convergence and enhance sampling quality, yet the mechanism driving this synergy remains insufficiently understood. In this paper, we investigate the connection between self-supervised spectral representation learning and diffusion generative models through a shared perspective on perturbation kernels. On the diffusion side, samples (e.g., images, videos) are produced by reversing a stochastic noise-injection process specified by Gaussian kernels; on the spectral representation side, spectral embeddings emerge from contrasting positive and negative relations induced by random perturbation kernels. Motivated by this, we propose a self-supervised spectral representation alignment method to facilitate diffusion model training. In addition, we clarify how joint spectral learning can benefit diffusion training from a geometric perspective. Furthermore, we find that the optimization of the spectral alignment objective is in an equivalent form of diffusion score distillation in the representation space. Building on these findings, we integrate a spectral regularizer into diffusion training objectives to improve the performance of diffusion models on multiple datasets. Experiments across images and 3D point clouds show consistent gains in generation quality. Code is released at https://github.com/yuehaowang/spectral-reg-diffusion.
Read the full abstract
Information Theory
Variational Bayesian Data Detection for Multiuser MIMO Systems Corrupted by Phase Noises
Phase noise (PN), arising from imperfect local oscillators, introduces multiplicative distortions that degrade the performance of communication systems. In uplink multiuser multiple-input multiple-output (MIMO) systems, this impairment is further compounded by the presence of independent oscillators at each transmit and receive antenna, each contributing an uncorrelated noise component. Existing PN compensation algorithms at the receiver either rely on linearization approximations that lose accuracy under severe PN conditions, or incur computational complexity that scales prohibitively with the number of antennas. To address these limitations, we propose a variational Bayes (VB) framework for joint PN estimation and data detection in uplink MIMO systems. We develop VB-based detectors that treat noise statistics as latent variables, and reformulate the inference problem by absorbing the transmitter PN into the transmitted signal, treating the resulting composite variable as the inference target. Under von Mises priors, this reformulation yields exact closed-form conjugate posterior updates, from which we derive an improved detector achieving superior performance at low complexity. Simulation results demonstrate that the proposed VB algorithm achieves lower symbol error rates than the Self-Interference Whitening (SIW) algorithm and conventional phase-noise-unaware detectors across a wide range of channel conditions, modulation orders, and PN severities, while remaining computationally scalable to large MIMO deployments.
Read the full abstract
Differential Geometry
A Fourier approach to Gromov's filling area conjecture
We prove that every compact connected Riemannian isometric filling $M$ of a circle of length $2\pi$ satisfies $\operatorname{Area}(M) \geq \frac{14\zeta(3)}{\pi} \approx 5.35677$, regardless of orientability or topological types. Our new approach uses the odd Fourier coefficients of the distance functions from boundary points. For orientable fillings, we use a cubic resonant perturbation to obtain $\operatorname{Area}(M)>5.40154$.
Read the full abstract
Robotics
CALIPER: Clean Scenes Cannot Rank Physical Inference in Pretrained Visual Representations
How far a pushed object slides depends on its mass and friction, which no single image reveals. Pretrained visual encoders are increasingly used as the perception front end of world models for manipulation, and their physical competence is assessed with perturbation benchmarks and linear probes, almost always in a clean, fixed-camera scene. We show that these assessments cannot distinguish an encoder that infers physics from one that does not. CALIPER (calibrate, then predict) is a direct test: an object of unknown mass and friction is struck twice at known speeds, a third strike is shown only up to the moment of contact, and a linear readout on frozen features must predict how far the object slides. Swapping in another object's calibration clips checks that the evidence is actually used. Across 2,000 simulated episodes and eight representations, from V-JEPA 2 to a randomly initialised ViT and raw pixels, calibration adds +0.50 R^2 and the swap removes it. Yet in the clean scene every representation lands within 0.02 R^2 of the ceiling set by true simulator state, because a fixed camera exposes the object's displacement directly in pixel coordinates. Resampling camera, lighting, and clutter for every clip spreads the same representations across 0.50 R^2; when the readout chooses a push speed for a goal distance, V-JEPA 2 misses by 4 mm and the random ViT by 20 mm, no better than ignoring the object. Linear probes track none of this: a change in frame aggregation moves a probe more than pretraining does, and erasing the probed mass direction from the same representation costs nothing in one scene and 0.35 R^2 in the other. Whether a benchmark can rank models is an empirical property, and we give three checks that establish it.
Read the full abstract
Classical Physics
Projection Angles of Projectiles in Sports: Qualitative Assessment of the Effects of Aerodynamic Forces or Run-Up
We examine two major factors that influence the optimum projection angle: aerodynamic forces and the effect of run-up. With respect to aerodynamics, we consider not only the drag but also the lift generated by spin during flight. By linearizing the equations of motion that include these forces, we derive perturbation solutions with respect to drag and lift coefficients and clarify their qualitative effects. The results show that both drag and lift reduce the optimum projection angle, with the latter exerting a stronger influence. To investigate the effect of run-up, we use an extended projection model in which the initial speed depends on the initial angle. Analysis of this model reveals that a stronger run-up increases the relative projection angle but decreases the launch angle observed from the ground. These findings provide a mechanical explanation for the release angle in shot put and the takeoff angle in long jump. The present study establishes a simple theoretical framework for clarifying the respective roles of aerodynamic and run-up effects in determining the optimum projection angles in sports.
Read the full abstract
Artificial Intelligence
Agentic ML Exploration (A-MLE) for Ads Ranking
Modern industrial ads ranking stacks are increasingly bottlenecked not by model capacity or training compute, but by the throughput of human ML iteration - the cycles of research, implementation, training, debugging, evaluation, and launch required to surface a single statistically significant improvement. A typical ranking stack contains numerous differentiated models with heterogeneous data, architectures, and infrastructure constraints, and each cycle takes days to weeks of senior engineer attention per model. As a result, techniques that have proven effective on one model diffuse into others slowly and unevenly, leaving substantial recoverable signal unexplored. We present Agentic ML Exploration (A-MLE), an autonomous LLM-agent system that systematically explores ML techniques across a portfolio of ads ranking models. A-MLE decomposes ML iteration into five stages involving hypothesis generation, exploration strategy, experiment execution, result analysis and shared knowledge substrate which are orchestrated by a single agent that invokes domain-specific skills and agentic workflows against a sandboxed execution layer, with human-in-the-loop checkpoints at each stage boundary. We deploy A-MLE across a representative set of large-scale ads ranking models and evaluate it along a tiered capability framework (tool availability, autonomous workflow execution, and open-ended exploration). We further report a controlled cross-LLM study using a fixed agent loop, which surfaces qualitative differences in execution reliability and exploration aggressiveness across the Claude Sonnet, Gemini, and GPT families. We discuss failure modes and the design choices that govern reliability. Our findings suggest that agentic exploration is a practical force multiplier for ML engineers in industrial recommenders, especially for the long tail of models that rarely receive expert attention.
Read the full abstract
Artificial Intelligence
zScore-N: A Neural Network for On-Chain Wallet Reputation Scoring
Wallet reputation scores decide who receives an airdrop, who can borrow, and who enters an allowlist across decentralised finance. They almost always begin as hand-written formulas: compositions of clamped logarithmic, linear and square-root transforms over behavioural features, with every threshold and point award set by hand. Such a formula is readable and deterministic, but it is piecewise and non-differentiable, it cannot improve as data accumulates, and it cannot distinguish a feature that is genuinely zero from one its pipeline failed to capture. We present zScore-N, the neural network that replaced ours in production. The formula served as its teacher: calibrated against 5,208,952 wallets sampled across 2019-2024 and verified to reproduce production output to within 2.3e-13, it supplies unlimited labelled training data at zero label noise. The trained network reproduces it to 0.58 points RMSE on the 1000-point scale (R^2 = 0.99997), against 2.25 for gradient-boosted trees and 28.04 for linear regression on identical features and splits. Trained with missing-value masks against uncorrupted targets, it halves the error that incomplete data introduces: at 10% feature-level missingness the formula drifts 51.4 points from its own complete-data output with a systematic -12.5 point bias, while the network drifts 17.9. The network carries the score at production scale, across a population of millions of wallets spanning six orders of magnitude in size and activity.
Read the full abstract
High Energy Astrophysical Phenomena
Revisiting EBL Constraints from Gamma-Ray Observations: A Critical Assessment and Methodological Improvements
The extragalactic background light (EBL), ranging from the infrared to the ultraviolet bands, is the second most intense photon field in the universe, surpassed only by the cosmic microwave background (CMB). It is primarily generated by starlight in galaxies, either directly or through absorption by dust and re-emission at longer wavelengths. Very high energy (VHE, $E$ > $100\,$GeV) photons can be absorbed via $\gamma \gamma$ interactions with the EBL during their propagation across cosmological distances, providing an indirect method to probe the EBL by studying its impact on the energy spectra of distant VHE sources. This paper examines the robustness of EBL constraints derived from gamma-ray data, critically assessing the assumptions made in previous studies about the intrinsic source spectra, the uncertainties of the observations and the validity of the analysis tools. We find that earlier studies likely underestimated the uncertainties in the EBL intensity constraints, underscoring the need to account for systematic uncertainties comprehensively. By employing a Monte Carlo (MC) simulation and a plausible model for unknown systematic errors, we compute more realistic uncertainties. Additionally, we discuss possible alternatives to set EBL constraints, relaxing the assumptions on the intrinsic spectra of gamma-ray sources, with the goal of achieving more robust constraints.
Read the full abstract
Analysis of PDEs
On the Quadrature Surface Free Boundary Problem on Riemannian Manifolds: Sufficient Condition of Existence
We study the quadrature surface free boundary problem on a compact Riemannian manifold $(\M,g)$ of dimension $N \geq 2$. Given a positive source $f \in L^2_g(\M)$ with compact support $K$, and $k>0$, one seeks a domain $\Omega \supset K$ on which the overdetermined problem $-\Delta_g u = f$ in $\Omega$, $u=0$ and $|\nabla_g u|_g = k$ on $\partial\Omega$ admits a solution. Following the geometric and variational framework of Djité–Seck, we prove that the integral condition \[ (NS)_g \qquad \int_C f\,dv(g) > k\,|\partial C|_g, \] where $C$ is the totally convex hull of $K$, is sufficient for the existence of a solution strictly containing $C$. We then show, by an explicit counterexample on the round sphere $S^n_R$, that $(NS)_g$ is not necessary in general under nonnegative curvature: the phenomenon arises because the perimeter is not monotone with respect to inclusion in positive curvature. Finally, we identify an additional geometric hypothesis—perimeter monotonicity within a normal convex neighborhood of $C$—under which necessity is restored, and we discuss the critical equality case $\int_C f = k|\partial C|_g$.
Read the full abstract
Machine Learning
CUNO: Curriculum and Preference Optimization for Stable Graph Unlearning under Mass Deletion
Graph unlearning removes the influence of designated training data from a trained graph model without retraining from scratch. However, existing methods suffer a sharp drop in model utility under large deletion ratios (mass deletion), a phenomenon we refer to as catastrophic unlearning. We find that a key cause is the uniform treatment of all deleted samples, which is particularly damaging in graph learning: structural dependencies cause different nodes to play vastly different roles in the learned model, yet existing methods apply the same forgetting operation to the entire forget set. Based on this insight, we propose CUNO, a curriculum-based graph unlearning framework that removes the forget set progressively, ordering samples by their estimated unlearning difficulty across multiple stages. CUNO further employs a distribution-level negative preference optimization (NPO) objective at each curriculum stage that steers the model away from its original behavior on the current forget subset while preserving retained performance. Our theoretical analysis shows that the curriculum design is most beneficial when the forget set spans a wide range of unlearning difficulty, a condition naturally satisfied under mass deletion. Comprehensive experiments confirm that CUNO consistently mitigates catastrophic unlearning: at 20% deletion, it retains 74% of the original utility compared to 26-53% for existing methods, and maintains more than half the original utility even at 50% deletion. Our code is publicly available at https://anonymous.4open.science/r/cuno-D4FF.
Read the full abstract
Number Theory
Replication Descent and $J$-Finality of Replicable Functions
We show that replication carries congruence symmetry down an explicit level tower. If a normalized replicable function $f$, holomorphic on the upper half-plane, is invariant under $\Gamma_0(N)$, then its $n$th replicate is invariant under $\Gamma_0(N/(N,n))$. Thus every replicate whose index is divisible by $N$ is the normalized modular invariant $J=j-744$. This gives a direct and classification-free proof of $J$-finality for congruence-invariant replicable functions. The argument uses only the replication identities and an elementary generation theorem for congruence subgroups; it requires neither complete replicability nor arithmetic hypotheses on the Fourier coefficients. We then apply the descent law to completely replicable functions of finite replication order. Their replication towers have a canonical terminal replicate, and the only possible terminal functions are $J$, $q^{-1}$, and $q^{-1}+q$. Moreover, a finite-order completely replicable function is $J$-final precisely when its terminal replicate is $J$, and any symmetry beyond translations forces this alternative.
Read the full abstract
Computer Vision
CS-CLIP: Compositional Scene Graph-guided CLIP for Robust Compositional Reasoning
Vision-language models (VLMs) demonstrate strong performance across compositional reasoning benchmarks, which require reasoning over semantic perturbations of objects, attributes, relations, and their interactions. However, our controlled analysis reveals that existing compositionality-aware VLMs exhibit element-specific biases, often underperforming vanilla CLIP on certain compositional elements. To address this, we propose Compositional Scene Graph-guided CLIP (CS-CLIP), which uses scene graphs to identify compositional elements and construct structured negatives via selective masking. We further retain negatives that are most contradictory to the original caption, forcing the model to rely on compositional structure rather than surface cues. CS-CLIP achieves state-of-the-art compositional reasoning with robust performance across compositional elements. It also preserves general vision-language capabilities such as cross-modal retrieval and downstream visual reasoning, while requiring fewer training samples than prior methods.
Read the full abstract
Atmospheric and Oceanic Physics
EastAsiaClimateExtremes: An AI-Ready Dataset of Weekly Atmospheric and Oceanic Extremes over East Asia for Subseasonal Prediction Research
Despite growing interest in AI-based prediction of climate extremes, event- or label-based AI-ready extreme climate datasets remain limited, constraining efforts to systematically characterize and forecast such phenomena. To address this gap, we present EastAsiaClimateExtremes, an open dataset that provides ERA5/OISST reanalysis-based weekly extreme labels and event-based metrics for anomalously high temperature (AHT), heavy rainfall (HR), and marine heatwaves (MHW) over East Asia, together with analysis workflows hosted on GitHub to facilitate reproducibility and adaptation. The dataset is fully documented and co-registered with ECMWF S2S hindcast outputs on a common spatial grid and temporal framework, thereby enabling direct comparison between reanalysis-derived labels and dynamical model forecasts. This unified dataset serves as a reference framework for East Asian climate extreme research and AI-based subseasonal-to-seasonal prediction. It supports both quantitative characterization of the spatiotemporal occurrence of regional extremes and systematic diagnosis of S2S model skill in reproducing extreme signals. Beyond these immediate applications, the dataset enables a broader range of studies such as extreme event attribution and compound risk analysis. The accompanying analysis workflows characterize the historical statistics of reanalysis-based weekly extremes-including occurrence frequency and mean and maximum intensity-together with their climatological means and trend characteristics, and further assess the skill of the ECMWF hindcast.
Read the full abstract
Databases
SmartANN: Object Causal Modeling Boosts Approximate Nearest Neighbor Diagnosis and Auto-Design
Approximate Nearest Neighbor (ANN) algorithms achieve high efficiency through interdependent phases across index construction and query execution. This coupling allows upstream performance loss to propagate downstream, affecting execution behavior and measurable outputs. Existing component-level analyses mainly compare isolated design choices, while end-to-end benchmarks report aggregate metrics; neither traces loss propagation across dependent phases, hindering root-cause attribution and automated redesign. We present SmartANN, a framework based on the object causal model (OCM) for ANN bottleneck attribution and automated redesign. SmartANN represents an ANN workflow as eight ordered, replaceable objects and diagnoses them with a sequential diagnose-and-replace loop. At each iteration, it identifies the first object deviating from expected behavior or output as a bottleneck. Because an upstream bottleneck can obscure downstream ones, SmartANN replaces it with a test oracle when available, or with an implementation producing a better outcome, then continues downstream diagnosis. From the diagnosed bottlenecks and failure causes, SmartANN composes compatible actions from a pluggable action library to generate an optimized end-to-end ANN design. We instantiate SmartANN for IVF-PQ and HNSW, covering partition-and-quantization and graph-based ANN families. Experiments on eight real-world datasets show that SmartANN improves Recall by 0.24–74.20%, and increases QPS by 28.8–256.5% at comparable Recall, with low diagnosis and auto-design overhead. The code is available at https://github.com/zhouyutong20/SmartANN.
Read the full abstract
Computational Physics
Beyond Residuals: Energy based solutions of partial differential equations using scientific machine learning
Energy-based approaches provide a natural and physically consistent framework for a large class of partial differential equations arising in solid and fluid mechanics, where the governing equations follow from variational principles. In contrast to residual-based physics-informed neural networks (PINNs) and their weak-form variants, which enforce the strong or weak form of the equations through loss minimization, the Deep Energy Method (DEM) directly computes the solution as the minimizer of an energy or incremental potential functional. This eliminates the need for residual weighting, avoids high-order derivatives, and enables the direct enforcement of physical constraints through the variational structure.In this work, we systematically revisit the Deep Energy Method, placing it in the broader context of physics-informed learning and variational modeling. We clarify the relationship between DEM, PINNs, and VPINNs, and identify the class of problems for which energy minimization provides intrinsic advantages in terms of stability, robustness and interpretability. Particular emphasis is placed on incremental variational formulations, which allow DEM to be applied to nonlinear, history-dependent and time-dependent problems, including phase-field fracture and dissipative systems. The variational structure underlying DEM further provides a natural foundation for optimization and inverse problems, where the energy functional acts as a physics-based constraint rather than a residual penalty. Through a series of numerical examples, we demonstrate that DEM offers a principled and effective alternative to residual-based methods for variational problems, highlighting its strengths and limitations relative to existing physics-informed approaches.
Read the full abstract
Differential Geometry
Orthogonal complex structures on flat tori
Superconductivity
Magnetochiral anisotropy of thermally activated phase slips in noncentrosymmetric superconductors
Resistive tails of narrow superconducting wires below the critical temperature are governed by phase slips, thermally activated near the transition and quantum at lower temperatures. We develop the theory of thermally activated phase slips in wires lacking inversion symmetry, modeled microscopically as a diffusive wire formed from a two-dimensional electron gas with Rashba spin-orbit coupling in an in-plane magnetic field. The phase-slip resistance acquires magnetochiral anisotropy: at fixed current it changes upon reversal of the field, and at fixed field upon reversal of the current, while the linear-response resistance remains field symmetric as required by Onsager reciprocity. The nonreciprocity originates from an asymmetry of the phase-slip activation barrier that is odd in both current and field. It is generated by the Lifshitz invariants of the free energy, the cubic gradient term and the momentum-odd quartic vertex, together with even gradient terms promoted to odd ones by the helical ground state, precisely the combination that determines the superconducting diode effect. The barrier asymmetry is therefore fixed by the diode efficiency of the same wire through a universal numerical ratio, with no adjustable parameters. The kinetic Lifshitz invariant, which makes relaxation and noise of the order parameter nonreciprocal, affects only the fluctuation prefactor at a parametrically subleading level. Because the asymmetry resides in the activation exponent, the magnetochiral signal is amplified relative to the diode asymmetry by the ratio of barrier height to temperature, and appears as a field-antisymmetric splitting of the exponential slopes of differential-resistance traces routinely used to identify thermally activated phase slips.
Read the full abstract
Artificial Intelligence
Style Over Substance: Content-Invariant Wrappers Flip LLM Safety-Judge Verdicts
Automatic safety judges – for example Llama Guard, or GPT-4o given a grading prompt – produce the numbers behind reported jailbreak success rates, defense evaluations, and safety leaderboards. We ask whether these judges grade what a reply contains or how it sounds. We place a fixed piece of text before or after a model reply (a style wrapper) and leave the reply itself unchanged byte-for-byte, so any change in verdict is an error of the judge. Over 600 JailbreakBench replies (300 harmful, 300 refusals), 9 wrappers, and 8 judges, using paired tests and a measured per-judge repeat-disagreement rate (the noise floor), we find that most judges barely move but specific judges have specific, cheap blind spots. A token refusal ("I can't help with that") placed before the unchanged harmful text flips 19.9% of GPT-4o-mini's correct "unsafe" verdicts to "safe" (noise floor 0.5%; 18.2% under majority-of-three re-scoring), yet moves Claude only 0.4%. The deployed Llama Guard 4 guard model is evaded by an "educational course" preface on 12.3% of harmful replies, and the flips persist across repeats, while a second deployed guard (gpt-oss-safeguard-20b) is immune. Changing only the grading prompt to ignore tone cuts the GPT-4o-mini attack roughly tenfold on the same model, so the weakness lives in the judge, not in the content. Pure tone-only decoration that adds no new claim still flips up to 2.1% of verdicts on the vulnerable judges, several times their noise floors. On this artifact set the judges also disagree about which target model is safest, and a bootstrap shows the ranking is unstable before any wrapper is added. We release the data, wrappers, code, and per-verdict labels.
Read the full abstract
Soft Condensed Matter
A unified description of flow-induced scission of wormlike micelles under shear and extensional flows
We investigate flow-induced scission of wormlike micelles under different flow kinematics using dissipative particle dynamics simulations of surfactant solutions in shear, uniaxial extensional, planar extensional, and biaxial extensional flows. The average lifetime of wormlike micelles is used to quantify the degree of scission. While flow-induced scission occurs in all flow types when the deformation rate is sufficiently high, the dependence of the average lifetime on the deformation rate differs among flow types. To provide a unified description, we introduce an effective extension rate determined by the velocity gradient tensor and micellar orientation. When expressed in terms of this effective extension rate, the average lifetimes obtained under all flow types considered collapse onto a single curve. These results demonstrate that a unified description of flow-induced scission requires not only the strength and kinematics of the imposed flow, but also the micellar orientation relative to the extensional direction.
Read the full abstract
Machine Learning (Statistics)
Distribution-free inference on the number of changepoints
Suppose we are given an ordered sequence of independent data whose distribution changes $K$ times at unknown locations, for some unknown $K \geq 0$. In this paper, we study the problem of performing distribution-free inference on $K$. First, we show an impossibility result: any distribution-free upper confidence bound on $K$ must be trivial and uninformative. Then, using conformal $p$-values, and under only the assumption that the data segments induced by the changepoints are exchangeable (within themselves) and mutually independent, we construct a finite-sample valid lower confidence bound on $K$, which we call the Conformal LOwer bound on Changepoint Count (CLOCC). We show that CLOCC is the only feasible way to provide a lower bound on $K$ under the stated assumptions, a property we refer to as its universality. We provide practical guidelines for choosing score functions that yield efficient and tight lower bounds. We evaluate CLOCC in several synthetic and real-data experiments, where it provides informative lower bounds on $K$, demonstrating its practical applicability.
Read the full abstract
Quantum Physics
Transversal Fanout for Fault Tolerant Distributed Quantum Computing: Analysis and Application
We study a resource-efficient approach for implementing logical fanout operations in fault-tolerant distributed quantum computing using transversal operations on quantum error-correcting code blocks. Logical fanout, comprising multiple controlled-NOT operations from a common control qubit to target qubits located at remote nodes, is an important primitive for distributed quantum computation but can require substantial non-local communication when implemented directly between encoded blocks. We exploit the structure of encoded blocks and the availability of transversal logical operations to construct distributed fanout circuits that reduce the required non-local operations while preserving the logical action of the fanout operation. The construction is developed for encoded quantum information and illustrated using Bivariate Bicycle (BB)-code blocks. We analyze the resulting physical gate, entanglement, circuit-depth, and ancilla requirements. The approach provides a systematic method for implementing large logical fanout operations across distributed error-corrected quantum processors. Also, we study a distributed implementation of the global gate GCZ involving logical qubits (encoded using BB-code blocks), exploiting the concurrency in transversal distributed fanouts.
Read the full abstract
Hardware Architecture
Routing Dense Layouts with History-Aware Offline Reinforcement Learning using LSTM
Detailed routing remains a dominant runtime bottleneck in physical design due to increasing complexity of design rules. Modern routers can struggle to resolve persistent violations under dense operating conditions. While recent work leverages reinforcement learning (RL) to dynamically select costs for each routing iteration, we find that this technique struggles with high-density designs where routing solutions are significantly harder. To address this, we present a history-aware offline RL policy which predicts iterative cost weights in these dense regimes to improve convergence across placement densities by utilizing readily available features from the router. Our policy uses conservative Q-learning similarly to prior work; however, our key insight is that addition of a lightweight LSTM architecture and additional features can retain sequence context and improve routing convergence across multiple densities and route guide qualities. Our policy can be integrated into any cost-based router with minimal pipeline changes, as it does not interfere with the core search algorithm. We evaluate our policy on held-out density and adjustment settings, including difficult operating points induced by dense placement and low guide quality. Our policy reduces design rule violations (DRVs) by an average of 92% over the top public baseline while simultaneously reducing runtime by 10%.
Read the full abstract
Distributed Computing
SemBridge: Compiling Consumer Observations into Cross-Stack Communication Plans
Distributed-tensor systems specify where values reside, while collective systems optimize how requested operations execute. At a boundary between vendor runtimes that cannot share a native communicator, neither abstraction states what a remote consumer must observe. SemBridge fills this gap by compiling graph and runtime facts into a typed contract for the consumer-visible result and its delivery obligations. The contract captures provenance, substitutability, completion, authority, demand, and native-domain locality. A deterministic lowerer constructs backend-neutral communication plans, and a symbolic checker validates each plan before execution across CUDA/NCCL and CANN/HCCL. An independent layout-only planner handles all 72 structural transitions but establishes only 54 complete obligations; a byte-only minimizer proposes 40 semantically invalid candidates, all rejected by SemBridge. On nine real edges, SemBridge produces distinct observation-aware plans that reduce startups on all nine and payload bytes on the three result edges. A live CUDA/CANN run derives and executes full-logit reconstruction, source projection, and owner-token delivery from log-probability, token-only, and owner-scoped requests. On a measured two-host 1-GbE capacity-spillover deployment, source projection cuts result traffic by more than 99.97% and increases throughput by 8.92-80.20% across Dense, MoE, and MiniMax workloads. All 18 MiniMax restart pairs at concurrency 1, 8, and 16 favor source projection. A Qwen3-14B MLP slice additionally verifies bitwise activation-shard delivery and HCCL completion of row-parallel partials. These results establish consumer observation as a semantic layer between placement and collective execution.