Systems and Control
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 14 September
Domain Randomization (DR) has been widely used to overcome the sim-to-real gap by training a controller on a distribution of simulated environments via reinforcement learning. While DR can achieve robust performance simply using controllers synthesized via policy gradient (PG) methods, the optimization landscape is not well understood, even in the case of linear quadratic regulator (LQR) objectives. To this end, we first study PG of domain randomized LQR over history-dependent policy classes, such as finite impulse response controllers, as they can extend the possibilities of simultaneous stabilization. Second, to find such a stabilizing controller, we propose a curriculum learning based algorithm which gradually expands the memory of the controller. Finally, we show that PG with the proposed algorithm converges globally to the minimizer of a sample average approximation of the DR objective under suitable bounds on the heterogeneity of environments. Empirical results support our findings and highlight promising directions for future work, including nonlinear domain-randomized control.
Programming Languages
Revisiting Soundness for Occurrence Typing, Semantically
Over the past two decades, numerous systems have brought some of the benefits of dependent typing to a wide variety of new programming languages, often by restricting which terms can appear inside types. Such techniques are known as refinement types, occurrence typing, liquid types, and path dependent types, among others. However, the restrictions adopted by these systems often break the substitution property, because they explicitly disallow the ability to substitute arbitrary terms for variables inside types. This leads to significant complexity in the design and metatheory of these systems, increasing the possibility of significant errors. We consider a specific line of work on occurrence typing, namely, the calculus underlying Typed Racket due to Tobin-Hochstadt and Felleisen 2010. We show that the fundamental challenge of substitution into types resulted in multiple flaws in the formalism and the syntactic type soundness theorem of this work. These flaws are replicated in several other papers building on this work, and also surface as a soundness bug in Typed Racket itself. We identify and repair these problems, revising the core calculus of Typed Racket and giving a semantic type soundness proof using step-indexed logical relations, formalized in Lean. We argue that this approach is simpler than it may seem, and easily scales to handle the complexity of the occurrence typing in Typed Racket.
Read the full abstract
Artificial Intelligence
Closing the Loop: Branch-and-Bound for Scalable Verification of Nonlinear Neural Feedback Systems
Despite recent advances in the verification of nonlinear neural feedback systems, scalability remains the central obstacle, as state-of-the-art solvers do not yet handle the network sizes and nonlinear dynamics of autonomy applications. Combinatorial solvers do not scale to large networks, whereas propagative solvers excessively sacrifice precision. This work seeks to improve the scalability of combinatorial solvers by formulating verification as branch-and-bound on an abstraction of the closed-loop system. We introduce \rail, an interface that exposes polyhedral enclosures of the dynamics to LiRPA-style bound propagation, and \clipper, a branch-and-bound algorithm that jointly refines enclosures and splits controller activations. This framework enables joint reasoning on the computational graph of the closed-loop system, preserving symbolic correlations across time steps. We present our construction and show that it yields significant improvements over the state of the art.
Read the full abstract
Probability
Renyi's Relative Entropy Lyapunov Functional for Convergence to Invariant Measure in Dissipative Stochastic Hamiltonian Systems
This paper is concerned with multivariable stochastic Hamiltonian systems governed by an ordinary differential equation for the position and an Ito stochastic differential equation for the momentum. The momentum dynamics involve both conservative and nonconservative forcing including Langevin viscous damping and the Wiener process as a random force. The setting allows for rotational degrees of freedom, a non-quadratic potential energy function and position-dependent mass, diffusion and damping matrices. In addition to modelling nonlinear physical dynamics subject to random forcing, the system is motivated by a stochastic optimisation viewpoint, where the search for global minima of the potential energy (for example, using the gradient descent) is carried out through convergence to a Boltzmann equilibrium measure which favours smaller values of this function. Under certain regularity conditions, we show that the $\chi^2$-divergence or, equivalently, Renyi's second-order relative entropy with respect to the invariant measure has a nonpositive time derivative and thus provides a Lyapunov functional candidate for this convergence. However, this derivative vanishes at some moments of time, which is related to the behaviour of the probability distribution of the momentum conditioned on the position and has a link to a family of auxiliary quantum harmonic oscillator Hamiltonians parameterised by the position space of the underlying classical system. This position-momentum conditioning is used in order to prove that such pre-equilibrium break times in the entropy dissipation are isolated, thus making the $\chi^2$-divergence a strictly decreasing Lyapunov functional for the Fokker-Planck-Kolmogorov equation which governs the joint position-momentum distribution in such systems.
Read the full abstract
Dynamical Systems
Asymptotic separation of periodic orbits and fractal dimension
For a totally bounded metric dynamical system $(X, d, T)$ we introduce a new critical value $\mathfrak s(X, d, T)$ which quantifies the asymptotic separation of periodic orbits. More precisely, $\mathfrak s(X,d,T)$ is defined to be the supremum of all $s\ge 0$ for which there exists a sequence of periodic orbits $\{\mathcal O_k\}_{k=1}^\infty$ in $(X,d,T)$ such that \[ \lim_{k\to \infty} \#\mathcal O_k = +\infty \quad \text{and} \quad \liminf_{k\to\infty}\#\mathcal O_k\cdot\eta(\mathcal O_k)^s >0, \] where $\eta(\mathcal O_k)$ denotes the smallest distance between distinct points in $\mathcal O_k$. When the dynamical system is induced by a self-similar iterated function system $\mathcal F$ satisfying the strong separation condition, we prove that this critical value is equal to the Hausdorff dimension $s_{\mathcal F}$ of its self-similar attractor. Furthermore, under a quantitative separation condition on periodic orbits we show that the associated empirical periodic measures converge weakly to the normalized $s_{\mathcal F}$-dimensional Hausdorff measure on the self-similar attractor.
Read the full abstract
Human-Computer Interaction
Intelligent Interaction Techniques (IIxT) - Proposal
Interaction techniques (IxTs) are the low-level, reusable components out of which user interfaces are designed, including menus, scroll bars, text input fields, and also copy-paste, text-entry, and selecting objects. The IxTs for graphical user interfaces (GUIs) were well established in the 1980s, with relatively minor additions and tweaks for smartphones in the 2000s. Most of today's AI user interfaces involve a chat window, which is an excellent interaction for some tasks, but is generally considered separate from the GUI IxTs. I argue for making the IxTs themselves more intelligent, so users can freely mix modalities, even within the same interaction. This will require research into new IxTs, and also into the infrastructure that will enable these intelligent IxTs (IIxTs) to be built. There are also significant security, privacy and economic implications to this vision.
Read the full abstract
Statistics Applications
Nationally Consistent, Locally Incomplete: A Bayesian Remote-Sensing Audit of Rooftop Photovoltaic Registries
Tracking the energy transition requires reliable statistics on renewable deployment. Rooftop photovoltaics (PV) are especially hard to track, owing to their decentralised nature, and the resulting inaccuracies in official statistics are known but not quantified. Remote sensing offers an independent way to identify rooftop PV systems. We introduce a Bayesian framework to estimate the ground-truth rooftop PV capacity from remote sensing detections, turning an imperfect detector into an uncertainty-aware measurement instrument. Applied to France, the corrected detections estimate a capacity of 4.03 GWp [3.96–4.11] (99% credible interval) of rooftop PV below 36 kWp, matching the transmission system operator's connection data within 3.3% nationally, while identifying local under-reports of up to 61% of local capacity. We also document and quantify a significant truncation bias in French rooftop PV open data. Beyond France, the approach paves the way for more reliable estimates of rooftop PV capacity worldwide.
Read the full abstract
Signal Processing
Discrete Coupling and Localized Motion for Pinching-Antenna Systems (PASS)
The practical implementation of pinching-antenna systems (PASS) is challenging due to hardware limitations in large-scale antenna movement and continuous radiation power adjustment. This paper proposes a practical PASS-enabled downlink multi-user multiple-input multiple-output communication framework that enables discrete radiation power control and localized discrete antenna movement. Specifically, a discrete coupling strength model is exploited to tune the radiation power at each pinching antenna (PA) through quantized coupling spacing levels. Moreover, each PA can only move among discrete locations within a limited region determined by the movement speed and duration. Based on the proposed framework, a joint optimization problem of the PA positions, coupling strength, and transmit beamforming is formulated. Considering waveguide attenuation, the total average power consumption is minimized, subject to each user's minimum SINR requirement and localized motion constraints. To address this coupled mixed-integer nonconvex optimization problem, a globally optimal branch-and-bound-based algorithm is first developed for the multi-waveguide single-user scenario. To further reduce complexity, a scalable genetic algorithm-assisted particle swarm optimization (GA-PSO) method is developed for the multi-waveguide multi-user scenario, where GA operations are incorporated to preserve population diversity and alleviate premature convergence. Simulation results demonstrate that the proposed design significantly reduces the power consumption compared with the conventional PASS schemes and MIMO architectures.
Read the full abstract
Nuclear Theory
Comprehensive study of $\Upsilon$($n$S) production in Oxygen+Oxygen collisions at LHC
We study bottomonium $\Upsilon$($n$S) production in ${\rm O}+{\rm O}$ collisions at $\sqrt{s_{NN}}=5.36$~TeV, including both cold nuclear-matter (CNM) effects and final-state interactions in a short-lived quark-gluon plasma. The CNM baseline is calculated using the EPPS21 nuclear parton distribution functions together with coherent energy loss and transverse momentum broadening. Hot-medium effects are evaluated with two transport approaches coupled to the same centrality-resolved 3+1D aHydroQP background: the QTRAJ-NLO open-quantum-system calculation, implemented with the QTraj code, and the TAMU-NP semiclassical rates, including primordial suppression and regeneration. The CNM effects give a moderate, state-independent suppression, dominated by the nPDF modification. The energy loss and transverse momentum broadening contribution remains at the few-percent level. Both transport approaches predict a sequential suppression pattern, $R_{AA}^{\Upsilon(1\rm{S})}>R_{AA}^{\Upsilon(2\rm{S})}>R_{AA}^{\Upsilon(3\rm{S})}$, with the strongest suppression around midrapidity in central collisions. The suppression of the ground state is similar in the two approaches but differs for the excited states. Since the CNM factor is common to all states, double ratios provide a direct measure of the relative hot-medium suppression. Our results support the conclusion that ${\rm O}+{\rm O}$ collisions at the LHC can produce a short-lived QGP capable of modifying bottomonium yields.
Read the full abstract
Nuclear Experiment
Modernization in processing and dissemination of experimental cross section data in EXFOR
The EXFOR library is a nearly complete database for experimental radionuclide production cross sections for light charged-particle induced reactions, and retrieval of the experimental cross sections in EXFOR is the first step in determination of the best estimate of the radionuclide production cross section. The EXFOR format was designed with 80-column punched cards as the recording media, but the format is still preferred for compilation and exchange of the nuclear reaction data. The format is suitable for reading and processing by EXFOR compilers, but it is not a good choice for reading by EXFOR users with modern computer languages such as Python. We have recently attempted to lower the barrier for modern software engineers to access the information stored in the EXFOR by developing tools converting an EXFOR file to (1) the HTML format for human reading (X4IVEW), and (2) the CX4 (Compact EXFOR) format for tabulation and plotting (X4TOCX). This article introduces X4VIEW and a preliminary version of X4TOCX with the radionuclide production cross section as an example.
Read the full abstract
High Energy Astrophysical Phenomena
Formation of Merging Black Hole Binaries Inside Massive AGN Stars
Stars embedded in the disks of active galactic nuclei (AGN) can grow to hundreds of solar masses, and the same disks are expected to host a population of stellar-mass black holes. A star may therefore capture passing black holes, turning its interior into a potential factory for forming compact binary black hole systems and, ultimately, for the gravitational-wave events seen by our detectors. We test how readily this channel operates using three-dimensional hydrodynamic simulations of black hole–star encounters. As a proof of principle, we adopt a star-to-black-hole mass ratio of $34\!:\!1$, and find that the capture does not disrupt the structure of the star: it heats the star by $10$–$30\%$, and the star loses only $\lesssim1\%$ of its mass throughout. The captured black hole loses its orbital angular momentum rapidly to dynamical friction, sinking to the stellar center within a stellar dynamical time $\lesssim10^{4}\,$s. If the star already harbors a black hole at its center, the two form a bound binary whose gravitational-wave coalescence time falls below $10^{4}\,$yr. Depending on the geometry and speed of the encounter, the resulting orbit may be nearly circular or retain substantial eccentricity. Our calculations confirm that black hole–star encounters in AGN disks are indeed a viable channel for assembling stellar-mass black hole binaries and supplying sources for gravitational-wave detectors.
Read the full abstract
Game Theory
Symmetric solution of the Bellman optimality equation for repeated harmony game
In social dilemma games, additional rewards or punishments have been studied as means of promoting cooperation. Therefore, it is important to investigate the ideal situation, in which such an additional payoff would change the game. In this study, we investigated the symmetric solution of the Bellman optimality equation for a repeated harmony game. The calculations showed that three types of symmetric solutions exist. One of them corresponds to the trivial All-C strategy, and another to the Win-stay Lose-shift strategy of the prisoners dilemma game. The nontrivial behavior of the strategy corresponding to the last solution is also discussed in detail. In addition, we numerically investigated which strategy the agents actually learn by the reinforcement learning algorithm.
Read the full abstract
Machine Learning
Drift Field Net: Learning Ocean Lagrangian advection fields from in-situ and satellite observations
The North Pacific Subtropical Gyre (NPSG) is a major accumulation zone for floating plastic debris, resulting from basin-scale convergent ocean circulation. Effective cleanup strategies in this region rely on accurate forecasts of Lagrangian particle drift. Here, we introduce Drift Field Net (DFN), a deep neural network that predicts ocean surface flow fields from operational satellite observations. DFN is trained using a novel two-stage strategy that combines pretraining on simulated data with Lagrangian fine-tuning based on an advection-consistent loss function. This physics-informed optimization directly improves the accuracy of particle trajectory predictions. We evaluate DFN against an operational physics-based forecasting system and demonstrate the potential of deep learning for ocean surface flow prediction. On in situ drifter trajectories, DFN reduces the mean positioning error by 20 km after a 7-day forecast compared with the operational model. Furthermore, Lagrangian fine-tuning with the proposed advection loss further reduces the positioning error by 10 km, highlighting the benefits of incorporating Lagrangian constraints into the training process.
Read the full abstract
Software Engineering
AgentGuard: Learning Execution Guardrails from Anomalous Coding-Agent Trajectories
AI coding agents increasingly rely on execution harnesses to interact with repositories and external tools. However, task success does not guarantee reliable execution. Agents may still modify unrelated files, rewrite tests, issue unsafe commands, or ignore failed validations, motivating behavioral guardrails for reliable execution. We present AgentGuard, an instruction-level guardrail framework that learns conditional execution constraints from anomalous trajectories of coding agents. Rather than relying on manually specified safety rules, AgentGuard automatically extracts recurring execution failure patterns, generalizes them into instruction-level behavioral constraints, and organizes them as a lightweight guardrail skill that dynamically activates only the rules relevant to the current instruction. This design enables behavioral guidance while minimizing unnecessary restrictions on normal execution. We evaluate AgentGuard using 642 documented failure traces collected from real coding-agent executions across 382 repository tasks. Guardrails are learned from 461 traces covering 282 tasks and evaluated on a disjoint set of 100 tasks. Using Claude Code with Claude Haiku 4.5 as the underlying coding agent, we compare the baseline agent with the same agent augmented by AgentGuard. Experimental results show that AgentGuard reduces the Abnormal Execution Rate from 69.0% to 26.7% and increases the Successful Task Completion Rate from 21.7% to 35.0%. These results demonstrate that execution guardrails learned from historical failures can substantially improve the reliability of AI coding agents while highlighting the remaining challenge of balancing safety and task completion.
Read the full abstract
Combinatorics
A Cheeger Inequality for Hypergraphs and Its Applications
Hypergraphs provide a natural framework for modeling higher-order relationships, but the development of spectral techniques with provable guarantees for general non-uniform hypergraphs remains challenging. Building on Banerjee's normalized adjacency matrix and Spiro's averaging-based diffusion framework, we develop a spectral framework for non-uniform hypergraphs and establish Cheeger's inequality for their conductance. A fundamental result in the spectral theory of hypergraphs asserts that, for every non-covering hypergraph, the second-smallest eigenvalue of its normalized Laplacian is at most one. This spectral characterization yields an improved Cheeger's inequality for non-covering hypergraphs, and we show that the resulting inequality is tight on both sides using cycle and cube hypergraphs. Our framework further yields higher-order Cheeger inequalities and provides theoretical guarantees for Fiedler's spectral partitioning algorithm, all in the setting of hypergraphs. Finally and most notably, we construct a new family of optimal hypergraph expanders that is tight for the Alon-Boppana bound.
Read the full abstract
Physics and Society
Higher-order interactions reveal synergistic backbones of cycling infrastructure networks
Infrastructure networks essentially underlie human mobility and transport. Improving the quality of single links increases network performance locally. However, efficient transport requires high-quality connected corridors across multi-link paths that do not emerge from independent single-link upgrades. Here, we introduce a framework for evaluating the impact of jointly upgrading multiple links as inherently higher-order interactions, enabling us to quantify link synergies in complex transport networks. Two links are synergistic if an upgrade of one increases the benefit of upgrading the other, promoting upgrades of topologically complementary links along the same path while discouraging upgrades of redundant parallel links. By expressing these synergies as second-order derivatives of overall network performance, we develop an efficient computational framework to identify synergistic links that form a connected network backbone. We apply our theoretical framework by combining empirical street network and cycling demand data for Hamburg, Germany, with a perturbed utility route choice model for urban bicycle traffic. Our results reveal synergies from higher-order interactions, thereby enabling strategic infrastructure planning that goes beyond local link importance in complex transport and flow networks.
Read the full abstract
Computer Vision
ProtoLIP: From Sentence-Level to Object-Level Evidence Disentanglement
Query-conditioned vision-language models enable fine-grained interpretation by revealing which visual content supports a given textual query and how this evidence changes across queries. However, semantically, sentence-level evidence does not necessarily decompose into object-specific contributions, while spatially, object-level evidence can remain entangled with co-occurring objects and surrounding scene context. Across multiple VLM architectures and independent benchmarks, we observe persistent object-level evidence entanglement. Moreover, exposed evidence maps do not necessarily correspond to the evidence that directly constitutes the model's prediction. To disentangle visual evidence at both semantic and spatial levels, we introduce ProtoLIP, a lightweight prototype-mediated evidence layer that organizes reusable visual prototypes into text-derived semantic families and uses coarse-to-fine evidence routing, where semantic families constrain prototype eligibility and the complete query determines fine-grained prototype contributions. Our studies show that ProtoLIP improves evidence localization and separation across query granularities, achieving average relative gains of 29% in Pointing and 43% in Energy across four object- and phrase-level OOD benchmarks. Its localization gains also transfer to independently pretrained VLMs, with larger improvements observed in several transfer settings. On the primary backbone, ProtoLIP also improves image-text matching discrimination while remaining competitive with a spatially supervised grounding model in object-level localization. Crucially, ProtoLIP constructs its image-text matching score directly from localized prototype evidence, enabling exact decomposition across prototypes, semantic families, and spatial evidence without spatial annotations or backbone retraining.
Read the full abstract
Machine Learning
Differentially Private Semantic Plans for Aggregate Insight Generation
URANIA provides end-to-end differential privacy (DP) for summaries of data-dependent clusters. However, its cluster–keyword release does not directly provide collection-wide aggregates for semantic concepts defined independently of the protected corpus. Records may express several concepts, records expressing the same concept may be assigned to different clusters, and cluster identities need not correspond across analyses. Consequently, cluster-level statistics do not directly provide comparable measurements of predefined concepts across collections or repeated analyses. We introduce DP-SPIN, a trusted-curator framework for aggregate measurement and summarization over semantic concepts fixed independently of the protected target records. Each record is mapped to a bounded sparse nonnegative vector over these concepts, whose sum forms a semantic sketch. A differentially private mechanism releases a semantic plan containing admitted concepts and noisy masses; normalized semantic-support values and support bins are obtained by post-processing. For user-level privacy, each user's aggregate contribution is clipped to a fixed bound. The language model receives only the plan and fixed decoding instructions, while a public verifier checks concept mentions, reported values, comparisons, and rank claims against the released plan. The final summary is differentially private by post-processing. We establish record- and user-level DP guarantees under add/drop and replacement adjacency. We evaluate DP-SPIN under record-level privacy on CFPB complaint narratives, Amazon All Beauty reviews, and Yelp restaurant reviews, and under user-level privacy on Amazon and Yelp. We compare DP-SPIN with non-private plan and summary references, DP keyword and category histogram baselines, and a URANIA-style baseline with a fixed public keyword vocabulary.
Read the full abstract
Machine Learning
Scaling Laws for Physics-Aware ACOPF Surrogate Learning
Learning-based surrogates for AC optimal power flow (ACOPF) promise large speedups over classical solvers, but their operational value depends on physical feasibility as much as predictive accuracy. Physics-aware objectives such as the augmented Lagrangian (AL) improve constraint satisfaction at additional per-step cost, yet how this trade-off behaves with scale is uncharacterized. We sweep model and dataset sizes under MSE and AL training, and measure how violation changes with network size. Both objectives improve as power laws: MSE prediction loss depends more on model size than on data, while the AL composite of prediction loss and violation improves comparably along both axes. Violation grows markedly more slowly with network size under AL. On matched hardware and equal training samples, AL reduces violation about $19\times$ with $2.4\times$ the training time and negligible added memory. The training objective shapes not only where a surrogate lands but how its quality evolves with scale.
Read the full abstract
Rings and Algebras
Regular Sublattices, O Closed Ideals and Dedekind MacNeille Completions
We study regular sublattices of lattices that satisfy the Join-Infinite Distributive Law (JID) under Dedekind MacNeille completion. Using order convergence, we identify the lattice of O-closed ideals with Crawleys completion, and compare it with the Dedekind MacNeille completion. When DM(L) satisfies (JID), this comparison yields extension results for lattice homomorphisms. In particular, if Y is a regular sublattice of a complete lattice L, and L and DM(Y) satisfy (JID), then the canonical extension lattice-embeds DM(Y) into L with join-regular image. We also give a sharp finite-dimensional obstruction: there exists an infinitely distributive regular sublattice of R3 whose Dedekind MacNeille completion is not distributive, and therefore cannot be lattice embedded in R3. In contrast, we prove that the inclusion of every join-regular sublattice of R2 with a least and a greatest element extends canonically to a lattice embedding of its MacNeille completion into R2.
Read the full abstract
Image and Video Processing
Semantic-Aware Neural Video Codec for Error-Resilient Low-Latency Transmission
Emerging physical AI systems require low-latency, task-oriented video communication over unreliable channels. We propose a semantic-aware multi-level neural video coding method for robust low-latency video transmission over unreliable channels that are abstracted as multi-level packet erasure channels. Built upon the real-time DCVC-RT neural video codec, the proposed framework introduces a semantic- and feature-aware coding strategy that partitions encoded representations into packets carrying different levels of semantic and latent-feature importance and assigns these packets to different streams, each associated with a priority level when transmitted over unreliable communication channels. We also developed an error-resilient entropy model that removes inter-packet dependencies, allowing each packet to be decoded independently under packet losses. The complete system is trained end-to-end over the abstracted multi-level packet erasure channels, enabling learning of channel-aware representations together with importance-aware packet assignment while facilitating the network for differentiated packet prioritization. Experiments show that the proposed framework significantly improves robustness over baseline DCVC-RT under packet erasures, achieving graceful degradation in less important regions while better preserving task-relevant visual content.
Read the full abstract
Superconductivity
Current - voltage characteristics in magnetic field for layered superconductor with intrinsic pinning of Josephson vortices
The results of measurements of the transverse voltage in the Hall configuration, Uy, under conditions of motion of the charge and vortex subsystems in orhtogonal magnetic and electric fields on epitaxial films of the layered high-temperature superconductor Nd2-xCexCuO4 are presented. the motion of Josephson vortices across CuO2 layers (along the c-axis) is analyzed considering the effects of intrinsic pinning. An exponential dependence of critical depinning current jc(B) for Josephson vortices is established in contrast to the power law dependences for Abrikosov vortices. With increasing external magnetic field after the depinning by a magnetic field at B=Bc, the onset of the vortex creep regime, which evolves into the flux-flow regime at crossover field B=Bf, is observed.
Read the full abstract
High Energy Physics (Experiment)
Probabilistic, Structure-Intrinsic Clustering with a Hierarchical Embedding (PSICHE) in Time and Space Applied In Particle Physics Jet Reconstruction
Jet reconstruction is an active and open area of particle physics research, with questions related to jet size, multiplicity, substructure, and experimental performance in the presence of pileup and noise. Analogous challenges are ubiquitous throughout other applications of unsupervised learning in the context of clustering and association. This paper introduces a new algorithm for clustering in the context of jet reconstruction: Probabilistic, Structure-Intrinsic Clustering with a Hierarchical Embedding (PSICHE). PSICHE addresses several challenges and shortcomings of existing jet clustering methods with novel features, including but not limited to: (i) dynamically-learned and variable jet (cluster) sizes, (ii) unsupervised, multi-scale learning of emergent features, like jet substructure and multiplicity, (iii) clustering in space and time, and (iv) incorporation of domain-specific experimental uncertainties. All of these properties are achieved in a self-consistent, probabilistic framework that is computationally tractable. Furthermore, demonstrations of these features are presented for a range of jet clustering examples in LHC scenarios for various physics phenomena, clustering inputs, pileup conditions, and detector performance parameters.
Read the full abstract
Systems and Control
An Integrated EMT Small-Signal Stability Analysis Tool for Power Systems with High Inverter-Based Resource Penetration
The ongoing replacement of synchronous generation by inverter-based resources (IBRs) introduces fast converter control dynamics whose characteristic frequencies extend beyond the classical electromechanical band. Conventional small-signal stability tools are commonly formulated in the phasor domain, representing electrical quantities as slowly varying phasors at the fundamental frequency, and therefore cannot resolve the sub-synchronous and converter-driven dynamics that increasingly arise in operation. Electromagnetic Transient (EMT) modeling captures these dynamics, but established EMT simulators produce time-domain waveforms rather than the modal indicators (eigenvalues, damping ratios, and participation factors) needed to assess stability risk. This paper presents EMT-SSA, an integrated tool suite for EMT-level small-signal analysis of IBR-rich power systems. From a standard PSS/E system snapshot (.raw/.dyr), it converts the model into an EMT representation with enhanced fidelity, solves an extended power flow for the steady-state equilibrium points of both the network and the device controllers, and linearizes a full-order EMT model to form the system state matrix, from which it produces eigenvalues, oscillation modes, and participation factor analysis results over a device library spanning synchronous generators, grid-following and grid-forming inverters, transmission lines, and loads. The EMT-SSA tool suite identifies poorly damped or unstable modes and attributes them to specific devices. The tool is demonstrated on the Kundur two-area system, with its modal results benchmarked against PSS/E NEVA.
Read the full abstract
Optimization and Control
Dimensional Decomposition and Column Generation for Bin Dimensioning in E-Commerce Fulfillment Centers
A fulfillment center (FC) is a specialized logistics facility where e-commerce inventory is received, stored, and processed. Its internal organization directly impacts space utilization and the performance of put-away and picking operations. Items are stored in bins following strict operational guidelines to ensure productivity in both activities, so bin dimensions must be tailored to the product profile of each facility to maximize space utilization. % We address the bin dimensioning problem arising in e-commerce fulfillment centers: given a set of candidate bin types and a large, heterogeneous inventory, determine the number of bins of each type and assign all items so as to minimize total bin volume, subject to operational constraints governing how products may be combined within a single bin. We develop a dimensional decomposition that exploits these constraints to reduce the three-dimensional packing problem to a one-dimensional block-positioning problem. Building on this decomposition, we formulate the problem as a mixed-integer linear program and propose two solution methods: a Best-Fit-Decreasing (BFD) heuristic and a column generation scheme that produces tight LP lower bounds. Computational experiments on four synthetic datasets, calibrated on proprietary data from an American e-commerce fintech and ranging from 1.5 thousand to 1.5 million SKUs, show that the BFD heuristic scales to instances with up to 1.7 million blocks, and that column generation certifies a BFD–LP gap of 2.26% on the smallest dataset. The experiments further reveal a structural divide in solution quality across datasets, driven by the interaction between operational constraints and each facility's product profile.
Read the full abstract
Multiagent Systems
Cheap Talk Stabilizes Strategic Interaction in LLM Agents
Large language models are increasingly deployed as interacting agents, making the persistence of their action policies across repeated interaction critical for reliable multi-agent operation. We investigate whether and how agent-generated, non-binding pre-play communication ("cheap talk") increases such persistence in four open-weight 7-9B-parameter LLMs. Our experiments span four repeated two-player games – Prisoner's Dilemma, Snowdrift, Stag Hunt, and Harmony – with incentive structures ranging from strategic conflict to alignment, each presented in six contexts. We observe unstable trajectories in all four games, although their prevalence and magnitude depend strongly on model and context. Across models, games, and contexts, cheap talk is predominantly stabilizing, with five corrected reversals concentrated in social or team framings; effects vary substantially by model and context. Controlled current-message interventions identify two separable output-level channels in Qwen: reduced action uncertainty and less between-round drift in action probabilities. Matched history-by-message counterfactuals further show that recent partner behavior conditions how mutual-benefit versus self-prioritizing language affects policy persistence. Finally, in Prisoner's Dilemma, we identify in Qwen and Falcon a history-balanced policy-content direction in late transformer layers; projecting out this direction increases realized switching during closed-loop play, demonstrating that complete trajectories are causally sensitive to this component. Together, these findings show that cheap talk can make individual trajectories more persistent across diverse incentive structures, while revealing that the magnitude and mechanisms of stabilization are model- and history-dependent.
Read the full abstract
Rings and Algebras
Relative commuting probability and BFC-type results for finite skew braces
We develop a probabilistic approach to the structure of finite skew left braces, motivated both by classical commuting probability in finite group theory and by the role of skew left braces in the study of set-theoretic solutions of the Yang–Baxter equation. We introduce relative commuting probability for skew left braces and establish analogues of several classical structural results, including relative BFC-type theorems. For natural classes of finite skew left braces, we show that a positive lower bound for the commuting probability forces the existence of a large section which is trivial up to bounded subgroups. We also study the probability that two elements generate a trivial sub-skew brace and a Sylow-local version of commuting probability, obtaining further structural consequences. Finally, we introduce a commuting probability for finite non-degenerate set-theoretic solutions of the Yang–Baxter equation. Under natural hypotheses on the associated structure skew brace $G(X,r)$, a positive lower bound for this probability yields a bound on $|G(X,r):\Soc(G(X,r))|$ depending only on the probability and on $|X|$.
Read the full abstract
Computation and Language
Spurious Tool Use: When RL Agents Learn the Wrong Reason to Act
Large language model (LLM) agents increasingly interleave natural language reasoning with external tools such as web search and code execution. These tool-use policies are often optimized via reinforcement learning (RL), which can amplify spurious correlations in the training data. In this work, we study when and why RL-trained agents learn shortcut tool-selection policies: invoking tools based on superficial prompt cues rather than genuine task requirements. We construct controlled synthetic environments combining factual question answering and mathematical reasoning tasks, and inject cues that are strongly correlated with specific tools during training but causally irrelevant to tool necessity. Across counterfactual evaluations where cues are present but the associated tools are not required, agents exhibit substantial shortcut behavior, with spurious tool invocation rates increasing by up to 39 percent. However, shortcut formation is not universal: across the conditions we test, it arises only when the agent has already learned to use the target tool reliably, suggesting that task competence, rather than dataset imbalance alone, is a key factor in shortcut learning. A swapped-cue analysis further shows that semantic alignment between cues and tools substantially amplifies this effect. To mitigate these failures, we introduce a dense, decision-level reward in which an LLM judge evaluates the necessity of each tool call. This tool-necessity reward effectively suppresses cue-driven tool use while preserving task performance, providing a practical approach to improving the robustness of LLM agent tool-use policies.
Read the full abstract
Machine Learning
The record is part of the task: matched-record evaluation of text classifiers across maintenance, safety and recall reporting
Many operational cases are documented more than once, at different workflow stages and for different purposes, yet model evaluations normally select one of these records before model comparison begins. We treat that selection as part of the evaluation and compare matched records of the same cases under fixed labels and splits in three systems: GE Aerospace repair events, NASA ASRS safety reports and NHTSA vehicle recalls. The design fixes, for any decision a classifier would support, the records that existed and could be read when it was taken, and compares them retrospectively, as written. Across the three GE fields, for events whose label comes from parts transactions independently of the narratives, held-out macro-F1 ranged from 0.33 to 0.91. A difference of 0.46 separated the customer report, written before shop work, from the technician report, written after diagnosis but before the transaction that generates the label. That difference is substantially larger than the representation and architecture differences tested on the same events. The public systems showed different patterns: the NHTSA defect summary remained strongest under every model family tested, whereas the ASRS analyst synopsis outperformed the reporter narrative under learned sequence models but not under lexical baselines. Secondary analyses showed that some model comparisons were also record-dependent. Evaluations should be run on the information available at the intended decision point and should report how both the record and the label were produced.
Read the full abstract
Quantum Physics
Towards Surrogate Based Dequantization of Quantum Reinforcement Learning
In recent years, the utility of parameterized quantum circuits as function approximators has been widely studied. In the context of reinforcement learning, this approach has led to variational quantum algorithms such as quantum Q-learning. While these methods show promising empirical results, and can provide provable advantages for artificial problems, it remains unclear whether they can provide a provable quantum advantage over classical approaches for problems of practical relevance. A natural way to investigate this question is through the lens of dequantization: The construction of efficient classical algorithms capable of matching the performance of quantum variational methods. Building on recent kernel-based dequantization results for supervised learning, we take steps towards extending this surrogate-based dequantization program to reinforcement learning. Specifically, we study the simplified setting of reinforcement learning with a uniform generative model in which uniformly random state-action samples are available, which models the regime of sampling from a large experience replay buffer after sufficient exploration. Within this setting, we provide finite sample guarantees for classical kernelized Fitted Q-Iteration, with classical kernels designed to match the inductive bias of particular parameterized quantum circuits. Using these results, we then provide a set of sufficient conditions, on the data-encoding strategy of a parameterized quantum circuit, the corresponding classical kernel, and the problem structure, under which kernelized Fitted Q-Iteration provides a meaningful dequantization of quantum Q-learning, in this simplified setting. Apart from providing rigorous dequantization guarantees when these conditions are met, these results also motivate the use of kernelized fitted Q-iteration as a dequantization heuristic when these sufficient conditions cannot be verified.