arXiv:2607.10310v3 Announce Type: replace
Abstract: Preparing for job interviews is important for securing desired positions, yet realistic practice remains difficult to access: real interviews are infrequent, expert mock coaching is costly, and self-practice offers neither adaptive dialogue nor structured assessment. Existing systems typically address only parts of this need through fixed question sequences, limited communication channels, or feedback with little supporting evidence. We present PolyInterview, an LLM-based platform for immersive mock interview practice with comprehensive multimodal assessment. PolyInterview uses the target job description and CV to generate questions tailored to the role and candidate, conducts multi-turn spoken interviews with a lip-synced digital human interviewer that asks answer-aware follow-up questions, and evaluates response content, vocal delivery, and non-verbal behavior. Four parallel evaluators produce 13 behavior-level features that are aggregated into 10 assessment aspects and two competency tracks. Guided by the KSA and STAR frameworks, the report links each score to behavioral evidence and actionable recommendations. PolyInterview is publicly accessible. Its current all-account snapshot contains 101 accounts, 1,564 interview sessions, 7,665 generated questions, and 1,422 five-stage question sets. Generated questions are more closely aligned with their matched job description than with cross-role job descriptions in 93.7% of sessions. An evaluation by ten experts found strong question plans and actionable feedback.
Science Journals
arXiv:2607.10350v3 Announce Type: replace
Abstract: Recent VLM and VLA systems have improved robotic perception and action prediction, yet long-horizon embodied agents still require a general runtime layer for reasoning, memory, tool use, verification, and cross-embodiment execution. We present ABot-AgentOS, a general robotic Agent Operating System that sits above low-level controllers and provides a deliberative agent layer for scene-conditioned planning, context-isolated skill execution, multi-stage verification, multi-modal memory, and edge-cloud collaboration. To evaluate such systems, we introduce EmbodiedWorldBench, an executable benchmark with 16 indoor, outdoor, and hybrid scenes, four difficulty levels, and over 200 tasks involving navigation, object search, NPC dialogue, dynamic events, and trace-grounded scoring. ABot-AgentOS further introduces Universal Multi-modal Graph Memory, a persistent source-grounded substrate that converts dialogue, visual observations, spatial context, temporal relations, and task traces into typed nodes and edges. A failure-driven self-evolution loop converts diagnosed memory failures into gated runtime evo-assets that are promoted only to later evaluation splits, preventing current-split ground-truth leakage while enabling continual improvement. On an initial EmbodiedWorldBench subset, ABot-AgentOS improves over a single-controller baseline in both task success and goal completion. Across memory benchmarks, ABot-AgentOS Static achieves 87.5 on LoCoMo, 59.9 on OpenEQA EM-EQA, 88.6 on Mem-Gallery, and 76.5 Acc@All on NExT-QA; self-evolution further improves LoCoMo to 88.7, OpenEQA to 60.4, and Mem-Gallery to 89.0. These results suggest that a general Agent OS layer can improve long-horizon embodied execution while providing persistent, auditable memory for continual interaction.
arXiv:2607.10383v3 Announce Type: replace
Abstract: Visual Language Navigation foundation models aim to unify deep reasoning for grounded spatial decisions with broad versatility for diverse embodied tasks. Current approaches typically achieve this integration via monolithic policies that map observations directly to actions, yet they often suffer from coordinate drift and poor handling of long-tail semantics. Furthermore, these black-box mappings lack interpretability, hindering the simultaneous achievement of generality, robustness, and transparency. We present ABot-N1, a step toward a general Visual Language Navigation foundation model, that addresses these challenges by decoupling cognition from control via a slow-fast architecture guided by dual visual-language signals. More specifically, a slow vision-language reasoner performs explicit Chain-of-Thought reasoning while producing a pixel goal. This compact set of image-space anchor points serves as a universal interface for diverse tasks, including point-goal, object-goal, poi-goal, instruction-following, and person-following. Subsequently, a fast action expert leverages both the textual cues and the pixel guidance to generate continuous waypoints at the native control frequency. By bridging high-level intents and low-level control through pixel-grounded anchors paired with explicit linguistic traces, our approach ensures robust, generalizable, and interpretable navigation across simulation and real-world benchmarks. ABot-N1 establishes new state-of-the-art records, delivering massive gains specifically in urban-scale navigation: boosting POI arrival by 35.0% (to 77.3%) and achieving 95.4%/92.9% SR in complex indoor and outdoor scenes. It also maintains superior robustness across object-reaching, person-following, and instruction-following tasks. New Point-Goal/POI-Goal benchmarks are released as open source to advance the field of urban-scale navigation.
arXiv:2607.10622v2 Announce Type: replace
Abstract: Kerr microcombs provide a compact route to broadband optical frequency grids, yet the primary comb states formed at the onset of Kerr-comb generation have received little attention as metrological objects. Here we characterize the coherence and frequency stability of pair-generated primary-comb tones in a silicon nitride microresonator using synchronized multi-channel frequency counting referenced to a hydrogen-maser-stabilized difference-frequency comb, enabling direct measurement of temporal fluctuations and correlations among the pump, signal, and idler tones. We show that the generated tones are strongly constrained by parametric energy conservation: under weakly locked conditions with MHz-level frequency excursions, the residual deviation from $2f_p=f_s+f_i$ remains sub-hertz in the mean, and the signal-idler regression deviates from the ideal $-1$ response by only $2.4\times10^{-9}$. When two of the three tones are tightly phase-locked, the energy-conservation residual of the full pump-signal-idler triad, equivalently the deviation of the measured idler from the value inferred from the locked pump and signal, reaches a fractional-instability floor near $6 \times 10^{-16}$ at $\tau\approx100~\mathrm{s}$. This demonstrates metrological-level preservation of the parametric constraint while revealing subtle mode-dependent noise transfer. Together, these results establish primary Kerr tones as a strongly correlated chip-scale parametric frequency triad suitable for demanding precision-frequency applications.
arXiv:2502.04808v3 Announce Type: replace
Abstract: Plasma spectroscopy is a fundamental tool for diagnosing laboratory and astrophysical plasmas. Accurate interpretation of spectra depends upon precise modeling and comprehension of Stark broadening and other mechanisms affecting spectral lines. In this context, computer simulations have emerged as valuable tools, offering idealized experiments with well-defined conditions. Molecular dynamics simulations, in particular, excel at replicating the particle interactions within the plasma and their impact on the state of a radiating atom or ion. However, these simulations present challenges in tracking electron capture processes, since setting an unambiguous criterion to distinguish between bound and free electrons is not trivial. In this paper we introduce a new algorithm that, within a classical framework, precisely identifies the scenario in which an electron is captured by an ion and then follows a stable orbit around it. The algorithm's applicability extends to emitters with charges Z >= 1. The procedure enables the correct identification of valid time-histories of the electric microfield perturbing the emitting ion, which will be used for subsequent line shape calculations. The ionization balance results obtained from the application of this algorithm are compared with an additional method based on the potential energy of the particles in the simulations and with atomic kinetic simulations. For both methods, we find good agreement, therefore validating the use of this approach.
arXiv:2503.07769v2 Announce Type: replace
Abstract: We study the problem of computing the diameter and the mean distance of a continuous graph, i.e., a connected graph where all points along the edges, instead of only the vertices, must be taken into account. It is known that for continuous graphs with $m$ edges these values can be computed in roughly $O(m^2)$ time. In this paper, we use geometric techniques to obtain subquadratic time algorithms to compute the diameter and the mean distance of a continuous graph for two well-established classes of sparse graphs. We show that the diameter and the mean distance of a continuous graph of treewidth at most $k$ can be computed in $O(n\log^{O(k)} n)$ time, where $n$ is the number of vertices in the graph. We also show that computing the diameter and mean distance of a continuous planar graph with $n$ vertices and $F$ faces takes $O(n F \log n)$ time.
arXiv:2607.15563v1 Announce Type: new
Abstract: Recent visual place recognition (VPR) methods based on vision transformers, particularly foundation models, have achieved remarkable recognition performance. However, these models process all visual tokens throughout the entire network, resulting in substantial computational overhead, which hinders their deployment in real-time and resource-constrained scenarios. A natural question thus arises: are all visual tokens necessary for VPR? To answer this question, we present the first systematic benchmark of token reduction for efficient visual place recognition. Our benchmark comprehensively evaluates representative token pruning, token merging, and hybrid pruning-merging methods across multiple state-of-the-art VPR models and diverse benchmark datasets covering urban, suburban, and natural environments. We further investigate token reduction from multiple perspectives, including recognition performance under different reduction configurations, computational complexity, inference speed, qualitative visualization, and deployment efficiency on edge devices. Through extensive experiments and in-depth analysis, our benchmark reveals multiple important characteristics of token reduction in VPR and provides several practical insights into the trade-offs between accuracy and inference efficiency. For example, token reduction can reduce computational cost by up to 29\% and improve throughput by up to 44\%, while incurring less than 1\% degradation in recognition accuracy. Overall, this work establishes a comprehensive foundation for future research on token-efficient VPR and efficient visual retrieval systems. Our codes and models will be available at https://github.com/Tong-Jin01/TokenReduction4VPR
arXiv:2607.15565v1 Announce Type: new
Abstract: Where should the question go in a vision-language model (VLM) prompt: before the image or after it? Intuition says before: knowing what is asked should tell the model where to look. Yet across visual question answering benchmarks, question-first prompting consistently underperforms the image-first ordering recommended for frontier VLMs, a phenomenon we term the question-first paradox. We trace the paradox to a conflict between two stages of VLM computation. Logit-lens and attention probes show the intuition is half right: a question placed before the image genuinely steers perception, moving image patch representations toward question-relevant concepts. The failure lies downstream. Stranded behind hundreds of image tokens, the question is barely attended by the answer token, which instead commits to image-driven (often wrong) answers; a causal attention knockout confirms that the answer reads the question only when the question follows the image. The diagnosis yields a training-free fix: question echoing, restating the question on both sides of the image so that one copy steers perception while the other is read out at answer time. The same division of labor appears in a fifty-year-old finding on human ``adjunct questions'', where repeating a question before and after a passage aids comprehension more than either position alone. Echoing the image as well brings further gains, restoring the whole-image view a causal decoder otherwise loses. The paradox holds across five open VLMs, costing up to 17.5 group-accuracy points. Echoed prompts close it and surpass the best single-pass ordering on NaturalBench, POPE, Winoground, and open-ended VQAv2, by up to 19 Winoground group-accuracy points, with no training, fine-tuning, or architecture change. The paradox reveals a trade-off between steering perception and preserving question access; echoing resolves it through prompt design alone.
arXiv:2607.15571v1 Announce Type: cross
Abstract: The $\mu$-basis is an algebraic tool originating from the theory of moving curves and moving surfaces, and it is widely used in the study of rational curves and surfaces. In this paper, we give the explicit formulas for the $\mu$-basis of planar quartic rational parametric curves based on redefined vector polynomials, and several illustrative examples are provided. Meanwhile, we also discuss the corresponding cases for quadratic and cubic curves.
arXiv:2607.16137v1 Announce Type: new
Abstract: Computing a minimum-weight perfect matching for a point set $P$ in Euclidean space is a classic geometric optimization problem. We consider the problem in a dynamic setting, where pairs of points may be added to or removed from the set $P$. Our focus is on maintaining an approximately optimal solution without making too many changes to the solution. More precisely, we are interested in $k$-stable algorithms, which change at most $k$ edges in the matching after each update to the set $P$. In other words, we consider an online setting (with insertions and deletions) with bounded recourse.
We study trade-offs between the stability of the algorithm and the approximation ratio of the maintained solution for point sets in $\mathbb{R}^1$. First, we present an $O(\sqrt{n})$-stable algorithm that maintains a $2$-approximation, which we show to be optimal among all algorithms with sublinear stability. Second, we prove that any $o(\log n)$-stable algorithm has unbounded approximation ratio. Our lower bounds hold even in the insertion-only case, while our algorithm works in the fully dynamic case. Moreover, our lower bounds also hold for the bipartite variant of the problem.
arXiv:2607.16138v1 Announce Type: new
Abstract: Improved Kernel Partial Least Squares (IKPLS) algorithms 1 and 2 are among the fastest PLS calibration algorithms. This article focuses on two shared steps, the computation of the $\mathbf{X}$ rotations, $\mathbf{R}$, and the $\mathbf{Y}$ loadings, $\mathbf{Q}$, and accelerates both. For $\mathbf{R}$, term-by-term accumulation is replaced by a direct evaluation strategy that requires the same number of multiplications but parallelizes better on modern hardware. For $\mathbf{Q}$, I identify - to the best of my knowledge, for the first time - equivalences showing that each $\mathbf{Y}$ loading is obtainable, up to explicitly derived constants, from quantities already computed earlier in the same iteration, and I exploit them in IKPLS to reduce the cost of each loading from $\Theta\left(KM\right)$ to $\Theta\left(M\right)$ operations whenever $M = 1$ or $2 \leq M < K$, with $K$ predictor variables (number of columns in $\mathbf{X}$) and $M$ response variables (number of columns in $\mathbf{Y}$). Both improvements provably yield exactly the same $\mathbf{W}$, $\mathbf{P}$, $\mathbf{Q}$, $\mathbf{R}$, and $\mathbf{T}$ as the original algorithms. Benchmarks with NumPy (CPU) and JAX (GPU) show speedups of up to two orders of magnitude for the isolated steps and of approximately $2\times$ (CPU) and $6\times$ (GPU) for entire fits. Both improvements are implemented in the free, open-source Python package \texttt{ikpls}.
arXiv:2502.06818v4 Announce Type: replace
Abstract: Recent works modify CLIP to perform open-vocabulary semantic segmentation in a training-free manner (TF-OVSS). In vanilla CLIP, patch-wise image representations mainly encode homogeneous image-level properties, which hinders the application of CLIP to the dense prediction task. Previous TF-OVSS works sacrifice globality to enhance the locality of CLIP features, by making each patch mainly attend to itself or its neighboring patches within a narrow local window. With their modifications,the ability of CLIP to aggregate global context information is largely weakened. Differently, in this paper, we rethink the global knowledge encoded by CLIP and propose GCLIP to answer how to extract and utilize beneficial global knowledge of CLIP for TF-OVSS. As the representation of each patch is finally determined by the attention weights and the Value embeddings, we propose to reshape the last-block attention and Value embeddings to aggregate useful global context into final features. Firstly, we aim to equip the last-block attention with image-level properties while not introducing homogeneous attention patterns across patches. To realize the goal, we fuse the attention from the global-token emerging blocks with the Query-Query attention. Secondly, we aim to make Value embeddings of the last-block attention module more semantically correlated. To realize this, we design a novel channel suppression strategy.Extensive experiments on five standard benchmarks demonstrate that our method consistently outperforms previous state-of-the-arts.
Holistic Fusion: Task- and Setup-Agnostic Robot Localization and State Estimation with Factor Graphs
arXiv:2504.06479v2 Announce Type: replace
Abstract: Seamless operation of mobile robots in challenging environments requires low-latency local motion estimation and accurate global localization. While most sensor-fusion approaches are designed for specific scenarios, this work introduces a flexible open-source solution for task- and setup-agnostic multimodal sensor fusion distinguished by its generality and usability. Holistic Fusion formulates sensor fusion as a combined estimation problem of i) the local and global robot state and ii) a (theoretically unlimited) number of dynamic variables, including automatic alignment of reference frames; this formulation fits countless real-world applications without conceptual modifications, offering a comprehensive solution beyond hard-coded/task-specific approaches. The proposed factor-graph formulation enables direct fusion of an arbitrary number of absolute, local, and landmark measurements expressed with respect to different frames by explicitly including them as states in the optimization and modeling their evolution as random walks. Moreover, local smoothness and consistency receive particular attention to prevent estimation jumps. Holistic Fusion enables low-latency and smooth online state estimation on typical robot hardware while simultaneously providing low-drift global localization at the IMU measurement rate. The efficacy of this released framework [1] is demonstrated in five real-world scenarios on three robotic platforms with distinct task requirements, highlighting the advantages of fusing multiple absolute measurement types [2]. [1] Code: https://github.com/leggedrobotics/holistic_fusion [2] Project: https://leggedrobotics.github.io/holistic_fusion
arXiv:2607.15910v1 Announce Type: new
Abstract: Accurate positioning of optical components is essential for maintaining beam alignment in free-space optical (FSO) communication systems. This work investigates reinforcement-learning-assisted tuning of cascaded position and velocity PID controllers for an optical deflector that moves the end of an optical fiber in the focal plane of an optical system. A Deep Deterministic Policy Gradient (DDPG) agent adjusts six PID coefficients through interaction with a physical experimental stand. The stand supports target-coordinate updates of up to $12$ kHz, while the agent and the controlled device are located approximately $200$ km apart and exchange data over UDP. After $5000$ training sessions, two fixed coefficient sets are selected and compared with a manually tuned baseline. For a pseudo-random target trajectory, the best RL-tuned set reduces the range of the radial positioning error from $119$ to $82$, corresponding to a $31\%$ reduction, and decreases its standard deviation from $15$ to $12$. For a constant zero target, the RL-tuned sets do not improve the radial error range. The results demonstrate the potential of DDPG for experimental PID tuning in dynamic positioning tasks and indicate the need for multi-regime optimization to achieve consistent performance under different operating conditions.
arXiv:2607.15554v1 Announce Type: new
Abstract: The recent breakthrough work of Chatterjee, Ghosh, Gurjar, Raj and Thierauf [CGGRT26] gives the first deterministic NC algorithm for the bipartite matching problem. They show how to detect as well as find perfect matchings in bipartite graphs in NC. In this note we present an arguably simpler-to-state variation of the NC detection criterion of [CGGRT26], with improved parameters.
arXiv:2607.16149v1 Announce Type: new
Abstract: A central problem in fluid mechanics is the identification of self-similarity, which reveals the underlying scaling behaviour of a flow. Recently, a data-driven framework proposed by Bempedelis et al. (2025 J. Fluid Mech., vol. 1020, A11) has enabled the extraction of single-scale self-similarity from data, without prior knowledge of the governing equations. However, many physical systems are inherently multiscale and therefore require more general approaches. Building on this foundation, we develop an algorithm for the systematic identification of multiscale self-similarity. The algorithm is first applied to two canonical fluid-mechanical problems: turbulent channel flow and late-stage homogeneous decaying turbulence characterised by classical dissipation laws. In both cases, the algorithm successfully identifies inner and outer self-similarity from numerical data and recovers the corresponding similarity expressions. In the intermediate region of both problems, the algorithm identifies similarity expressions that are independent of both the inner and outer scales, providing a novel data-driven route to recovering the corresponding scaling laws: the logarithmic law in turbulent channel flow and the -5/3 law in late-stage homogeneous decaying turbulence. Moreover, the algorithm uncovers corrections to both laws: one linked to the power-law scaling proposed by Barenblatt (1993 J. Fluid Mech., vol. 248, 513--520), and the other yielding an improved approximation of the energy spectrum. The algorithm is then applied to early-stage homogeneous decaying turbulence, characterised by the nonclassical dissipation law. It identifies the same inner similarity as in the late stage, but different outer similarity expressions.
arXiv:2607.15569v1 Announce Type: new
Abstract: While CXL presents a promising hardware substrate for Distributed Shared Memory (DSM), seamlessly scaling multithreaded applications across multiple nodes remains a formidable challenge. Existing CXL-based DSMs fall short: they require manual code modifications to share non-heap data, employ rigid data placement policies that fail under diverse and dynamic workloads, and suffer from severe page-fault processing overheads in sub-microsecond ($\mu\mathrm{s}$) environments.
We present xDSM, a full-space, elastic DSM system built over CXL that transparently scales unmodified multithreaded applications. To eliminate the burden of manual code rewrites, xDSM employs an OS-runtime co-design that establishes a globally coordinated address space, seamlessly sharing all memory segments. To mask CXL access penalties, xDSM abandons static placement rules in favor of a dynamic, latency-driven policy that actively balances data between local DRAM and CXL memory. Finally, to resolve the fundamental tension between high base-page fault overheads and severe huge-page false sharing, xDSM introduces spatial locality-aware elasticity, dynamically coalescing and splitting pages on the fly to amortize processing costs.
Evaluated across diverse workloads using 15 system configurations, xDSM outperforms CXL-only baselines by 1.5$\times$ to 2.2$\times$ and state-of-the-art hybrid DSMs by 1.1$\times$ to 2.2$\times$, while achieving near-linear scalability.
arXiv:2607.15570v1 Announce Type: new
Abstract: We study price-based distributed scheduling of flexible demand in an energy community, where a coordinator broadcasts electricity prices and individual households schedule their consumption. Household demand includes deferrable and non-deferrable loads, such as electric vehicle charging with completion deadlines and thermostatically controlled loads. The coordinator transacts with a distribution utility on behalf of community members under the regulated Net Energy Metering tariff. We formulate distributed demand scheduling as a bilevel stochastic dynamic program. The upper level optimizes the coordinator's pricing policy to minimize the community's energy costs subject to operating, revenue adequacy, and individual rationality constraints. The lower level involves stochastic dynamic programs that maximize households' consumption benefits subject to the availability of renewable generation. The computational cost of such a distributed stochastic dynamic program is prohibitive in general. By uncovering the structure of optimal centralized scheduling, we derive Threshold Pricing Rule (TPR) -- a simple community pricing policy with linear computational costs for the upper- and lower-level optimizations. Being independent of parameters of the underlying stochastic dynamic program, TPR is robust against modeling uncertainties and is shown to guarantee revenue adequacy for the community and individual rationality for community members. As the community size grows, TPR is shown to be asymptotically optimal.
arXiv:2607.15585v1 Announce Type: new
Abstract: Despite the success of diffusion models in Video Frame Interpolation (VFI), existing methods still suffer from two critical limitations. First, latent diffusion inevitably loses fine-grained details when reconstructing images from latent representations back to the pixel space. Second, multi-step sampling incurs prohibitive memory consumption and inference latency. To address these issues, we propose SPEED, a one-step pixel diffusion framework for high-quality VFI. Specifically, SPEED employs a progressive multi-stage architecture with dynamic patch scaling to effectively learn multi-scale motion, structural, and appearance representations. Furthermore, we propose a novel Noise-Update-Only Attention mechanism to prevent semantic degradation of the clean condition frames while reducing the computational overhead by nearly 50%. Besides, we introduce a Drift-aware Timestep Sampling strategy coupled with a tailored training objective to directly predict images in the pixel space, enabling one-step inference without compromising the quality of the generated frames. Extensive experiments show that SPEED achieves state-of-the-art performance. On SNU-FILM, SPEED reduces LPIPS by 8.8% while delivering 63.3% faster inference and 10.6% lower memory usage. On challenging 4K benchmarks, it further surpasses prior methods by up to 51.5% in LPIPS.
arXiv:2607.15856v1 Announce Type: new
Abstract: Naturalistic language comprehension requires listeners to process both local probabilistic expectations and contextual semantic relations. Surprisal has been widely used to quantify local word unexpectedness, but evidence that it robustly predicts fMRI BOLD responses during continuous comprehension has been mixed. This study investigates whether contextual semantic relevance, defined as how strongly an incoming word relates to its recent semantic context, predicts BOLD responses during naturalistic speech comprehension. We analyzed two public fMRI datasets, the Alice dataset and the Moth dataset, treating them as complementary rather than identical replications. Transformed BOLD responses were modeled with generalized additive mixed models (GAMMs) and original continuous BOLD time series were tested with FIR/deconvolution analyses. In Alice, semantic relevance was significant across all 12 ROIs (region of interest), whereas surprisal was not significant after FDR correction. In Moth, semantic relevance showed consistent negative effects across all 30 ROIs, while surprisal showed no comparable pattern. These findings suggest that semantic relevance is a promising BOLD-sensitive metric of contextual semantic fit. More broadly, our findings support the view that slow hemodynamic responses during naturalistic speech comprehension may be especially sensitive to contextual semantic integration, whereas local probabilistic prediction error may be more difficult to detect reliably with fMRI. In this sense, semantic relevance extends computational models of language comprehension from prediction alone toward context-sensitive semantic integration.
arXiv:2607.15432v1 Announce Type: cross
Abstract: Emergency department (ED) boarding occurs when admitted patients remain in the ED while awaiting inpatient beds. Boarding is a major driver of ED crowding and has been associated with poor patient outcomes. We propose a framework to help EDs reduce boarding time and length of stay by using information about current patients and bed availability to proactively request inpatient beds before admission decisions are finalized.
We formulate the problem as a Markov decision process in which predictions of each patient's admission probability and time to disposition are aggregated to guide early inpatient bed requests. This formulation leads to three data-driven policies based on approximate dynamic programming, reinforcement learning, and a newsvendor-type approach. Using a simulation model based on data from a large ED, we evaluate these policies across a wide range of settings. The simulation study shows that proactive aggregate bed requests can reduce average boarding times for admitted patients by 30-70\% and average length of stay for all ED patients by 6-15\%, while creating only modest idle time for prepared inpatient beds. The newsvendor heuristic provides the most attractive tradeoff between ED performance and inpatient bed idle time, whereas the reinforcement learning heuristic produces smoother bed-request patterns when stability in downstream hospital processes is especially important.
Our work shows how EDs can use prediction tools to make proactive bed-request decisions that improve ED operations while helping managers balance reductions in ED delays against inpatient bed idle time. Our findings also illustrate the value of evaluating both simple myopic heuristics and more sophisticated reinforcement learning-based approaches, since each can offer distinct advantages depending on the performance measures and implementation constraints most important to managers.
arXiv:2505.01360v3 Announce Type: replace
Abstract: Plate Tectonics requires strain localization over the entire thickness of the plates. However, modelling strain localization in the deep sections of the plates, which deform by ductile processes, remains a challenge, prompting the use of ad hoc schemes to model plate boundaries. We posit that the bottleneck for self-consistent generation of ductile strain localization in geodynamical models is poor representation of the intrinsic mechanical heterogeneity of rocks, in particular at small scales. This prevents its effects from being accounted for at larger scales, notably emergent properties that arise during upscaling, like anisotropy. To model this heterogeneity and its evolution, we adopt a stochastic description of the rheology, which evolves in time and space as a function of the local work-rate. This approach enables to reproduce the full variety of responses observed in nature, from heterogeneous deformation at the local scale, but homogeneous at the system-scale, with or without softening, to spontaneous development of system-scale shear zones. It enables, thereby, the construction of regime diagrams for ductile strain localization using three adimensional parameters. These parameters describe the degree of heterogeneity and potential for evolution of the rheology, function of (1) the constitutive equation, which represents the active deformation processes, (2) the evolution law for the rock mechanical properties, (3) the initial properties, and (4) the energy input to the system. This approach also enables the prediction of the intensity of localization and the resulting system-scale anisotropic softening, paving the way for self-consistent modelling of plate boundaries in geodynamics.
arXiv:2604.04593v2 Announce Type: replace
Abstract: Retrieval-augmented generation (RAG) grounds large language models in external medical knowledge, yet standard retrievers frequently surface hard negatives that are semantically close to the query but describe clinically distinct conditions. While existing query-expansion methods improve query representation to mitigate ambiguity, they typically focus on enriching target-relevant semantics without an explicit mechanism to selectively suppress specific, clinically plausible hard negatives. This leaves the system prone to retrieving plausible mimics that overshadow the actual diagnosis, particularly when such mimics are dominant within the corpus. We propose Contrastive Hypothesis Retrieval (CHR), a framework inspired by the process of clinical differential diagnosis. CHR generates a target hypothesis $H^+$ for the likely correct answer and a mimic hypothesis $H^-$ for the most plausible incorrect alternative, then scores documents by promoting $H^+$-aligned evidence while penalizing $H^-$-aligned content. Across three medical QA benchmarks and three answer generators, CHR outperforms all five baselines in every configuration, with improvements of up to 10.4 percentage points over the next-best method. On the $n=587$ pooled cases where CHR answers correctly while embedded hypothetical-document query expansion does not, 85.2\% have no shared documents between the top-5 retrieval lists of CHR and of that baseline, consistent with substantive retrieval redirection rather than light re-ranking of the same candidates. By explicitly modeling what to avoid alongside what to find, CHR bridges clinical reasoning with retrieval mechanism design and offers a practical path to reducing hard-negative contamination in medical RAG systems.
arXiv:2607.15472v1 Announce Type: cross
Abstract: Oscillatory dynamics arise ubiquitously in nonlinear systems, yet identifying a physically interpretable phase and phase dynamics in nonlinear, high-dimensional oscillations remains a central unresolved problem. Here we establish the principle of a universal dynamical clock, a physical perspective in which oscillations of arbitrary dimensionality and geometry are equivalently represented as uniform rotation through an equant-induced nonlinear viewing coordinate, inspired by Ptolemy's equant and formalised through an areal-uniformity principle reminiscent of Kepler's second law. Using a machine-learning framework, we demonstrate the existence of such an equant for a broad class of oscillatory dynamics and construct the associated dynamical clock and phase dynamics under additive forces, including noise, periodic perturbations, and coupling. Its value in uncovering new physical rules and phenomena is demonstrated by four findings: (i) collective oscillations in Escherichia coli populations obey a previously unexplained superlinear scaling law, resolving a long-standing open problem posed in 2004; (ii) the response mechanisms of engineered genetic circuits to changes in gene expression and environmental conditions; (iii) a classical-mechanics counterpart of the Berry geometric phase emerges naturally from the phase of the dynamical clock; and (iv) optimal equant non-uniformity provides a geometric early-warning signal for critical transitions and enables prediction of critical parameters. By providing operational and system-agnostic phase dynamics that can be constructed directly from data, the dynamical clock enables principled classification, comparison, and control of oscillatory systems, and offers a new route to understanding how specific dynamical regimes support distinct functional behaviours in networked systems.
arXiv:2505.19909v2 Announce Type: replace
Abstract: Neural Network-based Quantum Monte Carlo (NNQMC), an emerging method for solving many-body quantum systems with high accuracy, has been mainly applied to small systems due to demanding computation requirements. In this work, we introduce a framework based on local pseudopotentials to break through such limitation, improving the computational efficiency and scalability of NNQMC. The incorporation of local pseudopotentials reduces the number of electrons treated in neural network and also achieves better relative energy accuracy than all electron NNQMC calculations for complex systems. This counterintuitive outcome is made possible by the distinctive characteristics inherent to NNQMC. Notably, by avoiding costly integration terms, this approach is also substantially more efficient than its widely used semilocal counterparts. Our approach enables the reliable treatment of large and challenging systems, such as the $\text{Fe}_4 \text{S}_4 (\text{SCH}_3)_4$ iron-sulfur cluster. Overall, our findings demonstrate that the synergy between NNQMC and local pseudopotentials substantially expands the scope of accurate ab initio calculations.