Symplectic Geometry
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Wednesday 9 September
Let $\Omega$ be the Lorentz future cone in $\mathbb{R}^{d+1}$ with respect to the Lorentz product and let $T^M$ be the $M$-fold product of the future tube $T=\mathbb{R}^{d+1}+i\Omega$. The Lorentz group $\mathrm{SO}_0(1,d)$ acts diagonally on $T^M$, and its complexification $\mathrm{SO}(1,d)^\mathbb{C}$ acts on $\mathbb{C}^{(d+1)\times M}$. We prove that the domain $G^\mathbb{C}\cdot T^M$ is a Stein manifold for any connected unipotent subgroup $G$ of $\mathrm{SO}_0(1,d)$.
High Energy Astrophysical Phenomena
Anisotropic wind in tidal disruption events
Over the coming years, the number of tidal disruption events (TDEs) is expected to substantially increase with observations from the Vera Rubin Observatory (g and r band) and ULTRASAT (near UV) wide-field surveys. These future samples have great promise to characterize the bottom end of the massive black hole mass function, but existing detections of intermediate mass black hole TDEs are primarily in X-rays, leaving their optical/UV emission largely unexplored. We present a time- and angle-dependent analysis of the outflow produced by dissipation near pericentre in a three-dimensional end-to-end radiation-hydrodynamics simulation of a TDE by a $10^4 M_\odot$ black hole with realistic parameters, run with the code RICH. We find that outflow anisotropy produces viewing-angle-dependent observables. Towards the poles and the pericentre region, mass-loss rates are low and bolometric luminosities reach $\sim2$–$3$ times the Eddington luminosity. Towards the stream, the properties show a stronger dependence on latitude: the mass-loss rate increases and the bolometric luminosity decreases as the line of sight approaches the orbital plane. These denser regions favour H$\alpha$ and H$\beta$ emission. Despite these variations, all viewing directions show a common spectral evolution, with an initial soft X-ray flare followed (around $1.25t_{\rm fb}\approx3$~days) by the reprocessing of shock-powered emission into the UV and optical bands. Although the optical/UV luminosities we predict for this TDE are likely too dim for past surveys (e.g. ASAS-SN, ZTF), they are within the detection capabilities of LSST and ULTRASAT to horizons of $\sim 790$ and $\sim 340$ Mpc, respectively, for the brightest viewing directions.
Read the full abstract
Superconductivity
Linear and nonlinear Edelstein effects in Rashba superconductors
We formulate a quasiclassical theory of the Edelstein effect in superconductors that incorporates both intraband and interband contributions. To describe the interband contribution, which is absent from the conventional leading-order quasiclassical formulation, we derive augmented Eilenberger equations in the presence of antisymmetric spin-orbit coupling. The intraband contribution is evaluated using multiband Eilenberger equations. We apply these formulations to supercurrent-induced surface spin magnetization in $s$-wave Rashba superconductors and investigate its dependence on temperature, distance from the surface, spin-orbit coupling strength, and supercurrent. The intraband contribution originates from a supercurrent-induced asymmetry of quasiparticles with opposite momenta and spin polarizations, whereas the interband contribution arises from the anomalous-velocity term generated by the momentum derivative of the Rashba spin-orbit potential. In the helicity basis, this anomalous-velocity term is expressed in terms of the Berry connection associated with the momentum dependence of the Rashba eigenstates. The intraband contribution increases linearly with the spin-orbit coupling strength, whereas the interband contribution exhibits a nonmonotonic dependence and is maximized when the Rashba spin splitting is comparable to the superconducting gap. Moreover, within the clean $s$-wave Rashba model considered here, we find that the nonlinear dependence of magnetization on the supercurrent arises solely from the interband contribution. Thus, although the intraband contribution dominates the linear Edelstein effect, the nonlinear Edelstein effect can serve as a useful probe of the interband contribution originating from quantum geometry.
Read the full abstract
Quantum Physics
Finite-Particle Quantum Reduction of Thermodynamic Irreversibility
Thermodynamics is useful because it lets us predict and control complex systems from a few accessible quantities when microscopic reconstruction is impractical. We ask whether this operational reduction can be less costly in a finite quantum system than in a matched classical one. We compare a few-boson Bose–Hubbard chain with its number-conserving classical-field counterpart under the same driving and the same coarse spatial record. The quantum system relaxes more in the observed particle distribution, yet the retained mean energy resolves more of the microscopic structure hidden by that record, leaving less information unusable and smaller entropy generation. We then close the process into a heat-engine cycle whose controls use only the measured mean energy and spatial record, not full state tomography; at matched thermal resources, the smaller entropy generation gives more work and higher efficiency. The advantage fades as the particle number grows toward the classical-field regime.
Read the full abstract
Computation and Language
Through the Looking Glass: Directly Reading and Writing Transformers
How many of a transformer's components decide a token? Counted by the absolute value of each unit's and channel's contribution to the logit, one prediction rests on thousands to hundreds of thousands of them. But contributions are signed, and across eighteen models the mass pushing away from the predicted token is a median of seven times the mass carrying it. Divide by the net and the count is dozens: on the baseline, 53 components carry ninety percent of a prediction, 13 it cannot survive losing, and 8 suffice to produce it alone. Across twelve models trained elsewhere, 124M to 7B parameters, the sufficient set runs from two components to sixteen, and what a prediction draws on, followed all the way back, is one to three percent of the model, a share that does not grow with size. Three quarters of a layer's update is a fixed linear map of the state it received. Everything is read from the model's own parameters and activations, with nothing trained or fitted, and it names a component on both sides: what it writes, from the predictions it drives, reaching close to half of every model; what it reads, from its weights in the frame of its own layer, at 58.9 percent above chance over its eight strongest inputs. Sorting the remainder by upstream source yields grammatical categories the embedding cannot see. A name can be acted on. An association the model does not hold installs into one spare unit, key and value read from the weights, for a quarter of a percent of held-out loss, a fortieth of what a rank-one update costs. An installed attention head and a unit two layers above it make an edit fire only where a token occurred earlier in the context, and a unit the model trained for itself is driven from two layers upstream, 86 percent of the effect passing through it. An order-preserving activation puts a unit's inputs at the instrument's ceiling, at the price of a two-part install.
Read the full abstract
Classical Analysis and ODEs
A proof of Askey's convexity conjecture
For $-1<\alpha\leq1/2$, let $J_\alpha$ be the Bessel function of the first kind and let $j_{\alpha,2}$ be its second positive zero. Define $\beta(\alpha)<\alpha+1$ by $$ \int_0^{j_{\alpha,2}}u^{-\beta(\alpha)}J_\alpha(u)\,du=0. $$ We prove that $\beta''(\alpha)>0$ for $-1<\alpha\leq1/2$, including the left second derivative at $\alpha=1/2$. The continuous extension $\beta(-1)=0$ is strictly convex on $[-1,1/2]$, strengthening Askey's 1993 convexity conjecture. The analytic argument uses a positive expansion of a primitive and a vanishing weighted sum. Increasing ratios of consecutive weights give a negative covariance term, while one comparison controls slope and curvature. The remaining algebraic step proves eight rational inequalities by finite exact polynomial calculations, reproduced in an appendix.
Read the full abstract
Differential Geometry
The measure contraction property on Grushin spaces
We determine the sharp measure contraction exponents of two families of Grushin-type metric measure spaces. The radial Grushin space $\mathbb{G}^{n+m}$ is $\mathbb{R}^{n}\times\mathbb{R}^{m}$, equipped with Lebesgue measure and generated by $X_i=\partial_{x_i}$ and $Y_j=|x|\partial_{y_j}$, for $1\leq i\leq n$ and $1\leq j\leq m$. We prove that $\mathbb{G}^{n+m}$ satisfies $\operatorname{MCP}(K,N)$ if and only if $N\geq n+4m$ and $K\leq 0$. We also show that, for $\alpha\geq1$, the $\alpha$-Grushin plane generated by $X=\partial_x$ and $Y_\alpha=|x|^\alpha\partial_y$ satisfies $\operatorname{MCP}(K,N)$ if and only if $K\leq0$ and $N\geq N_\alpha$, where \[ N_\alpha := 1+\max_{L>1} \frac{(2\alpha+1)L}{(L-1)^{2\alpha+1}+1}. \] This resolves the conjecture posed in arXiv:2010.16350 and, for integer $\alpha\geq2$, provides the first examples of real-analytic sub-Riemannian structures with noninteger curvature exponent. Both results recover the known curvature exponent $5$ of the classical Grushin plane, corresponding respectively to $n=m=1$ and $\alpha=1$.
Read the full abstract
Signal Processing
Covariance-Aware MM-PGD for Mixed Near-/Far-Field Activity Detection
Grant-free activity detection with mixed near-field (NF) and far-field (FF) devices is an important problem that can be addressed via covariance-based detectors. The difficulty is that NF users induce device-specific structured spatial covariances, whereas FF users are well approximated by isotropic covariances. Under a unified Rician model, we first formulate activity detection as a relaxed maximum-likelihood problem in the full LM-dimensional vectorized observation space. We then develop a covariance-aware majorization-minimization projected gradient descent (MM-PGD) detector. It updates the full activity vector jointly and avoids the per-coordinate high-rank subproblems that arise in the NF regime. Numerical results over SNR, antenna-count, and NF-ratio sweeps show that MM-PGD achieves up to 20x lower miss-detection probability than the strongest coordinate-wise baseline. The advantage is most pronounced at high NF ratios, while in the all-FF case MM-PGD performs on par with the strongest coordinate-wise baseline.
Read the full abstract
Dynamical Systems
On the Absence of Anosov Factors for DA Local Diffeomorphisms
We give a class of local diffeomorphisms which are homotopic to toral hyperbolic endomorphisms, but which are not topologically semi-conjugate, within the homotopic class of the identity, to any Anosov local diffeomorphism. In particular, we show that for a non-invertible partially hyperbolic $C^1$-smooth local diffeomorphism with expanding directions on the $2$-torus, if it is semi-conjugate to an Anosov local diffeomorphism, then it is also an Anosov local diffeomorphism.
Read the full abstract
Solar and Stellar Astrophysics
MHD Modelling of magnetic reconnection heating and jets in the solar corona
Magnetic reconnection is seen as the lead mechanism to explain the persistent and bursty heating of the solar corona that keeps its thermal structure. Lately the observations and modelling of magnetic reconnection outflows, i.e. nanojets, has empowered our capabilities to scrutinise this mechanisms, however more needs to be understood on the relation between the energetic and kinematic events. The aim of this paper is to study in details the properties of the heating events (nanoflare) and the kinematic counterparts (nanojets) in order to augment detection capabilities of magnetic reconnection events. We use MHD simulations of the interaction of two magnetic flux tubes where magnetic reconnection occurs and such reconnection outflows are generated. We vary key parameters, such as the critical current and the size of the domain to study the heating and slingshot mechanisms. We find that the properties of the nanojets can give away much information on the magnetic reconnection that generated it, while it is difficult to directly link the amount of heating and the jet speeds. We also find that such reconnection model can explain a wide range of heating events. Magnetic reconnection still is invariably linked to both heating and outflows and our simulations show the complex interplay between these two byproducts of the magnetic reconnection and how their properties can be studied in observations.
Read the full abstract
Chemical Physics
Direct separation of intra- and inter-molecular contributions in pulse dipolar EPR experiments
We present a direct method of separating intra- (form factor) and inter-molecular (background) contri- butions in pulse EPR dipolar spectroscopy (PDS) experiments. The form factor/background separation is accomplished by rearranging data measured at two different biradical concentrations, without any fitting fundamentally required. This new data processing method allows to analyse PDS data from spin concen- trated samples for which background separation was previously unrealistic. In particular, the presented method would be of high importance for studies of spin labeled biomolecules, especially in respect to the liquid-liquid phase separation (LLPS). Furthermore, we propose a simple measurement and data processing protocol to separate the PDS contributions from each phase in LLPS samples. The protocol for the form factor/background separation for RIDME experiment is also discussed
Read the full abstract
Probability
Moderate Deviations for Nonlinear Hawkes Processes
A Hawkes process is a simple point process whose intensity depends on its history; the resulting dynamics are generally non-Markovian. We establish a sample-path moderate deviation principle for a nonlinear Hawkes process in the full moderate regime. Since a Poisson cluster representation is unavailable for nonlinear Hawkes processes, we use the past configuration as a Markov state and construct a potential, or Poisson corrector, for the centered stochastic intensity. A monotone Poisson coupling shows that the add-one increment of the corrector is uniformly bounded. The centered counting process is consequently the sum of a martingale with bounded jumps and an exponentially negligible boundary term. Exponential stabilization of the predictable quadratic variation follows from the process-level large deviation principle for nonlinear Hawkes processes. The martingale moderate deviation theorem then yields the result for every scale between the central-limit and large-deviation scales. The same construction gives a response formula for the asymptotic variance and, in particular, verifies that the variance dominates the stationary mean intensity in the self-exciting case.
Read the full abstract
Systems and Control
IMU-Centric Moving Horizon Estimation for Lateral Dynamics Estimation Across Vehicles and Grip Conditions
Accurate estimation of lateral vehicle dynamics near the adhesion limit is important for stability control and high-performance driving, but lateral velocity is rarely measured directly because sensors such as optical sensors are costly. This paper presents an inertial measurement unit (IMU)-centric Moving Horizon Estimation framework that reconstructs lateral velocity using standard onboard signals, without relying on exteroceptive odometry or detailed tire-parameter tuning. Experimental validation on human-driven sports cars and an autonomous open-wheel race car across tracks, maneuvers, and conditions demonstrates accurate and robust lateral velocity and lateral acceleration estimates. The proposed framework is available at https://github.com/Aseuffo/IMU-Centric-MHE
Read the full abstract
Optimization and Control
Distributed Vector Assembly via Selector-Anchored Diffusion with Local Innovation
We introduce Selector-Anchored Diffusion with Local Innovation (D+LI), a distributed vector-assembly primitive that constructs complete local replicas from independently owned authoritative blocks. Orthogonal selectors preserve block identity, local innovation anchors each owner's contribution, and Laplacian diffusion propagates these contributions across the network. We show that pure averaging with raw additive selector sources cannot isolate this replica. The D+LI error operator decomposes coordinatewise as the identity minus an owner-grounded Laplacian, yielding global contraction with a factor independent of the total and block dimensions. For time-varying outer targets, we establish a finite-T input-to-state stability (ISS) bound and a self-consistent estimator-controller gain condition for Lipschitz coordination maps, including non-potential ones. For semi-algebraic potential problems, we prove Kurdyka-Lojasiewicz (KL) finite-length convergence of the coupled decision-assembly iteration with a fixed finite communication budget per outer iteration under explicit timescale-separation conditions. Three numerical experiments demonstrate contraction tightness and finite-horizon tracking, verify an admissible KL timescale, and illustrate discrete and non-potential outer maps.
Read the full abstract
Machine Learning
Robust Beam Prediction for V2X Networks with Multi-Modal Sensing
Integrated sensing and communication (ISAC) provides a promising foundation for beam prediction in future vehicle-to-everything (V2X) networks. However, existing sensing-assisted beamforming methods still rely heavily on radio-frequency sensing, which may become unreliable in complex vehicular environments. Meanwhile, the growing availability of heterogeneous sensors, such as cameras and LiDAR, offers new opportunities to improve beam prediction through richer environmental perception. Motivated by this, this paper proposes a multi-modal beam prediction framework for V2X networks. Specifically, we develop BeamTransFuser, a hierarchical Transformer-based architecture that progressively fuses camera, LiDAR, radar, and GPS observations for robust beam prediction. In addition, to handle possible missing modalities in practical deployment, we introduce a generative module that reconstructs missing modality features from the available observations. Experimental results on a real-world multi-modal V2X dataset show that the proposed framework consistently outperforms representative baselines, while the generative module further improves robustness under incomplete sensing conditions.
Read the full abstract
Human-Computer Interaction
Seeing the Voice, Preserving the Self: A Participatory Design Approach to Deaf-Centric Text-to-Speech
We describe a participatory design approach toward developing Deaf-centric text-to-speech (TTS) technologies. While TTS is growing rapidly in the mainstream, it has received little attention to date in the deaf and hard of hearing (DHH) technology space. Critical problems have remained unaddressed for DHH users, including the ability to manipulate tone, emotions and delivery via non-auditory means. Verifying that the generated speech matches intent and is appropriate for a given situation without having to listen to it is another challenge. Respecting cultural and identity factors in the generated speech is also important. This work explores the design space with DHH participants through two focus groups, three co-design sessions, and four one-on-one early-stage design evaluation sessions. Participants included people both familiar and unfamiliar with TTS, as well as DHH content creators. We describe key findings, design ideas, results, and implications for future Deaf-centric TTS development. We also identify unmet technology requirements that pose barriers to adoption of Deaf-centric TTS technology.
Read the full abstract
Computation and Language
Politics of Feelings: Emotional Expression and Legislative Effectiveness in the U.S. Congress
Emotions are a pervasive feature of political communication, yet existing research has focused primarily on describing patterns of emotional expression rather than examining whether they are associated with consequential legislative outcomes. We address this gap by investigating the expression and correlates of discrete emotions in more than 1.7 million speeches delivered in the U.S. Congress between 1973 and 2024. Using a transformer-based emotion classifier, we measure eight discrete emotions: anger, fear, disgust, sadness, joy, enthusiasm, pride, and hope. We examine how these emotions vary over time, across policy topics, legislator characteristics, and their relationship with legislative effectiveness. We find that congressional speeches are becoming emotionally expressive over time. Emotional expression also varies systematically across policy domains and ideological positioning of legislators. Notably, the relationship between emotional expression and legislative effectiveness depends on the specific emotions expressed: enthusiasm and pride are positively associated with effectiveness, whereas anger exhibits a negative association. Emotional valence and emotional diversity are positively associated with legislative effectiveness, while emotional intensity is negatively associated with legislative effectiveness. These findings demonstrate that computationally derived measures of discrete emotions can provide insight into affective dimensions of legislative speeches and facilitate our understanding of how legislators communicate, interact, and perform within democratic institutions.
Read the full abstract
Differential Geometry
Immortal Curve Shortening Flow on Analytic Surfaces
We show that on any closed orientable real-analytic Riemannian surface, any smoothly immersed closed immortal curve-shortening flow converges to a unique non-constant closed geodesic.
Machine Learning
An Exponential Deterministic–Randomized Gap in ERM-Oracle Complexity for Thresholds on an Unknown Order
Attias, Hanneke and Ramaswami (NeurIPS 2025) asked whether randomization provably reduces the oracle calls needed for online learning when the class is accessible only through an oracle. We study the instance they singled out: transductive online learning of thresholds on an unknown total order of T instances, with a consistency-type ERM oracle that returns a full concept consistent with a queried labeled set (or reports non-realizability). Our main result is a separation for a fixed natural oracle. When the oracle is the minimal-prefix rule (or the maximal-prefix rule), every deterministic learner makes M mistakes and Q calls with $M+Q\ge T-\varepsilon$ on some instance ($\varepsilon\in\{0,1\}$, according to whether the empty prefix is a concept), and the constant is exact; hence $O(\log T)$ mistakes cost $T-\varepsilon-O(\log T)$ calls, whereas that paper's randomized learner achieves $O(\log T)$ expected calls and mistakes under the same rule. The randomized order is optimal: on an explicit hard distribution under the minimal-prefix rule, every learner has expected mistakes at least $((T+1-\varepsilon)\,128^{-\mathbb{E}[Q]}-1)/2$, so $\Omega(\log T)$ expected calls are necessary for polylogarithmic mistakes. The separation is governed by the oracle's selection rule, not by the class alone: for a legal feasible-median ERM rule a deterministic learner achieves $O(\log T)$ calls and mistakes, while a global-median rule again forces linear total cost. The same linear bound holds when the oracle's answers are chosen adversarially and then frozen into a memoryless oracle. We add partial tradeoff results for fixed query budgets (the middle regime is open) and an interface contrast: with only a weak consistency oracle, returning a realizability bit, both deterministic and randomized learners need $\Theta(T)$ calls.
Read the full abstract
Analysis of PDEs
A geometric approach to nonlocal 2-Hessian equations
We study a nonlocal 2-Hessian equation, given by an infimum of linear deformations of the fractional Laplacian, that is, $\inf_{A\in {A}_2}\Delta^s (u\circ A)(A^{-1}x)$. We characterize the class of coefficient matrices ${A}_2$, which determines the behavior of the operator, and provide a detailed geometric description of the possible degeneracies. Our main theorem shows that the nonlocal 2-Hessian equation remains uniformly elliptic for strictly positive right-hand sides, which leads to regularity estimates. The results hold under weaker hypotheses than those previously considered in the literature, and for the full range $s\in(0,1)$. In particular, no convexity on the solutions is required. Our hypotheses can be interpreted as nonlocal counterparts of the local notions of semiconcavity and 2-convexity. Moreover, all the results are stable as $s\to1$, recovering the local case. The geometric methods developed here are new, even in the local setting, and may be relevant to a broader class of nonlocal fully nonlinear equations and curvature-type problems.
Read the full abstract
Materials Science
Beyond Impedance Additivity: A Systematic Nonlinear Perspective on Memristor Associations
We investigate the validity of the superposition principle and impedance additivity in AC circuits containing a memristive device connected in series with a resistor, capacitor, or inductor. While the series association of impedances is a cornerstone of linear circuit theory, its applicability to memory-bearing nonlinear systems remains largely unexplored. Using a state-dependent memristive model, we numerically analyze the stationary current response under sinusoidal excitation and characterize the resulting harmonic spectra, Bode diagrams, and Nyquist plots. To assess whether the fundamental response can still be interpreted through an equivalent-circuit framework, we introduce the concept of an apparent memristor, whose effective parameters are extracted directly from the composite impedance. We show that, although the fundamental harmonic can be accurately reproduced by an apparent equivalent circuit over selected parameter ranges, the effective parameters differ substantially from those of the isolated memristor, revealing a renormalization induced by the coupling to the passive element. More importantly, we identify parameter regimes in which the apparent-circuit description breaks down altogether, particularly for capacitor and inductor-coupled systems, demonstrating that the composite impedance cannot generally be expressed as the sum of independent impedances. These results establish boundaries for the use of equivalent-circuit models in memory-enabled electronic systems and provide practical guidelines for the interpretation of impedance spectroscopy in nonlinear devices exhibiting memory.
Read the full abstract
Molecular Networks
Are You Learning Biological Signal or Shortcuts? Auditing and Mitigating Bias in Protein-Protein Interaction Datasets
Protein-protein interaction (PPI) databases do not faithfully reflect biological realities. Instead, they are influenced by study and technical biases that distort certain protein and interaction attributes. Machine learning models can exploit these as learning shortcuts if the negative dataset is not constructed with care. So far, the shortcuts introduced during PPI dataset construction have only been examined in isolation. Here, we systematically characterize both reported and, to our knowledge, previously unreported biases in PPI datasets that lead machine learning models to learn shortcuts instead of biological signal. We analyze HIPPIE, IntAct, and STRING, dedicated PPI databases, as well as two datasets derived from 3D-structural information in the Protein Data Bank (PDB). We show that random data splitting introduces strong topological shortcuts. When train-test protein overlap is removed, the resulting datasets still retain usable shortcuts stemming from self-interactions, taxonomic identity, and functional relatedness, whose prevalence interestingly depends on the data source. We further show that sampling negatives from a set of high-confidence non-interactors, an intuitively appealing choice, can amplify the shortcut stemming from functional relatedness. To detect and mitigate these biases, we provide an open Nextflow pipeline that combines similarity-aware, data-loss-minimizing dataset splitting with bias-minimizing negative sampling, both formulated as integer linear programs. Its key concept of quantifying biases to minimize them through optimization-based negative sampling can, in principle, be extended to any machine learning problem where the pool of negative candidates is much larger than the positives and is thus of interest also beyond PPI prediction.
Read the full abstract
Computation and Language
Who Argues What? Joint Argument-Entity Detection and Classification in Political Debates
Political debates are often analyzed through Argument Mining (AM) to investigate the key arguments that drive them. However, political arguments are rarely interpretable from argumentative spans alone, as claims and premises generally depend on the entities (e.g., people, events, locations, parties) they mention. Existing AM resources and methods typically annotate argumentative spans and roles, but do not provide a paired debate-entity layer for asking which Debate Named Entities (DNE), e.g., actors and events, are invoked within debates. In this work, we address these data and methodological gaps by (i) introducing DNE-ElecDeb, an entity-enriched version of the USElecDeb dataset that adds DNEs in both argumentative and non-argumentative spans and defines Debate Named Entity Recognition (DNER) as the task of detecting DNEs, and (ii) proposing Joint Argument and Entity Tagging (JAET), a generative framework that fine-tunes decoder-only LLMs to insert inline argument and entity tags into debate turns while preserving the original transcript. Under BIO-tagging evaluation, JAET improves relative F1 on the joint AM+DNER task by +27.3%, resp. +41.9%, under the untyped, resp. typed setting over the strongest sequential AM-DNER pipelines, demonstrating that such gains cannot be recovered by composing two independent modules. Notably, similar margins replicate on Persuasive Essays (+26.6%, resp. +52.7%), showing effective generalization to domains orthogonal to political debates. By unifying argumentative and entity-level representations within a single view, our contributions pave the way for richer political debates understanding.
Read the full abstract
Complex Variables
Stein neighborhood bases for complex–totally real unions
We study Stein neighborhood bases for unions of closed complex and totally real submanifolds, with particular attention to a linear configuration with noncompact intersection. We prove that the set $X=\{w=0\}\cup\{\Im z=\Im w=0\}\subset\C^2$ admits a Stein neighborhood basis. The proof starts from the explicit plurisubharmonic defining function $\rho(z,w)=(\Im w)^2+\sinh^2(\Im z)|w|^2$ and obtains the necessary non-uniform control at infinity by adding a locally uniformly convergent series of nonnegative plurisubharmonic terms constructed by Oka–Weil approximation. We also give an example of a closed complex submanifold and a closed totally real submanifold in $\C^2$ with compact intersection whose union has no Stein neighborhood basis, showing that the unrestricted union problem has a negative answer.
Read the full abstract
Materials Science
pyeCE: A Python Implementation of the Embedded Cluster Expansion
The cluster expansion is a widely used approach for predicting the finite-temperature thermodynamics of alloys from zero-kelvin first-principles calculations, but its conventional formulation becomes intractable for materials with more than three or four chemical species. High-entropy alloys have therefore remained largely out of reach. We present pyeCE, an open-source Python library that implements the embedded cluster expansion (eCE), in which machine learning maps many chemical species onto a smaller set of effective species and thereby limits the growth in the number of cluster functions. pyeCE provides the complete modeling workflow, including the construction of symmetry-adapted descriptors with a learnable per-sublattice chemical embedding, a neural-network energy model, ladder-based training, uncertainty quantification, and finite-temperature simulations through Monte Carlo sampling. Built on the PyTorch and pymatgen libraries, it supports systems with multiple species on multiple sublattices and runs on graphics processing units. We demonstrate the package on two material systems. In the first, a single model spanning the full composition space of a 9-component refractory alloy resolves short-range order and order–disorder behavior. This model enables rapid screening for compositions with strong Cr clustering, a feature linked to the formation of a continuous, corrosion-resistant oxide scale. In the second, a model of hydrogen dissolution in a Mo–Nb–W alloy reproduces the composition dependence of hydrogen uptake and resolves the interstitial environments that hydrogen occupies. The modular design of pyeCE allows it to be extended to problems beyond alloy thermodynamics, including kinetics, defect energetics, and the coupling of chemical order to magnetic and vibrational degrees of freedom.
Read the full abstract
Chemical Physics
Efficient Deterministic-Stochastic Representation of the Coulomb Operator in Real Space
We present an efficient mixed deterministic-stochastic approach for constructing the Coulomb operator in real space that preserves accuracy across the full spectral range. The dominant long-range components of the interaction are captured deterministically via a compact, low-rank approximation, constructed using Chebyshev-filtered subspace iteration, while the remaining spectral tail is treated using unbiased probing with a small number of stochastic vectors. The resulting operator is benchmarked within self-consistent field calculations for the extended Hubbard Hamiltonian on periodic, perturbed, and non-periodic three-dimensional lattices. We account for stochastic bias in observables by implementing a jackknife correction. The method systematically improves with deterministic rank and stochastic sample count, while substantially reducing the computational cost of constructing the Coulomb matrix.
Read the full abstract
Soft Condensed Matter
Topology-dependent mixing of knots in flexible polymer chains
The organization of multiple knots along a polymer contour can depend sensitively on the topology of the constituent knots. Using coarse-grained Brownian-dynamics simulations, we investigate topology-dependent mixing of two knots in a flexible polymer at controlled normalized wall separation. We compare homogeneous ($3_1\otimes 3_1$) and heterogeneous ($3_1\otimes 5_1$) knot pairs using a contour-based coordinate that distinguishes mixed and demixed configurations. Starting from partially overlapping configurations, the $3_1\otimes 5_1$ pair is observed in mixed configurations more frequently over the simulation time window than the $3_1\otimes 3_1$ pair at intermediate wall separation. We further find that, over the intermediate range of normalized wall separation, mixing of the $3_1\otimes5_1$ pair produces a composite knotted region that occupies substantially less polymer contour than the two constituent knots collectively occupy in demixed configurations. The corresponding reduction in contour occupancy is much smaller for the $3_1\otimes3_1$ pair. This indicates that the heterogeneous constituent topologies can be accommodated within a smaller portion of the polymer contour in the composite configuration. These results establish that constituent topology can strongly influence multiple-knot organization in flexible polymers.
Read the full abstract
Computer Vision
3rd Place Solution to Human Motion Challenges in Real-World and Clinical Settings (MoCha) @ECCV2026: Language-Aligned Motion Representations for Domain-Generalizable UPDRS-Gait Severity Estimation
In this work, we introduce language-aligned motion representations for domain-generalizable UPDRS-Gait severity estimation, aiming to learn semantically structured motion features that generalize across heterogeneous clinical domains. We first learn motion representations using a Bi-GRU backbone that captures the temporal dynamics of SMPL sequences. Prior to model training, motion captions are generated offline using Qwen2.5-7B-Instruct. The backbone is then trained with both classification and text-alignment objectives to learn discriminative and semantically structured motion representations while accounting for the class imbalance present in the training data. We subsequently adapt the learned backbone independently to each source domain so that the model can capture domain-specific motion characteristics. The resulting source-specific models are then merged at the parameter level to consolidate complementary knowledge across source domains into a single domain-generalized model. To further mitigate class imbalance, we perform GPT-5.5-based pseudo labeling, and our final merged models for each site do not use any class-prior correction during inference. The resulting model is evaluated under the unseen-site setting of the MoCha Challenge, using Macro F1 as the primary evaluation metric. Our method achieves a macro-F1 of 0.57 on the hidden test set with only 637K active parameters at inference, ranking 3rd among 58 leaderboard entries in the MoCha 2026 Challenge. The challenge attracted 1,669 submissions from 112 participants and offered monetary prizes sponsored by Machine Medicine Technologies.
Read the full abstract
Strongly Correlated Electrons
Fast, Accurate, and Scalable Fermionic Neural Networks via Translation Equivariance
We demonstrate that designing a neural quantum state to be an exact eigenstate of the Hamiltonian's symmetries significantly improves both training speed and final variational energy. For the 2D electron gas, we design TorFormer, a neural network wavefunction which is an exact eigenstate of the total momentum. TorFormer describes both the Fermi liquid and Wigner crystal with no supervision and significantly outperforms Psiformer-based references up to large system sizes. For $r_s = 30.0$ and $40.0$ at $N=91$, we compare TorFormer trained for $8\mathrm{K}$ steps against the previous best NQS, which required $100\mathrm{K}$ training steps. Our improvement to the total energy at $r_s = 40.0$, excluding the trivial Madelung part, is $0.12\%$—enormous compared to the tiny differences separating phases. Relative to Slater-Jastrow-backflow diffusion Monte Carlo, TorFormer's energy decrease is roughly $9.8$ times that of the previous best NQS. Our work demonstrates that neural quantum states can both accurately and efficiently solve large-scale problems.
Read the full abstract
Systems and Control
From Laboratory Aging Studies to Field Predictions: Quantifying Uncertainty in Battery Storage Lifetime Predictions
Predicting how long a battery energy storage system will last is critical for warranty design, maintenance planning and investment decisions, yet degradation models are mostly deterministic and rarely validated against real field data. We apply an open-source probabilistic degradation framework, combined with a cell-to-system approximation, to bridge the gap between cell-level laboratory aging models and system-level field predictions for residential battery energy storage systems with quantified uncertainty. The framework predicts cell-level state-of-health to within 0.4 % mean absolute error, roughly half the error of prior models for this dataset. When applied to field operation data, the framework's predictions are consistent with all three available system-level capacity measurements - a benchmark rarely available for open probabilistic degradation models. Cell-level heterogeneity is approximated by two bounding stress scenarios differing only slightly (a 5 % spread in temperature and a 9 % spread in current). The mean degradation trajectories of the two scenarios reach end-of-life 10 months apart, while the full predicted end-of-life range across both scenarios spans approximately three years, about a third of the expected system lifetime. We link this uncertainty to two drivers. One is a systematic mismatch between laboratory test conditions and field-representative operating stress. The other is variability in the training data itself. These insights translate into concrete, resource-efficient recommendations for future aging study design, supporting more confident predictions of battery lifetime under real-world conditions.