Peer Revenue Play credits. Real papers.

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Tuesday 8 September

Information Theory

Stability of Fork-Join Systems with Redundancy and Heterogeneous Servers

We consider the stability problem of fork-join systems with redundancy (FJR) and heterogeneous servers under both static and dynamic capacity-allocation policies. In an $(n,k)$ FJR system, each arriving job is split into $n$ independent tasks, with one task assigned to each of $n$ parallel servers. Once $k \le n$ tasks have been processed, they are joined and the corresponding job departs the system; the remaining $n-k$ unprocessed tasks are then removed and are therefore termed redundant. We first identify the nominal traffic intensity and characterize the maximal stability region, defined as the set of traffic intensities for which there exists an admissible policy that stabilizes the system. We then establish conditions under which this maximal stability region is attained for two classes of policies: static and dynamic. Specifically, we show that for static allocation policies, in which service capacities remain fixed over time, maximality is achieved whenever the fastest server is allocated no more than $1/k$ of the total service capacity. For dynamic allocation policies, in which a fixed total service capacity may be repeatedly reallocated among the servers, we show that maximality is achieved whenever the cumulative capacity allocated to the $j$ shortest queues does not exceed $j/k$ of the total capacity for every $j=1,\ldots,k-1$. Our analysis is based on a projection of the $(n,k)$ FJR system onto a simpler $(k,k)$ system that has no redundancy, together with a novel sample-path comparison argument for multidimensional processes based on the generalized Schur-convex order.
Read the full abstract

Physics and Society

Identifiability of Latent Space Network Models on Anisotropic Thurston Geometries

A latent space network model places the nodes in a metric space and lets the probability of a tie decrease with distance. In a space of constant curvature, pairwise distances determine the positions up to an isometry. In the products and in the three remaining three-dimensional model geometries they do not. We study the three geometries that are neither of constant curvature nor products: the Heisenberg group, the solvable group and the universal cover of the unit tangent bundle of the hyperbolic plane. We ask what one network identifies about positions in them and when their geometry is detectable. Two anchors remove the isometry ambiguity. Small configurations are not determined by their distances, and generic local identification holds beyond a finite threshold, certified in the Heisenberg group. We derive the posterior on the quotient by the isometry group. For small configurations, the divergence to the nearest product or constant-curvature competitor is the stress component of the curvature difference and vanishes at high order in the scale; undirected ties therefore detect the geometry only in large networks with large enclosed areas. Directed ties expose it at first order: in the two twisted geometries, asymmetric preferences circulate around triangles in proportion to enclosed area, which no additive ranking produces. On dense competitive-game counter networks, the coupled model beats rankings on every geometry, degree-corrected rankings and free antisymmetric terms. A Euclidean model with the same coupled term matches it, so the gain is the coupling of similarity and circulation through shared coordinates. An additive-and-multiplicative-effects model predicts better still.
Read the full abstract

Robotics

TANGO: Humanoid Navigation in Cluttered Environments with a Whole-Body Vision-Language-Action Model

We study the problem of navigating cluttered indoor environments with a humanoid robot. Unlike conventional methods that model navigation as a 2D path planning problem, humanoid traversal in cluttered environments requires continuous geometry-aware whole-body adaptation, including coordinated arm placement, torso adjustment, and gait modulation for collision-free movement through complex 3D spaces. We introduce TANGO, the first whole-body vision-language navigation framework for language-conditioned humanoid traversal in cluttered environments. Given a natural-language instruction and egocentric RGB observations, TANGO directly predicts 29-DoF joint-space actions for downstream whole-body control. We train TANGO entirely in simulation by synthesizing diverse collision-free traversal behaviors via global path planning, kinematic whole-body motion generation, obstacle-aware motion editing, and RL-based tracking. This pipeline provides dynamically feasible action supervision for learning language-conditioned whole-body policies. In extensive simulation experiments, TANGO demonstrates state-of-the-art performance in vision-language navigation, while outperforming strong modular baselines in navigating challenging scenes requiring obstacle negotiation. Lastly, we deploy TANGO zero-shot on a Unitree G1 humanoid robot, and observe robust language-guided traversal in cluttered real-world scenes without training on any real-world navigation data.
Read the full abstract

All authors Anqi Li, Yuxin Chen, Zhaobo Li, Zhuo Cao, Junli Ren, Masayoshi Tomizuka, Dhruv Shah

Machine Learning

Learning Length-Extrapolatable Recurrent Models

Recurrent models provide a natural path to long-context modeling, yet models trained with backpropagation through time (BPTT) often fail beyond their training horizon. Classical analyses emphasize gradients that vanish or explode along temporal paths. However, dense per-token losses can still train a shared recurrent rule despite severe decay, showing that decay alone does not determine whether learning fails. We instead study state credit: the signal through which future losses reach earlier recurrent states before contributing to parameter updates. Accordingly, we intervene directly on state credit and propose Credit Stabilization through Time (CST). During backward propagation, CST locally rescales the state-credit signal to stabilize its norm without rotating the component being corrected, while leaving the forward computation unchanged. Because controlled synthetic tasks and real data exhibit different credit dynamics, we specialize CST to each regime. In both settings, CST improves performance beyond the training horizon, with gains observed at up to 128x the training length.
Read the full abstract

Computation and Language

ReCite: Agentic Reasoning for Faithful Citation

Accurate citations are the foundation of academic writing, tracing intellectual origins and substantiating core claims. However, manually navigating the growing volume of scientific literature is increasingly difficult, prompting reliance on automatic citation recommendation. While modern retrieval-augmented architectures have largely mitigated the fabrication of non-existent papers, current systems relying on semantic similarity struggle with misattribution, often citing authentic papers that fail to logically support the author's claim. To address this challenge, we argue that accurate citation requires a shift from similarity-based search to active, claim-level reasoning. We propose ReCite, a decoupled agentic framework that orchestrates location perception, intent-aware query planning, and reflective verification. Trained on synthesized reasoning trajectories, our agent verifies claim-evidence consistency and triggers self-correction loops when retrieved candidates lack logical support. Experiments demonstrate that our lightweight framework outperforms state-of-the-art massive generative models in strict citation accuracy. By grounding literature matching in verifiable logic rather than semantic overlap, ReCite establishes a reliable foundation for automated academic writing.
Read the full abstract

All authors Yuyang Huang, Bobo Li, Jiajia Song, Yuzhe Ding, Chong Teng, Fei Li, Donghong Ji

Computer Vision

SyncWorld: Visual Calibration Enables World Models as Zero-Shot Simulators

World models are increasingly used as policy-in-the-loop imagination environments, where reliable rollouts require fine-grained controllability with respect to low-level robot actions. A key obstacle to scaling such models in robotics is that actions are not a universal language in pixel space: changes in visual environment, camera view, robot placement, or embodiment alter how the same numerical action manifests visually, leading to conflicting supervision under mixed training and brittle generalization at deployment. We introduce SyncWorld, an action-conditioned world model that serves as a zero-shot simulator across unseen environments without any additional training. SyncWorld leverages a visual calibration episode—paired frames and actions that showcase all the controllable degrees of freedom—to specify the setup-specific Action–Visual Mapping in context. Training with visual calibration contexts teaches the model to interpret actions through visual evidence and to leverage interaction history when explicit calibration is unavailable. Experiments show that SyncWorld can accurately simulate action outcomes in previously unseen settings, and that its capability of simulating rollouts enables test-time policy improvement without training.
Read the full abstract

All authors Yuncong Yang, Zhengtao Han, Furkan Ozyurt, Zeyuan Yang, Han Yang, Junyi Cao, Haoyu Zhen, Yilun Du, Chuang Gan

Algebraic Geometry

Vector Bundles of Coinvariants for Admissible Affine Vertex Operator Algebras

Simple affine vertex operator algebras $L_k(\mathfrak{g})$ at non-integral admissible levels $k$ are generally neither $C_2$-cofinite nor rational. Despite the absence of these standard finiteness conditions, we prove that the sheaf of coinvariants (and, dually, of conformal blocks) of ordinary modules in the category $\mathcal{O}_k$ forms a vector bundle over the moduli space $\overline{\mathcal{M}}_{0,n}$ of stable $n$-pointed genus-zero curves. The proof relies on establishing the finite-dimensionality of genus-zero coinvariants, a vanishing theorem that isolates ordinary admissible modules at the boundary, and a smoothing theorem enabled by partial strong identity elements in the mode transition algebra. Consequently, admissible affine vertex operator algebras supply a rich class of non-rational, non-$C_2$-cofinite examples in which vector-bundle behavior is preserved for a well-behaved subcategory of modules.
Read the full abstract

Artificial Intelligence

Procedural Graphs: Self-Evolving Execution Structures for LLM Agents

Large language models are increasingly deployed as agents that plan over long horizons and act through external tools. Most agents select actions through unconstrained generation over an accumulating history, leaving implicit the procedural knowledge of what to do, in what order, and under which conditions. As trajectories lengthen, agents can lose track of their objectives, invoke tools out of order, and repeat unproductive actions. We introduce the Procedural Graph: just as a knowledge graph organizes factual knowledge into (entity, relation, entity) triplets for what-is questions, a Procedural Graph organizes procedural knowledge into (procedure, relation, procedure) triplets for what-to-do questions. At each decision step, the framework localizes the agent's active node, and a guidance model translates the surrounding subgraph into step-level situational guidance that biases the solver's next action without dictating it. The graph is self-evolving: an LLM refiner contrasts failed trajectories with successful ones and edits the graph's topology and attributes, committing edits that preserve or improve held-out validation performance while retaining rejected ones to discourage repetition. Starting from a minimal skeleton, the loop builds graphs that match or surpass hand-designed ones. It can also repair a flawed expert prior. Across multiple datasets, task types, and LLMs, the Procedural Graph delivers consistent gains over memory-based baselines, and self-evolution further improves performance without manual engineering.
Read the full abstract

All authors Yuxing Lu, Yicheng Chen, Shanchan Wu, Sercan Ö. Ar{\i}k

Optimization and Control

Silver Rate Is (Almost) Optimal for Gradient Descent

We study how far gradient descent (GD) can be accelerated by predetermined stepsizes in smooth convex optimization. Writing $p_{\mathrm{sil}}=\log_2(1+\sqrt{2})$, we prove an $\Omega\left(n^{-p_{\mathrm{sil}}-O(\sqrt{\log\log n/\log n})}\right)$ non-anytime lower bound. In the anytime setting, every infinite schedule has infinitely many horizons with error $\Omega\left(n^{-\frac{2p_{\mathrm{sil}}}{1+p_{\mathrm{sil}}}-O(\sqrt{\log\log n/\log n})}\right)$. Together with the silver-schedule upper bound [Altschuler and Parrilo, 2025] and the anytime upper bound [Zhang et al., 2025], our results determine the optimal polynomial convergence exponents in both settings.
Read the full abstract

Cosmology and Nongalactic Astrophysics

Sizing the Universe with DESI Galaxy Sizes: Plain Fundamentals of Fundamental-Plane Lensing

Weak gravitational lensing provides a powerful way to map cosmic structure, but most current measurements rely on galaxy shape distortions from deep imaging surveys and are affected by systematics such as intrinsic alignments, photometric-redshift uncertainties and shape-measurement biases. Here we present a spectroscopic galaxy-galaxy lensing magnification measurement using Fundamental-Plane (FP) size residuals, $\delta_r\equiv\Delta\log_{10}R_\mathrm{e}$, from 3.26 million DESI luminous red galaxy (LRG) sources behind DESI Bright Galaxy Survey lenses. The FP-like relation predicts the intrinsic sizes of LRGs from lensing-invariant quantities, including velocity dispersion $\sigma_0$ and surface brightness $I_\mathrm{e}$, with a scatter of about 0.06-0.07 dex. Lensing magnifies LRG sizes, giving the direct convergence response $\delta_r(\kappa)=\kappa/\ln 10$, but we show that the full lensing response is modified by magnification bias, because fitting $R_\mathrm{e}$ with $I_\mathrm{e}$ inevitably induces a magnitude dependence in $\bar{\delta}_r(m)$. After calibrating this response, we recover surface-density profiles with uncertainties comparable to those from individual Stage-III shear surveys using 5-20 million higher-redshift sources. The corresponding excess surface-density profiles agree with shear-based measurements. We further show that the estimator is robust to size-measurement uncertainties, with a convergence multiplicative bias only $\simeq -0.2$ times the size bias. FP lensing therefore provides a clean, spectroscopic and complementary probe of cosmic structure.
Read the full abstract

All authors Kun Xu, Jiaqi Wang, Ravi K. Sheth, J. Aguilar, S. Ahlen, F. Beutler, D. Bianchi, D. Brooks, A. Carnero Rosell, T. Claybaugh, A. de la Macorra, P. Doel, J. E. Forero-Romero, E. Gaztañaga, G. Gutierrez, K. Honscheid, C. Howlett, M. Ishak, J. Jimenez, S. Juneau, R. Kehoe, D. Kirkby, O. Lahav, M. Landriau, L. Le Guillou, M. Manera, A. Meisner, R. Miquel, J. Moustakas, S. Nadathur, J. A. Newman, W. J. Percival, F. Prada, I. Pérez-Ràfols, C. Ravoux, G. Rossi, R. Ruggeri, E. Sanchez, C. Saulder, D. Schlegel, M. Schubnell, H. Seo, J. Silber, M. Siudek, G. Tarlé, B. A. Weaver, H. Zou

Multiagent Systems

Copying explains the collective behavior of AI agents in the wild

In June 2026, thousands of AI agents found that a small public wiki would accept edits from inside their sandboxes, and started using it to help one another pass a timed test. Each agent lived for about an hour and remembered nothing afterwards. Nobody asked them to cooperate, and the wiki had not been built for them. The complete record of what they wrote is public, and it is unusually informative, because it preserves not only what each agent wrote but what that agent could see before writing. We use it to follow the three decisions an agent had to make on arrival: where to write, what to call itself, and how to word its message. One rule governs all three. An agent takes an option with a probability close to the share of that option in what it can see, and the share that matters is the one on the page in front of it, then the one in the stream of recent edits, and only weakly anything older. Three minimal copying models, one per decision and with a single free parameter each, reproduce the heavy-tailed distribution of how many agents met on a page, the frequency of the pieces from which the agents built their names, and the patchwork of pages that are internally consistent and different from one another. Copying whatever the environment happens to show is enough to produce most of the collective structure of this population. It is also what makes such a population easy to steer, since whoever writes first, or writes while the others are quiet, sets the convention for everyone who comes later.
Read the full abstract

Cosmology and Nongalactic Astrophysics

Differential Polarization Calibration: A Consistency Test for Cosmic Birefringence

The cosmic microwave background birefringence angle $\beta$ is exactly degenerate with a common instrumental polarization-angle offset. The Minami–Komatsu(MK) likelihood separates these quantities using Galactic foreground polarization, so its inferred detector angles merit a diagnostic that does not reuse the foreground-$EB$ model. We develop a birefringence-blind differential estimator from antisymmetric and symmetric cross-spectra and apply it for the first time as an explicit consistency test of a foreground-assisted birefringence analysis. For a map pair, the common rotation cancels identically and the estimator measures $\Delta\alpha_{ij}=\alpha_i-\alpha_j$; combining all pairs in a gauge-fixed network reconstructs the observable detector-dependent calibration pattern. Applied to eight Planck NPIPE detector-set maps, the differential reconstruction agrees with the MK pattern, giving $\chi^2=8.18$ for seven degrees of freedom (PTE=0.32), with all shared-data residuals below $1.8\sigma$. Because both reconstructions use the same maps, this is an independent estimator and modeling route rather than an independent-data confirmation. The agreement therefore strengthens confidence in the calibration foundation of the MK-inferred $\beta$ without providing an independent absolute measurement of it. A conditional common-mode reconstruction gives $\beta=0.37\pm0.12^\circ$. The differential test uses existing data and can serve as a validation layer for future $EB$-based birefringence analyses.
Read the full abstract

Robotics

Proxy Policy Steering

Generalist robot policies carry broad manipulation priors from large-scale data, but specializing them to a new task remains the deployment bottleneck. This requires eliciting task-specific behavior from limited demonstrations without degrading their broad capabilities. We introduce Proxy Policy Steering (PPS), an inference-time adaptation method that resolves this challenge by training two lightweight proxy policies whose calibrated velocity-space difference steers the frozen base sampler. A reference proxy models the frozen base's behavior on target-task observations, and a task proxy, initialized from the reference, captures how this behavior changes under task supervision. Their difference forms a calibrated velocity-space residual that steers the frozen base sampler at every denoising step. We identify the conditions under which this residual isolates the change induced by task supervision, and validate them empirically. Because the base is never directly modified, its broad capabilities remain available at inference, including behaviors such as recovery from failure that the demonstrations themselves do not exercise. Adaptation requires only forward velocity predictions from the base, making PPS lightweight to train and applicable even without access to the base's parameters. On 8 real-world and 4 simulation manipulation tasks, PPS lifts the state-of-the-art pi 0.5 base policy by 53% absolute success rate on average, with zero-to-one gains on tasks the base never solves, while preserving the base's broad capabilities. PPS outperforms LoRA fine-tuning, from-scratch specialists, residual policies, and prior inference-time steering methods.
Read the full abstract

All authors Chuanruo Ning, Tianrui Wang, Wei-Chiu Ma, Kuan Fang

Strongly Correlated Electrons

Green's Functions from Sample-based Krylov Quantum Diagonalization: An Impurity Solver for Dynamical Mean-Field Theory

We generalize the sample-based Krylov quantum diagonalization (SKQD) method from ground-state calculations to the evaluation of single-particle Green's functions. By constructing and sampling unitary Krylov subspaces in the N +/- 1 particle-number sectors and evaluating all sector-connecting overlaps classically, the approach reconstructs the Green's function via a Lanczos continued fraction while retaining the shallow-circuit, ancilla-free character of SKQD. The quantum device is required only to prepare and sample short-time evolutions. Applied to the particle-hole-symmetric single-impurity Anderson model in chain geometry, with the discrete bath representation used in dynamical mean-field theory (DMFT), the method recovers the spectral function using a relatively small fraction of the full Hilbert space. Across a range of interaction strengths that spans the metal-insulator transition, the main spectral features are reproduced. These results suggest that SKQD-based Green's-function calculations may allow DMFT impurity solvers with a larger number of bath sites on near-term quantum hardware than is currently practical.
Read the full abstract

Discrete Mathematics

The exact asymptotic constant in the metric dimension of Jaccard space

Let $X$ be a finite set with $|X|=n$ and let $\mathrm{Jac}(a,b)=|a\,\triangle\, b|/|a\cup b|$ be the Jaccard distance on the power set $2^X$. Lladser and Paradise recently proved that the metric dimension of $(2^X,\mathrm{Jac})$ is $\Theta(n/\ln n)$, with the constant left open; their bounds are $(\ln 2)\,n/\ln n\lesssim \beta(2^X,\mathrm{Jac})\lesssim 2\ln(2e)\,n/\ln n$. We determine the constant: \[ \beta(2^X,\mathrm{Jac})=\frac{2n}{\log_2 n}\,(1+o(1))=(2\ln 2)\,\frac{n}{\ln n}\,(1+o(1)). \] The proof identifies the problem, on each “slice” of subsets of fixed cardinality, with the Erdős–Rényi coin-weighing problem for a spring scale (the problem of detecting matrices). The lower bound is the Erdős–Rényi entropy argument applied to the middle slice; the upper bound follows from the explicit detecting families of Lindström and of Cantor and Mills, augmented by a single extra landmark that reveals cardinality.
Read the full abstract

Computer Vision

Point4D: Long-range 4D Motion Reconstruction

We introduce Point4D, a feed-forward model for 4D reconstruction of long-range video sequences. Point4D is able to reliably infer dense per-point 3D trajectories across multi-hundred-frame videos, unlike existing 4D methods that are limited to short input windows of at most a few dozen frames. A key innovation that enables this is our flexible 3D query-based motion decoder that decouples trajectory prediction from image-plane visibility. The predicted 3D endpoints are then directly re-queried in the next chunk without re-projection or matching. Furthermore, we show that extracting and reusing a visual descriptor from an arbitrary frame where the point is visible leads to better performance than relying solely on the source patch. Overall, Point4D achieves state-of-the-art performance across diverse long-video tracking benchmarks spanning over 200 frames and largely outperforms previous feed-forward 4D method. Project page: https://point-4d.github.io
Read the full abstract

All authors Minsik Jeon, Jay Karhade, Deva Ramanan, Shubham Tulsiani

Cosmology and Nongalactic Astrophysics

Baryonification IV: Constraining baryonic feedback with X-ray gas fractions

Baryonic feedback redistributes gas around dark matter halos, suppressing the matter power spectrum at scales now probed by weak lensing surveys. X-ray observations directly trace this hot gas, and are one of the main probes of its distribution and properties. We present a forward-modelling framework, built on the baryonification model, linking the three-dimensional gas density and temperature profiles of groups and clusters to observed X-ray surface brightness and luminosity profiles on one side, and to matter power spectrum suppression on the other. We validate the model against independent three-dimensional density reconstructions from the literature, and examine our temperature and metallicity treatment in the group-scale regime. Applying this framework to the SZ-selected CHEX-MATE and X-ray-selected eFEDs samples, we measure gas fractions across the group-to-cluster mass range while accounting for X-ray selection effects, with the first published gas fractions based on CHEX-MATE data. Combining both samples, we derive a joint constraint on the hot gas fraction retained by groups and clusters as a function of mass and on the baryonic suppression of the matter power spectrum. We find $f_{\rm gas} = 0.029 \pm 0.006$ at $M_{500c} = 3\times 10^{13}M_\odot$, $f_{\rm gas} = 0.078 \pm 0.004$ at $M_{500c} = 3\times 10^{14}M_\odot$, and suppression of 6% at $k=1\,h/\rm Mpc$ and 23% at $k=5\,h/\rm Mpc$. Our findings are consistent with recent kinematic Sunyaev-Zel'dovich results, hinting at strong feedback. We also show that the $L_X$-$M$ relation is degenerate with feedback strength, and that different feedback scenarios produce distinct X-ray profile shapes that map onto the same $L_X$-$M$ point. This work is a first step toward extending the framework to forward-model diffuse X-ray emission at the map level for simulation-based inference in upcoming wide-area X-ray surveys such as eROSITA.
Read the full abstract

All authors Jozef Bucko, Andrina Nicola, Aurel Schneider, Michael Kovač, Sambit K. Giri, Robert Reischke, Esra Bulbul, Nicolas Clerc, Ang Liu, Iacopo Bartalucci, Alexandre Refregier

Computer Vision

Studying Image Tokenizers as Visual Languages in Unified Multimodal Models

Image tokenizers define the “visual language” of unified multimodal models, yet are commonly studied through isolated metrics or generation-/understanding-only evaluations. These evaluations do not fully capture how visual tokens behave when modeled jointly with text. We build a controlled pure-autoregressive testbed and track task-specific validation losses during multimodal continual pretraining across text, image, text-to-image (T2I), and image-to-text (I2T) prediction. We examine how these losses scale and relate to downstream performance, then use them to study multimodal learnability—how well image and text tokens are jointly modeled—and tokenizer design. We find that (1) losses should be analyzed by task, since they exhibit distinct scaling behavior and rank tokenizers differently. (2) The loss–performance relationship depends on the predicted token space: for a fixed tokenizer, T2I and I2T losses correlate with generation quality, but across tokenizers, the T2I loss–performance relationship shifts with the image-token space, whereas I2T loss, computed over a shared text vocabulary, provides a more consistent signal. I2T loss also correlates with both generation and visual understanding performance after supervised finetuning. Using losses as a lens, we show that (3) better reconstruction does not necessarily yield lower task-specific losses or stronger downstream performance, and that (4) image tokenizer choice can affect text modeling under joint optimization. As case studies, we revisit three tokenizer design axes—the discriminator, semantic supervision, and vocabulary size—to examine their effects on joint modeling and downstream performance. Together, our testbed offers a complementary perspective on image tokenizers as visual languages, highlighting their interplay with text in joint multimodal training.
Read the full abstract

All authors Siting Li, Zhengyang Wang, Simon Shaolei Du, Xi Chen, Yang Liu

Combinatorics

Are trees really just butterflies in disguise?

As a generalisation of the Erdős-Sós conjecture about graphs, Addario-Berry, Havet, Linhares Sales, Reed and Thomassé conjectured that every digraph on $n$ vertices with more than $(k-1)n$ arcs contains every antidirected tree with $k$ arcs. We prove a dense, approximate version of this for trees with bounded maximum degree, as well as for trees whose layers are evenly distributed. We use a regularity based approach, centred around finding a copy of a given tree in the blow up of a caterpillar.
Read the full abstract

Algebraic Geometry

The border Waring rank of $x_1\cdots x_n$ is $2^{n-1}$

In this paper, we show that the border Waring rank of $x_1x_2\cdots x_n$ over fields of characteristic zero is exactly $2^{n-1}$. As a consequence, the classical polarization identity is an optimal Waring decomposition even if we allow limits. As a symmetric tensor, this monomial is identified with the $n\times n$ permanent tensor. However, the lower bound is proved via the higher-order Koszul flattening of the $n\times n$ determinant tensor, which is not symmetric.
Read the full abstract

Machine Learning

NOAH: Learning the Full Patient Journey. A Longitudinal Multimodal Time-Aware Model for Representation and Forecasting

The digitization of healthcare has generated vast, longitudinal, and multimodal patient records over a lifetime, yet fully exploiting these data to represent and predict patient state trajectories remains a critical challenge. Current AI models often struggle to capture the complex, irregular temporal dynamics and inherent stochasticity of real-world multimodal patient data. Existing AI approaches for modeling longitudinal patient records are predominantly discriminative, limited to a few modalities, constrained by closed categorical vocabularies, treating time as a monotonic inductive bias, or they are limited in forecasting future patient states. We introduce NOAH, a time-aware, task-agnostic, generative transformer model representing and forecasting the full multimodal patient journey. NOAH features a novel bidirectional time integration and a variational latent space to capture the continuous evolution of patient states and the stochasticity of clinical trajectories. Built from over 559 million clinical events from 431,000 hospital visits of 299,000 patients across the MIMIC dataset family, NOAH natively processes medical images, time-series and numeric signals, categorical events, as well as structured and unstructured clinical records. NOAH is the first truly holistic generative model in its field, enabling autoregressive forecasting with optional time control, zero-shot classification, and counterfactual intervention simulation. It generates highly informative and predictive patient state representations that demonstrate strong performance in probing for clinical outcomes, 15 ICD chapters, and 29 comorbidities, as well as in time-to-event prediction. Seamlessly handling diverse modalities and complex temporal dynamics, NOAH provides a versatile, task-agnostic, scalable foundation for intelligent predictive systems in personalized clinical care and digital medicine.
Read the full abstract

All authors Tobias Susetzky, Raphael Rehms, Dmitrii Seletkov, Özgün Turgut, Michelle Espranita Liman, Lisa Steinhelfer, Rickmer Braren, Daniel Rueckert

Statistical Mechanics

Large-scale dynamics of integrable quenches

The Ballistic Macroscopic Fluctuation Theory (BMFT) is a path-integral based approach to treat large-scale correlations in interacting integrable systems which, so far, has only been applied to quasi-equilibrium settings. In this paper we extend this approach to integrable quench problems and, to illustrate it, we compute the time evolution of the full-counting-statistics (FCS) of a conserved charge. We identify the relevant path integral, solve it via saddle-point, and obtain a set of partial differential equations describing the dynamics of the FCS. We solve these equations explicitly for free fermionic systems and for the interacting cellular automaton Rule 54. In both cases, we recover the known exact results. For Rule 54, our construction moreover completes the solution across the entire ballistic window.
Read the full abstract

High Energy Physics (Phenomenology)

Mixing-suppressed inelastic dark matter: a minimal model for the LZ 248 keV event

We construct a minimal Majorana singlet-vector-like-doublet model for the single nuclear-recoil-like event reported by LZ near 248 keV. At the splitting inferred from the recoil energy, an electroweak-strength $Z$ transition predicts thousands of events; singlet-doublet mixing suppresses the rate without moving the recoil spectrum. Fixed-coupling interpretations instead require a larger splitting in the extreme halo tail, but solar gravitational acceleration removes this suppression, while the larger splitting shifts recoils toward LZ's empty high-energy sideband. In our model the same mixing suppresses solar capture and annihilation, while the small mass gap required for coannihilation weakens Higgs-mediated cooling, allowing the captured population to remain extended and out of equilibrium. Imposing the relic abundance and normalizing the rate to one event leaves a two-dimensional mass-splitting parameter space. A thermalized population gives the conservative IceCube limit $\delta=$300-301 keV; with our nonthermal cooling ansatz and elastic scattering treated at tree level, the nominal limit is $\delta=$331-341 keV near the 730-733 GeV relic-density endpoint, which is also favored by the recoil spectrum. We outline a gauged $U(1)_N$ origin for the parity and splitting and a candidate $R$-symmetric supersymmetric embedding.
Read the full abstract

Artificial Intelligence

A Data-Driven Framework for Identifying and Prioritizing RPA Opportunities in Healthcare Processes

Robotic Process Automation (RPA) is widely used to reduce administrative burden in United States hospitals, yet an estimated 30-50% of RPA initiatives underperform because processes are selected informally, without a repeatable method to catalogue candidates, prioritize them, match each to an automation tier – a Python bot, an open-source orchestrator such as n8n, or an enterprise platform such as UiPath – and forecast financial return before committing resources. We propose a four-module, data-driven framework unifying these decisions: a Process Taxonomy of twenty recurring hospital processes across five value streams; a Prioritization module deriving an Automation Suitability Index from an Analytic Hierarchy Process matrix with an explicit consistency check; a Tool-Tier Selection module recommending the least-cost technology sufficient for a process complexity, integration, and compliance profile; and a Return-on-Investment module quantifying labor savings, error-cost avoidance, payback, and net present value. Applied to a synthetic portfolio spanning all twenty processes, plus a reference data-flow architecture linking it to hospital EHR/payer/ERP systems: 12 of 20 clear the prioritization threshold; the ranking is robust to +/-20% weight perturbation (Spearman correlation 0.83, top-5 set preserved 97.7%, 2,000 Monte Carlo trials); an Automation Risk Index flags four qualifying processes as Critical risk; a budget-constrained portfolio optimization shows diminishing marginal NPV as spend scales from $400K to $1.03M; and a second Monte Carlo analysis shows portfolio NPV stays positive at its 5th percentile. The framework is a conceptual synthesis of the literature rather than an instrument calibrated on primary hospital data; we discuss HIPAA governance and a research agenda for empirical validation. A supplementary Python implementation accompanies the paper.
Read the full abstract

High Energy Physics (Phenomenology)

A Warped Extra Dimensional Candidate for the LZ 248 keV Event

The recent result by the LUX-ZEPLIN (LZ) experiment of a single nuclear-recoil event at $248\,\mathrm{keV}$ suggests an interpretation in terms of inelastic scattering between dark matter (DM) and the xenon target. The widely considered thermal Higgsino DM with $m_{\chi}\sim$~TeV and mass splitting $\delta \sim 350$~keV is essentially excluded as it would be captured efficiently by the Sun, with a resulting neutrino flux from its annihilation in excess of the IceCube limit, and expected recoils above the LZ nominal search window, where none were seen. However we argue that although these tensions apply to a Higgsino with constrained cross section, they need not apply to the broader set of inelastic candidates that alternative solutions to the hierarchy problem might suggest. We show that a new massive gauge boson, $Z'$ with a candidate DM sector and Majorana mass splitting of order $\sim 10^{-7}$ of the Dirac mass, evades current constraints. We propose an inelastic dark matter (iDM) candidate that can arise naturally in a Randall-Sundrum warped extra-dimensional model. The DM candidate involves a vector-like fermion confined to the TeV brane and a bulk field. The brane Lagrangian contains Dirac mass terms that are also naturally exponentially suppressed. The left-handed component of the bulk field has a Majorana mass with origin on the UV brane, while the interaction with Standard Model nucleons is mediated by a massive dark Kaluza-Klein (KK) gauge boson whose mass is also automatically in the right range. We propose a concrete model of thermal DM with $m_{\chi}\sim$~TeV in an RS geometry with warp factor $\pi kR\simeq 31.7$ and the $Z'$ a KK mode with natural TeV-ish scale mass in which the ratio of neutrino to quark coupling is also natural, giving a mass splitting of $\delta \sim 300$~keV, and show how it could explain the LZ signal while being consistent with existing constraints.
Read the full abstract

Machine Learning

Entropy-Regularized Rank-Masked Policy Optimization for Test-Time Reinforcement Learning in Code Generation

Existing methods for test-time reinforcement learning (TTRL) derive rewards from answer-level self-voting on unlabeled test-time tasks with canonical answers, but this breaks down for code generation because programs cannot be compared by surface form and therefore do not directly provide a usable training signal. To make TTRL applicable to code generation, we propose probe-driven TTRL, which constructs output-free probe inputs from the problem statement, executes candidate programs on these probes, and defines a Probe Consensus Reward (PCR) from the resulting behavioral agreement. PCR provides a behavioral training signal for open-vocabulary programs, but it is not a fully reliable verifier and remains susceptible to reward hacking through spurious consensus. We therefore introduce Entropy-Regularized Rank-Masked Policy Optimization (ERPO), which converts low PCR into conservative negative updates through rank masking and controls policy drift with an entropy ceiling. On coding benchmarks, ERPO substantially improves pass@1 and pass@k in both in-domain adaptation and zero-shot transfer.
Read the full abstract

All authors Jiacheng Xu, Feng Chen, Xiuneng Xu, Bo An

Artificial Intelligence

Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails

Agent harnesses (the system prompt, tool set, execution hooks, and context-management scaffolding around a model) are a critical determinant of agentic task success. Automated harness evolution can enable smaller models to perform well on domain-specific tasks at a fraction of frontier-model cost. Since both the harness and model weights shape behavior, we ask how harness evolution and lightweight fine-tuning should be combined. Across seven enterprise agent tasks, we first evolve a harness with the weaker model, then find that a stronger expert often uses it more effectively, suggesting expert supervision could close the remaining gap. However, training the weaker model on the expert's complete trajectories under the evolved harness backfires: performance regresses on all seven tasks by 4 to 30 points across Qwen3-Coder and Gemma 4, even though the same procedure helps under the unevolved harness. Our analysis shows that imitation transfers knowledge and increases scaffold usage, but disrupts model-harness fit: the weaker model adopts the expert's planning strategy without the competence to execute it and no longer matches the harness evolved around its native planning style. We therefore develop an on-policy expert-correction pipeline, automated by a meta-level MLE agent, that localizes the failing turn in the weaker model's own rollout and asks the expert to rewrite only that turn. This preserves the model's planning style and combines the gains of harness evolution and model adaptation. Our results identify and resolve a source of contention between harness and weight updates, yielding a compatibility-preserving recipe for economical co-evolution on domain-specific enterprise tasks.
Read the full abstract

All authors Zhou Yu, Bin Bi, Shiva Kumar Pentyala, Shubham Mehrotra, Sougata Chaudhuri, Shilpa Bhagavath, Zeyuan Chen, Ran Xu, Phil Mui, James Zhu, Sitaram Asur

Artificial Intelligence

ExecCritic: Learn to Test, Test to Improve for Coding Agents

Execution feedback can guide coding agents toward correct repository repairs, but only when the tests capture the behavior requested by the issue. Agent-generated tests can encode incomplete or incorrect behavioral targets; when the same trajectory writes both the patch and the test, their errors can agree and create false confidence. We introduce ExecCritic, combining a test–verify–revise scaffold with a role-specific reinforcement learning recipe for training agents within it. The scaffold separates test construction from source-code repair: a Test agent independently generates repository-native tests, a fail-closed harness qualifies and freezes them, and a Repair agent revises source code from their execution feedback without changing the tests. Both roles use Qwen-3.5-35B-A3B as the backbone and are trained separately. In Learn to Test, the Test agent learns to produce behaviorally valid tests that distinguish correct from incorrect patches. In Test to Improve, the Repair agent learns both direct task resolution and feedback-guided revision. On SWE-bench Verified, test quality determines whether feedback helps: holding the base Repair agent fixed, tests from the base Test agent reduce resolved rate from a no-test baseline of 61.2% to 57.3%, whereas tests from GPT-5.6-sol raise it to 65.3%. Role-specific post-training raises the Qwen Test agent's Base-to-Gold success from 22.2% to 62.2%; composing the two post-trained Qwen agents reaches 72.6%, an 11.4-point gain over the original no-test baseline without stronger-model or Oracle feedback at evaluation time. Code is publicly available at https://github.com/MSR-Orchard/execcritic.
Read the full abstract

All authors Leitian Tao, Baolin Peng, Haorui Wang, Hang Wang, Hao Cheng, Wenlin Yao, Qianhui Wu, Tao Ge, Sharon Li, Jianfeng Gao

Quantum Physics

Near-optimal synthesis of non-Gaussian phase gates via qubit-oscillator Rabi control

Non-Gaussian gates remain a key bottleneck for universal continuous-variable (CV) quantum computation because the nonlinearities they require are difficult to engineer. To address this challenge, we develop an efficient qubit-oscillator Rabi synthesis scheme for polynomial phase gates, with a total interaction time that scales polylogarithmically with the inverse target error \(\varepsilon\). Specifically, for a class of readily preparable initial states, we show that a degree-\(R\) phase gate can be approximated by an analytically constructed Rabi sequence with total time \(O(\log^{(R-1)/2+o(1)}(1/\varepsilon))\). This construction requires no numerical optimization and therefore extends naturally to arbitrarily large multimode systems. We further establish a total-time lower bound of \(\Omega(\log^{(R-1)/2}(1/\varepsilon))\), showing that the synthesis is near optimal. As applications, we use this scheme to simulate representative CV quantum dynamics and implement a CV quantum algorithm for solving linear partial differential equations. These results establish qubit-oscillator Rabi control as an efficient, analytically compilable, and near-optimal primitive for CV quantum information processing.
Read the full abstract

All authors Zhen Yang, Shan Jin, Zi-Wen Liu, Xiaoting Wang

Cosmology and Nongalactic Astrophysics

Mitigating baryonic effects in weak lensing with higher-order statistics

Weak gravitational lensing is a premier cosmological probe, but its small-scale statistical power is compromised by baryonic feedback. Higher-order statistics capture non-Gaussian information that the power spectrum misses, yet their sensitivity to feedback remains a concern for Stage IV surveys. We quantify how unmodeled feedback biases the cosmological parameters inferred from the angular power spectrum (PS), starlet peak counts, and the starlet $\ell_1$-norm, and we determine the scale cuts needed to remove that bias. We also test the Bernardeau-Nishimichi-Taruya (BNT) transform as a strategy for more precise scale cuts. Our analysis is based on the cosmoGRID V1 suite, which imprints feedback on dark matter convergence maps with a baryon correction model, and on simulation-based inference with neural posterior estimation, carried out across footprints ranging from Stage III-like to the full sky. We find that biases grow with survey area, exceeding $2\sigma$ for all three statistics at Stage IV-like footprints, and removing them costs a substantial fraction of the signal. Restricted to these “baryon-safe” scales, the starlet $\ell_1$-norm still reaches a figure of merit almost twice that of the PS. The BNT transform localizes the baryonic sensitivity to the lowest transformed redshift bin and improves the PS figure of merit by a factor of $\sim$1.4, while its linear mixing of shape noise inflates the contours of the map-based higher-order statistics. Higher-order statistics therefore deliver a substantial gain over the power spectrum with no baryonic modeling at all, and an even larger one as modeling improves.
Read the full abstract

All authors Andreas Tersenov, Sacha Guerrini, Jean-Luc Starck, Martin Kilbinger