arXiv:2606.14288v1 Announce Type: new Abstract: This work investigates how restitution modeling affects the dynamics of rocking blocks subjected to harmonic excitation. While several studies have reported discrepancies between experimentally observed impact behavior and the predictions obtained using the classical Housner restitution coefficient, the implications of adopting alternative restitution formulations on the global dynamics of rocking systems remain largely unexplored. The system is formulated as a hybrid non-smooth dynamical model and analyzed through bifurcation diagrams, Lyapunov exponents, and basins of attraction for different slenderness ratios. By comparing the classical restitution model proposed by Housner with the alternative formulation of Mao et al., we show that the choice of restitution model strongly influences the predicted system response. The alternative formulation leads to an earlier onset and greater prevalence of complex oscillations, as well as changes in the type, stability, and accessibility of attractors compared to the classical model. However, as the slenderness ratio increases, the dynamical features produced by both formulations progressively converge, indicating a reduced sensitivity to the restitution model for taller blocks. These results provide a dynamical perspective on why alternative restitution formulations, which predict impact responses closer to experimental observations, can produce markedly different behaviors from those obtained using the classical Housner model.
Science Journals
arXiv:2606.14302v1 Announce Type: new Abstract: LLM-based agents trained with reinforcement learning optimize step-wise action prediction but lack metacognitive awareness of task progress, inducing a gap that hinders long-horizon scaling. A pilot study reveals that online progress prompting hurts performance while retrospective demonstrations help, yet this capability cannot emerge from outcome-reward training alone. We present RePro, Retrospective Progress-Aware Training, a framework that trains agents to self-generate progress signals via a forward-then-reflect rollout paradigm: the agent executes actions online, then retrospectively reassesses its step-wise progress given the completed trajectory and known outcome. RePro initializes with a Retrospection Warmup that teaches reflection format from minimal external demonstrations, then further trains through RePro-PO with a composite reward that produces self-generated signals without continuous external supervision. Experiments on WebShop, ALFWorld, and Sokoban show that RePro enhances the Qwen family's performance, with up to $12\%$ absolute success rate gains.
PLRTune: Importance Pre-Sampling and LLM-Guided Reinforcement Learning for Automatic Database Tuning
arXiv:2606.14312v1 Announce Type: new Abstract: Configuration tuning is critical to database performance, yet automatic database tuning remains challenging due to high-dimensional knob spaces, substantial online tuning cost, unreliable textual hints derived from Large Language Models (LLMs) or community documents, and the difficulty of exploiting the remaining optimization room after initialization. Hence, we propose PLRTune, a staged database tuning system that leverages workload-specific domain knowledge to identify a reduced search space and further optimize within this promising region. First, we develop an importance pre-sampling and reranking strategy to identify the dominant knob subset in a workload-specific manner and derive a compact state representation. Second, we design an execution-guided hint refinement technique to improve the initialization quality of documentation-guided tuning. Finally, we propose a post-tuning refinement stage that leverages Twin Delayed Deep Deterministic Policy Gradient (TD3) to explore the dominant knob subset and further exploit the remaining optimization room. We evaluate PLRTune on MySQL and PostgreSQL across diverse benchmark workloads. Compared with state-of-the-art approaches, PLRTune achieves the best final result on all tested workloads, improving over the corresponding best-performing alternative by 9.50% on average. Moreover, PLRTune reaches the strongest baseline's best performance level 9.03 times faster on average across workloads, demonstrating its practical runtime efficiency without sacrificing final tuning quality.
arXiv:2606.14188v1 Announce Type: new Abstract: We present CORD-SLS, a real-time control method for safe deformable object manipulation, with a focus on ropes and cloth. At its core is a GPU-parallel differentiable simulator with contact smoothing which enables efficient gradient-based planning through intermittent contact. To robustly satisfy constraints under model and sensing uncertainty, we develop a real-time, GPU-parallel output-feedback robust model predictive control (MPC) algorithm that plans with this simulator. We further show that the simulator accelerates model-based RL for training neural manipulation policies. To improve real-world robustness, we use conformal prediction to calibrate visual-feedback and perception-error bounds for MPC, producing reachable tubes that enable high-probability safe control. We evaluate CORD-SLS on high-dimensional, contact-rich rope and cloth manipulation tasks in simulation and hardware, including obstacle avoidance, routing, folding, and smoothing. Across settings, CORD-SLS achieves millisecond-speed planning, exceeding baselines in safety, speed, and task success.
arXiv:2606.14315v1 Announce Type: new Abstract: The ML literature contains many distinct concepts falling under the heading of 'AI alignment'. After noting three concepts of AI alignment in the context of their corresponding research programs, we claim that realistic interventions may promote 'AI alignment' under one conception while being actively counterproductive from the perspective of others. We suggest that tensions between alignment ideals emerge due to differences in background threat-models, alongside differences in normative orientations. In light of our analysis, researchers aiming to further the goal of 'AI alignment' should do five things. First, they should not conflate distinctions of policy and distinctions of scientific scope; second, methodological disagreements should be acknowledged explicitly; third, researchers should distinguish between 'AI alignment' as a high-level ideal and specific 'alignment proxies' used in empirical research; fourth, they should use more granular concepts to identify both the source and nature of possible AI harms/benefits; fifth, they should explicitly acknowledge the diversity of 'alignment' concepts in both empirical work and in communication with non-technical audiences.
arXiv:2606.14316v1 Announce Type: new Abstract: Classical high-order backward differentiation formula (BDF) methods for the Landau-Lifshitz-Gilbert (LLG) equation often suffer from restrictive stability constraints, requiring small time steps and imposing stringent lower bounds on the damping parameter. These limitations become particularly severe for schemes of order higher than three. In this paper, we develop a class of high-order generalized BDF (GBDF) schemes for the LLG equation, including both semi-implicit and fully explicit treatments of the gyromagnetic term. The proposed schemes significantly improve stability properties and substantially relax the damping parameter constraints, but introduce essential difficulty in its analysis compared to the classical BDF schemes. We construct a novel multiplier which enables us to carry out a energy-based error analysis. This approach yields optimal-order error estimates under considerably weaker assumptions on the damping parameter than those required for classical BDF schemes. Numerical experiments are presented to confirm the theoretical results, and demonstrate that the proposed GBDF schemes achieve higher accuracy, enhanced stability, and much wider admissible damping regimes compared to classical high-order BDF methods.
arXiv:2606.14324v1 Announce Type: new Abstract: Instantaneous pitch estimation plays an important role in analyzing steep pitch variations such as speech prosody and singing techniques. Conventional approaches estimate instantaneous frequency after isolating the fundamental waveform from signals that contain harmonics and noise, which makes the accuracy sensitive to imperfect fundamental filtering. In this study, we formulate fundamental waveform filtering as a speech enhancement problem. Specifically, we train a Wave-U-Net model to extract a fundamental waveform from an input speech signal. The instantaneous pitch is then obtained by computing the instantaneous frequency from the analytic signal of the estimated fundamental waveform. Experimental results show that the proposed method outperforms conventional deterministic approaches and provides accurate and robust instantaneous pitch estimation across diverse domains, including speech, singing voice, musical instruments, and degraded speech signals.
arXiv:2606.14331v1 Announce Type: new Abstract: At the intersection of rising wealth inequality and intensifying environmental pressures, we investigate a reverse causal relationship that has received comparatively little attention: wealth inequality may not only be a consequence of environmental crises, but also act as a structural obstacle to the ecological transition itself. We develop a stylized agent-based model in which heterogeneous agents, whose initial wealth follows a Pareto distribution, allocate their income between either a Brown or a Green sector through a utility function. The function is designed to capture the trade-off between short-term returns and exposure to long-term systemic risks. A central ingredient is that wealthier agents perceive themselves as less vulnerable to environmental shocks, thereby reducing the amount of resources available for the transition. We show that, beyond inequality thresholds compatible with those observed in most developed countries, the economy remains locked in a Brown regime, even when a substantial share of agents is sensitive to externalities. We then assess a set of stylized fiscal policies (basic income, carbon taxation, Green incentives, and a combined scheme) and find that their effectiveness depends strongly on the inequality regime and on the regressivity embedded in the fiscal mechanism, revealing multidimensional trade-offs between transition speed, cumulative environmental destruction, growth, and fiscal pressure.
arXiv:2606.14343v1 Announce Type: new Abstract: Chiroptical sensing is central to gain fundamental insight into electronic, vibrational and rotational degrees of freedom of chiral molecules, and is a cornerstone for nanomedicine and drug discovery platforms. Current chiral sensing technologies to assess the enantiomeric imbalance of chiral pharmaceutical compounds are sensitive to ml volumes but are time-consuming and cannot be integrated on a chip, thus creating a major bottleneck for drug discovery and nanomedicine. Here, we propose a novel chiroptical sensing approach based on optical rectification in a photonic micro-cavity filled by a drug solution with nl volume. We theoretically demonstrate that, upon optical excitation by intense pulsed laser light, such a nonlinear effect produces a chirally-sensitive nV voltage burst at the electrically-gated micro-cavity boundaries, with sign depending solely on the drug enantiomeric imbalance. Our results shed light on the potential of optical rectification as a robust platform for innovative lab-on-a-chip devices enabling chiral sensing with nl sensitivity.
arXiv:2606.14394v1 Announce Type: new Abstract: Public blockchains continue to struggle with scalability because improving throughput is not as simple as increasing block size or reducing block interval. Larger blocks increase validation and transmission cost, while shorter intervals raise the likelihood of propagation delays, forks, and stale blocks. These limits motivate sharding, where transaction processing is divided across multiple parallel shard groups. In this work, we present a configurable SimPy-based discrete-event simulator for evaluating sharded blockchain architectures under controlled workload and network assumptions. The simulator models mining, verification, inter-shard coordination, block dissemination, measured throughput, average block time, and communication overhead. Our simulator achieves 1.6M TPS at 256 shards under a local datacenter-like setup and 0.6M TPS in a global WAN setup, showing strong throughput gains from parallel execution. However, the gains are not unbounded: beyond a certain number of shards, coordination traffic, synchronization, and network overhead begin to dominate, leading to diminishing returns.
arXiv:2606.14401v1 Announce Type: new Abstract: This article studies the stability of feedback interconnections of linear time-invariant systems based on frequency-dependent passivity indices. Using these frequency-dependent passivity indices, we show that the feedback interconnection of two systems can be certified to be stable even if both systems have a lack of passivity in terms of their scalar passivity indices. The main contribution of this paper is a new stability theorem based on frequency-dependent passivity indices. Moreover, we discuss the connection of the proposed feedback stability theorem to prior results based on scalar passivity indices. A numerical case study showcases the advantages of frequency-dependent passivity indices over scalar indices for feedback interconnections of linear systems.
arXiv:2606.14406v1 Announce Type: new Abstract: Clinical decision-making is augmented by decision-support systems (DSSs). To counter overreliance on DSSs, several methods have been proposed that create friction in order to promote cognitive engagement and reflection. In this paper, we investigate how two such forms of friction, namely data-driven questions and `what-if' analysis, are perceived by medical experts. For a real-world decision task, we replicated a DSS used in clinical practice and gathered clinicians' feedback on a prototype through in-situ interviews (n=7). Our findings suggest that while the questions were perceived as unhelpful for reflective thinking, they could serve as reminders to consider relevant information. Furthermore, inspecting `what-if' hypotheticals was found useful for potentially improving patient care. Clinicians saw our prototype as a promising training tool for novice clinicians. From the clinicians' feedback, we make recommendations for designing friction in work practices. Our work contributes to human-AI interaction research, which aims to encourage reflection to mitigate AI overreliance.
arXiv:2606.14411v1 Announce Type: new Abstract: We design and evaluate Fabula, an interactive app for fiction writers. Fabula uses detailed narrative plans informed by general narratological theory. Stories are structured hierarchically into scenes and beats that can be (re)generated and revised at script and story plan level. Using participatory AI, we critically evaluate and improve Fabula with casual and published writers, via design interviews and writing sessions with 42 experts, and large-scale internal and external testing. We interrogate our design choices: (1) whether a language model-based auto-evaluator, optimized on human experts' preferences, can improve story quality, (2) whether users want UI that exposes the detailed narrative plan alongside the story script, (3) to what extent our narratology assumptions fit localised storytelling traditions and serve screenwriters or playwrights, and (4) whether convergent iteration over the story plan supports writers' creativity. Building on critical feedback and concerns, we use Fabula as a cultural probe in adversarial design, and identify potentials for writing feedback and for interactive storytelling.
arXiv:2606.14195v1 Announce Type: new Abstract: Kalman filters based on the Embedded Latent Transfer Operators (ELTO) emerge as novel statistical tools for sequential state estimation. However, a critical limitation stems from their use of simplified noise models, which fail to dynamically adapt to non-stationary processes. To address this limitation, we introduce an ELTO-based Bayesian filtering approach with a new structured parameterization for the filter's noise model. This parameterization enables structured noise adaptation, which couples the data-driven learning of an optimal time-invariant noise model with dynamic parameter adaptation that responds to changes in dynamics within non-stationary processes. Empirical results show that our structured noise adaptation improves the filter's dynamic state estimation performance in noisy, time-varying environments.
arXiv:2606.14427v1 Announce Type: new Abstract: Most TinyML hardware accelerators focus on supporting Quantized Neural Networks (QNNs) to meet stringent constraints on power consumption and size. Despite this, the security aspects of quantization within TinyML hardware remain largely unexplored. Although previous studies indicate that QNNs demonstrate similar or enhanced robustness when compared to full-precision Deep Neural Networks (DNNs) against typical evasion attacks, no attack strategies tailored specifically for TinyML hardware have been proposed yet. This paper addresses this shortfall by demonstrating how a two-step attack pipeline can surpass the current state-of-the-art in the QNN context and shows the need for more hardware-aware security research.
arXiv:2606.14428v1 Announce Type: new Abstract: Potential functions are a key tool in theoretical computer science with applications ranging from the runtime analysis of algorithms and data structures, through the analysis of the expected behavior of random processes and search heuristics, to proving the existence of equilibrium states in strategic games. Typically, proofs that employ potential functions are short, elegant, and easy to verify, yet very powerful. Moreover, potential functions are essential ingredients for constructive proofs, in particular in algorithmic game theory. There, a key question is the existence of equilibrium states, but the most powerful theorem in the field -- Nash's theorem -- is unfortunately non-constructive. For many strategic games, potential functions come to the rescue by enabling constructive proofs that sometimes even yield efficient algorithms for finding equilibria. We add to this by providing a novel class of entropy-inspired log-multinomial potential functions for natural game-theoretic settings where rational agents of different types strategically choose actions to maximize their utility. In particular, we consider utility functions that are based on the fraction of same- and other-type agents taking the same action. We demonstrate the versatility of the new potential function class by presenting simple equilibrium existence proofs for two recent game-theoretic models, for which only involved technical proofs were previously known. Even better, the new potential function class yields efficient algorithms for constructing equilibria for much more general models. Thereby, we positively resolve several open problems.
arXiv:2606.14430v1 Announce Type: new Abstract: Partial differential equations (PDEs) regulate the behaviour of countless spatiotemporal systems in the physical and life sciences. In many cases, they encode the coupling between the system's degrees of freedom, leading to nonlinear equations whose solution space is challenging to explore exhaustively. Systematic approaches to PDE model exploration are a holy grail of computational science. In this article, we formulate a criterion for increasing the diversity of a search campaign, based on the PDE residual behaviour under solution deformation. We develop a practical formalism to compute this property and illustrate its role in a few cases of interest.
arXiv:2606.14434v1 Announce Type: new Abstract: A nonlinear orbital station-keeping solution for the circular and elliptic versions of the Earth-Moon Restricted Three-Body Problem (R3BP) is developed via a backstepping technique. Formal guarantees for global asymptotic stability (GAS) are attained, as shown through Lyapunov's stability theory. The adequacy of the proposed control law is evaluated through the means of numerical trials over closed periodic solutions of the circular and elliptic R3BPs. The ramifications of the control gain choice are carefully studied and simulated.
arXiv:2606.14448v1 Announce Type: new Abstract: Cooperative massive multiple-input multiple-output (MIMO) promises large gains over cellular deployments, but existing comparisons of different architectures often mix antenna distribution, inter-site coordination, and processing assumptions. This paper introduces a graph-based framework for fair comparison of cellular, coordinated, and cell-free massive-MIMO systems. We differentiate between two key properties, namely antenna distribution and inter-site cooperation, which yields seven representative system types. We derive compatible uplink and downlink spectral efficiency (SE) expressions, including an uplink bound for detectors with mixed instantaneous and statistical effective channel state information (CSI), and adapt scalable user association and processing rules to all considered architectures. We evaluate these systems using extensive numerical simulations and show that for a fair comparison much larger simulation areas (at least 2.5 $\times$ 2.5 km2) than commonly used are required. We introduce the relative capacity, which measures how closely each architecture approaches centralized cell-free processing. The results show that coordinated, phase-aligned beamforming across spatially distributed antennas is the main source of cooperation gains. In dense deployments with few antennas per access point (AP), coordinated Distributed Antenna System (DAS) and hybrid cell-free architectures achieve much of the centralized cell-free performance while requiring substantially weaker midhaul assumptions.
arXiv:2606.14453v1 Announce Type: new Abstract: Independent Redistricting Commissions (IRCs) are a promising tool for bottom-up redistricting, but their public testimony processes are vulnerable to adversarial manipulation. We propose using differential privacy to draw redistricting plans that incorporate community of interest (COI) testimonies while remaining robust to adversarial input. Treating individual testimonies as data points, we use the marked edge walk to sample from differentially private distributions of redistricting plans via the exponential mechanism. We introduce two score functions and demonstrate that both can be targeted by MEW across a range of privacy budgets. Applying this method to Missouri's mid-cycle redistricting using 808 COI testimonies, we show that COI-informed sampling outperforms an uninformed baseline and the enacted plan. An adversarial experiment demonstrates that the method can be robust to attacks under certain privacy budgets and may perform better in practice than formal group privacy guarantees imply. We also find that stronger COI preservation tends to spread minority and Democratic representation more evenly across districts.
arXiv:2606.14459v1 Announce Type: new Abstract: Modern Automatic Speech Recognition (ASR) systems have made remarkable progress on standard benchmarks, yet performance gaps have emerged under real-world distribution shifts, caused by recording conditions, accents, speech impairments, and noise. Existing datasets and benchmarks typically isolate these factors, which overlooks their co-occurrence in real-world applications. In this paper, we argue that model robustness can be treated as a dynamic capability that continually develops, and we introduce MoDiCoL, a Modular Diagnostic Continual Learning dataset designed for controlled analysis of linguistic content, speaker characteristics, and acoustic environments. Furthermore, we propose a real-world-inspired continual learning curriculum to simulate incremental updates and study how robustness is acquired, transferred, and forgotten. We evaluate three continual learning strategies and provide detailed insights into robustness under evolving conditions.
arXiv:2606.14460v1 Announce Type: new Abstract: Transformer-based clinical language models are increasingly integrated into high-stakes clinical decision support pipelines, yet the computational mechanisms through which demographic associations encoded in medical documentation propagate into model probability distributions remain empirically underspecified. We present a systematic computational audit of representational bias in ClinicalBERT (Alsentzer et al., 2019), a BERT-based model pretrained on MIMIC-III discharge summaries, employing two complementary probing methodologies: Log Probability Bias Analysis (LPBA), which quantifies demographic descriptor-induced shifts in masked token probability distributions across behavioral and evaluative semantic categories, and Masked Language Model-based analysis (MLM), which probes internal representational structure for demographic agency attribution encoding across 98 real clinical sentence templates and eight intersectional race-gender combinations. Corpus frequency analysis operationalizes the distinction between statistical disparity and bias amplification by benchmarking model outputs against empirical term frequencies in the MIMIC-III training corpus. Of 32 statistically significant findings, 65.6% contradict observed corpus distributions, rising to 80% for Black patients and 87.5% for agency attribution under MLM probing, providing direct empirical evidence that representational bias in ClinicalBERT operates predominantly through model-internal amplification rather than training data inheritance. Keywords: natural language processing, clinical documentation, algorithmic auditing, representational bias, health equity 1
arXiv:2606.14463v1 Announce Type: new Abstract: Neurosymbolic (NeSy) models integrate neural networks and symbolic reasoning for robust and interpretable AI. State-of-the-art NeSy models require that the symbolic component is expressed in a differentiable way, often complicating the use of approximate inference. We propose EM-NeSy which casts probabilistic NeSy learning as an instance of the Expectation-Maximization (EM) algorithm. In the expectation step, we compute the posterior over the neurally predicted symbols conditioned on the label via probabilistic inference. In the maximization step, we update the neural parameters based on this posterior using gradient descent only through the neural component. This formulation unlocks the full potential of the EM algorithm for NeSy learning. It allows NeSy to extend naturally to approximate reasoning without any additional modifications or differentiability requirements of the symbolic component. Furthermore, it recovers the standard end-to-end gradient-based NeSy setting under exact inference. Our experimental results demonstrate the scalability and computational efficiency of EM-NeSy.
arXiv:2606.14471v1 Announce Type: new Abstract: Feedback optimization is a control approach for driving a dynamical system to the solution of an optimization problem by interconnecting the plant with an algorithm. Existing stability guarantees typically rely on timescale separation, enforced by conservative gain bounds that limit transient performance and require a pre-stabilized plant. This paper revisits the robust control perspective on feedback optimization. We formulate the plant-optimizer interconnection as a generalized plant, where the cost gradients are characterized by Zames--Falb Integral Quadratic Constraints. Classical timescale-separation bounds are recovered as a special case of static multipliers, with dynamic multipliers yielding substantially tighter stability margins. The formulation also enables IQC based synthesis of dynamic output feedback controllers that jointly stabilize the plant and optimize transient performance, with possible model uncertainty absorbed into an uncertainty channel. For constrained problems, the framework extends to dynamic controllers that generalize projected gradient flows. Numerical examples illustrate the benefits and flexibility of the proposed approach.
arXiv:2606.14480v1 Announce Type: new Abstract: We describe the GPU implementation of the Seeding and Matching algorithms, developed for the first level trigger of the LHCb experiment and key to reconstruct long and very displaced tracks at 40 MHz. The algorithms have been participating in the data taking during the full Run 3 of LHCb with a very high throughput, increasing the physics reach of the experiment. The Seeding is a standalone pattern recognition algorithm aiming at finding charged particle trajectories in the most forward tracker of LHCb. These trajectories are then extrapolated backward by the Matching algorithm which combines them with stubs formed from hits in the first tracker in order to form what we call Long tracks. Hits in the second tracker are then searched for to better define the trajectory and improve the track momentum resolution. This backward approach, complementary to the approach of extrapolating the stubs in the first detector to the forward tracker through the magnetic field, improves the Long track efficiency at low transverse momenta, increasing the potential of key physics decay channels.