Papers

Science maps describe relationships between papers. Text embedding models build these relationships by turning a paper's text into a list of numbers that gives it a position. Would another model give the same picture? I compare ten models using titles and abstracts of 500,000 publications from the OpenAlex bibliographic catalogue, published in 2000-2024 across 26 disciplines, including Medicine and Physics. Representing each discipline by its papers' average position produces similar arrangements across models, including when other papers are selected. Yet within a discipline and period, two models share, on average, only eight of the 25 papers closest to a given paper. Supplying titles alone also changes much of that list, with effects that vary between models. Models can also disagree about which discipline's papers are more spread out. One measure of spread compares angles between lines drawn from a common point to each paper's position. Moving that point can reverse which discipline appears more spread out, even when selecting other papers leaves the answer unchanged. The broad arrangement can remain similar while answers about nearby papers and differences between disciplines depend on the model, the text it receives and how its output is processed.

Quantum aging transition has recently been reported, whereby the oscillatory dynamics of a network of globally coupled quantum self-sustained oscillators is degraded sharply when the fraction of inactive nodes exceeds a critical value. Here we explore the reverse process: the revival of rhythmic dynamics from an aged state in a quantum oscillator network by introducing a controlled incoherent single-boson pumping channel. Using a self-consistent quantum master equation in the Lindblad form for globally coupled active–inactive quantum Stuart–Landau oscillators, we show that increasing the pumping strength progressively suppresses the aging transition and eventually eliminates it, restoring rhythmicity even when all oscillators are inactive. Remarkably, complete compensation of the coupling-induced dissipative loss is not required; a substantially weaker pumping strength is sufficient to eliminate the aging transition. We show that the revival is governed by a redistribution of the Fock-state population of the inactive oscillators toward higher-occupation states. Our results establish controlled incoherent pumping as an effective mechanism for restoring dynamical robustness in quantum oscillator networks.

We present the first dynamic maximal independent set (MIS) of a graph with $n$ vertices and up to $m$ edges under edge updates in the prediction model, where a predicted update sequence is available for offline preprocessing before the real updates arrive online. We consider the incremental (insertion-only), decremental (deletions-only), and fully dynamic (insertions and deletions) models. Without predictions, deterministic algorithms need worst-case update time of $O(n)$ for decremental and $O(m)$ for incremental and fully dynamic settings. Bernstein et al.~[SODA26] proved under the combinatorial BMM conjecture, combinatorial incremental algorithms against an adaptive adversary require $\Omega(n^{1-o(1)})$ amortized update time. We present deterministic incremental, decremental, and fully dynamic algorithms whose update times depend on the prediction error. This error is measured by the number $\eta_e$ of mispredicted updates, and by the number $\eta_v$ and total degree $\eta_v^*$ of vertices whose MIS membership was mispredicted. With $O(mn)$ preprocessing, our decremental and incremental algorithms achieve worst-case update times $O(\eta_v)$ and $O(\eta_e+\eta_v^*)$, respectively. Our fully dynamic algorithm achieves $O(\eta_e+\eta_v^*)$ worst-case update time with $O(un)$ preprocessing for $u$ updates. We also give space-efficient variants using $O(m)$ space for the decremental and $O(m\sqrt{m})$ for the incremental setting, at the cost of a logarithmic overhead in worst-case update time. Our algorithms satisfy consistency, robustness, and smooth degradation, the required properties of algorithms with predictions. They can also utilize the general error correction framework of Khan et al.~[Arxiv26], relaxing the error measures from first point of difference to prefix matching (for partially dynamic) and set edit distance (for fully dynamic), maintaining the same bounds.

State space models provide efficient linear-complexity processing for long sequences, but their performance frequently degrades as context length grows. We propose SuPER-Mamba, a training-free framework that uses Surprisal as a model-intrinsic signal for long-range state routing in pretrained Mamba models. SuPER-Mamba computes position-wise Surprisal from next-token predictions, aggregates neighboring responses through Local Pooling, and uses the resulting scores to prioritize Global-Channel updates without query-dependent scoring or additional training. We evaluate SuPER-Mamba on three state-space models across the LongBench-E benchmark. The results show that SuPER-Mamba improves the average score from 8.06 to 15.69, corresponding to a 94.7% relative gain over the Base Models and a 13.1% improvement over the state-of-the-art prior baseline. Ablation and length-wise analyses further support the effectiveness of the proposed routing framework.

Natural-language self-consistency and executable programs provide complementary test-time evidence for mathematical reasoning, yet either channel may support the same semantic error. Thus we propose ACED, a reference-anchored framework that treats the self-consistency answer as the default and executable outputs as supporting evidence rather than interchangeable votes. ACED converts source-specific answer counts into smoothed supports, combines them through a weighted logarithmic pool, and overrides the reference only when a calibrated relative-evidence margin is exceeded. The same rule supports both candidate selection and admission of executable-only answers. Across five math word-problem benchmarks, ACED matches or outperforms the corresponding baselines. On GSM8K, ACED-CANDIDATE recovers 89.7% of the attainable candidate-set gain with only three damaging overrides, while ACED-ALL reaches 86.9% accuracy versus 82.6% for direct Program-of-Thought aggregation. These results show that preserving evidence provenance enables more reliable integration of natural-language and executable reasoning.

Recent Vision-Language-Action (VLA) models increasingly use natural-language instructions to condition autonomous driving behavior. Yet existing closed-loop benchmarks primarily measure route completion, safety, and rule compliance, leaving a fundamental question unresolved: does language actually determine which feasible behavior the vehicle executes? A trajectory may be safe and successful while still violating or even ignoring the passenger's instruction. To address this, we introduce \ltd, a benchmark designed to directly test this capability. Its key evaluation unit, the paired counterfactual unit (PCU), holds the initial driving scene fixed while changing only the instruction. This design enables us to distinguish isolated instruction success from counterfactual paired success across four instruction-following scopes. To test whether this distinction persists across evaluation regimes, we instantiate Language2Drive in complementary closed-loop CARLA simulation, reconstructed real-world scenes, and open-loop driving benchmarks, including our released CARLA-Language2Drive simulator with purpose-built maps. Evaluating representative driving VLAs across these settings reveals a striking mismatch between strong driving competence and weak language control: 93.2% of successful rollouts do not have a successful counterpart under the alternative feasible instruction. This failure is dominated by one-sided completion, where a model repeatedly executes one preferred behavior but fails when the same scene requires the counterfactual alternative. These results expose two fundamental limitations of current evaluation: isolated success and driving scores, which are widely used by conventional driving benchmarks, can substantially overstate language-following capability, while current driving VLAs themselves lack reliable instruction following under matched counterfactual tests.

Using bibliometric data from publication records, we investigate the institutional broadening of Japanese astronomy and astrophysics over 1996-2025. The number of Japan-affiliated publications increased from 588 to 2,156, and the number of unique participating Japanese institutions increased from 123 to 227, indicating a nominal broadening of the institutional base. However, institution-publication links remained concentrated among leading institutions, and the effective number of institutions, a frequency-weighted diversity measure, increased more gradually than the number of unique participating institutions. These results show that Japanese astronomy and astrophysics broadened institutionally, but not through a complete equalization of institutional participation. The observed change is best interpreted as an expansion of the participating institutional base around a persistent core of leading institutions.

Memory-based long-video generation aims to combine user control with long-range visual consistency. We organize this process into three functional components: a planner, an adapter, and a generator. Representative multi-shot systems couple memory to generator-specific interfaces or use predefined agent workflows, rather than independently training a memory policy for cross-shot objectives. We introduce REEL, a general-purpose visual-memory planner. Through a unified tool interface, REEL learns to retrieve references, create missing visual evidence, and update persistent memory. Backend-specific adapters translate its evidence plans into supported conditioning inputs, allowing the same policy to guide different black-box video generators. To train REEL, we construct two high-quality datasets, REEL-SFT-4k and REEL-RL-1k, covering diverse interaction patterns and prompt types. Training proceeds in three stages: supervised fine-tuning, single-shot Group Relative Policy Optimization (GRPO), and multi-shot GRPO. Multi-shot GRPO carries policy-generated memory across consecutive shots, so later outcomes guide earlier reference selection and memory updates. REEL-Bench tests identity preservation and intended state changes under text-only requests and timed reference-image uploads. On JoyAI-Echo~1.5 and MiniMax-H3, REEL improves cross-shot consistency on ST-Bench by 15.1% and 20.7%, and Core-100 on our proposed REEL-Bench by 2.41 and 8.32 points, respectively, over the no-memory baseline.

Non-exchangeability and radial asymmetry have got substantial attention in the literature, but relation between the two has not been studied yet. We propose the question: if a copula has a given degree of radial asymmetry, what is the range of possible values of the measure of non-exhangeability. We determine the set of all points whose first coordinate is the radial asymmetry of a copula and the second coordinate is its non-exchangeability. We show that this set is a non-convex hexagon.

Accurate long-term air quality forecasting is challenging due to long-range transport, meteorological variability, and nonlinear chemical reactions underlying secondary pollutant formation. Purely data-driven models often suffer from long-horizon drift and lack physical consistency, while existing physics-informed approaches commonly follow open-loop designs that fuse neural and physical components only at the output level. We propose AeroChem, a closed-loop physics-informed state space framework for chemically reactive air quality forecasting. AeroChem adopts a two-branch co-correction design: a UDE-based process branch provides a structured but imperfect physical prior, while a Mamba-based virtual measurement branch captures long-range temporal context and produces probabilistic virtual observations. Neural Kalman Fusion couples the two branches by correcting the latent physical state at each forecasting step and feeding the posterior state back into the simulator, reducing long-horizon drift. We further introduce HanoiAir, a real-world dataset integrating pollutant concentrations, meteorology, and emission inventories. Experiments on HanoiAir and a public benchmark show that AeroChem improves long-horizon stability and sudden-change forecasting.

Halogen bonding in which a covalently bound fluorine acts as halogen-bond donor (F-bonding) is rare, and its open-shell variant with radical electron donors remains unexplored. We introduce SEFB-15 dataset (15 open-shell DF$\cdots$CH$_3$ complexes) exemplifying single-electron F-bonding (SEFB) and CSFB-14 dataset (14 closed-shell DF$\cdots$NH$_3$ complexes) representing closed-shell F-bonding (CSFB). Reference binding energies (BEs) at DF-RMP2-F12/CBS//DF-RMP2-F12/VDZ-F12 are validated against CCSD(T), CCSD(T)-F12b, W1-F12 and W2-F12 for selected systems (DF = F$_2$, NCF, CNF, NH$_3$F$^+$ and NH$_2$F$_2^+$), with MA%D of 4.5% (DF$\cdots$CH$_3$) and 3.4% (DF$\cdots$NH$_3$) against CCSD(T)-F12b/CBS//CCSD(T)-F12b/VDZ-F12. BEs from W1-F12 and W2-F12 are identical. Reference F-bond distances at DF-RMP2-F12/VDZ-F12 are validated against CCSD(T)-F12b, CCSD(T), CCSD, and DCSD for the same systems, with MA%D of 0.9% (SEFB) and 0.4% (CSFB) against CCSD(T)-F12b/VDZ-F12. M06-2X is recommended for closed-shell halogen bonding with heavier halogens but has never been assessed for open-shell F-bonding. M06-2X/AVTZ systematically overbinds all SEFB-15 complexes and underestimates their F-bond distances, with MA%Ds of 75.8% on BEs and 8.9% on F-bond distances. M06-2X predicts comparable BEs for F$_2\cdots$CH$_3$ and F$_2\cdots$NH$_3$, contrary to all ab initio results, an anomaly persisting across all 16 basis sets. Neither basis set nor BSSE correction resolves the overbinding. BSSE correction worsens BEs for DF$\cdots$NH$_3$ complexes. For CSFB-14, M06-2X/AVTZ yields MA%Ds of 16.0% for BEs and 3.3% for F-bond distances. M06-2X is reliable for CSFB but unreliable for SEFB regardless of basis set. Companion paper II reports performance of other Minnesota functionals and ab initio methods for these datasets.

Adaptive modulation and coding (AMC) improves the spectral efficiency of underwater acoustic (UWA) links, but its feedback is inevitably stale: sound propagates slowly, and by the time channel state information (CSI) arrives, the fast time-varying multipath channel has moved on. Predictive link adaptation therefore needs more than a point forecast: it needs a distribution over each candidate action's consequences. We propose PRB-WM, a physics-regularized Beta world model for predictive per-subcarrier link adaptation on real time-varying UWA channels. A persistence-anchored stochastic latent-space dynamics model learns how the channel evolves; a counterfactual outcome head evaluates, for every candidate modulation and coding scheme (MCS), a Beta-binomial distribution of bit errors on shared imagined futures; and a chance-constrained receding-horizon model predictive controller maximizes expected spectral efficiency subject to a bounded probability of violating the bit error rate (BER) target. To the best of our knowledge, PRB-WM is the first world model for UWA communications. In closed-loop replay of six measured sea-trial links from the public UWA Channel Library against a 24-method pool spanning the full methodological spectrum, PRB-WM attains a mean goodput 87% above the strongest baseline, ranks first on all six time-varying links at the longest lookahead, and is the only method whose realized frame risk decreases systematically with lookahead, meeting the BER-violation contract where no baseline does. Ablations confirm that rollout sampling dispersion is essential to the chance constraint and that the physics anchor is necessary for trainability.

Existing human-centred XAI research predominantly evaluates explanations after they have been generated. However, this fails to provide actionable insights into improving accessibility. This study incorporates human-based feature understandability into feature selection to improve the generation of human-centred explanations. Natural language explanations for machine learning models trained on tabular data often cite influential features in the input data to explain a given decision. As a consequence, users' ability to understand the produced explanation hinges on their understanding of the cited features. To quantify this natural-language feature understandability, Rossberg et al. proposed the Feature Understandability Scale, which assigns numerical `understandability scores' to features based on user feedback. To evaluate the degree to which these scores can improve machine learning explanation quality, this study collects understandability scores across two datasets, proposes an understandability-aware feature selection algorithm to select features that are both understandable and informative, and trains a random forest based on these feature subsets. The SHAP-based explanations generated by this model are evaluated using user feedback, and changes in explanation quality are assessed. The individual evaluation of the explanations showed that while the understandability-based explanations scored higher on average, the difference was not statistically significant. In contrast, the comparative evaluation found that participants preferred these explanations significantly more frequently than expected at the chance level. It is concluded that the integration of understandability can yield more understandable explanations at minimal cost to classification performance, although changes to quality require elucidation in larger cohorts.

This article aims to introduce the bosonic asymmetric quantum conditional mutual information (QCMI), $I_{B}(A; B|C)$, as a rigorously defined information-theoretic quantity for continuous-variable (bosonic) systems. After defining $I_{B}(A; B|C)$ through quantum instruments on infinite-dimensional Fock spaces, we prove that Bose-Einstein condensation (BEC) induces a logarithmic divergence $ I_{B}(A; B|C) \geq \alpha \log_2 (N_{0}^C+1)$, which further establishes $I_{B}(A; B|C)$ as a quantum-information order parameter for BEC with power-law scaling near the critical point. We provide the complete five-step proof of this result, connecting measurement back-action, off-diagonal long-range order, and Jensen's inequality. In the condensed phase, the effective causal propagation velocity is set by the Bogoliubov sound speed $c_s$ in the long-wavelength limit, and vanishes continuously at the superfluid–Mott-insulator transition. Applications to composite bosons, the BEC–BCS crossover, and a quantum-gas-microscope measurement protocol are presented. Finally, we demonstrate how the reduced bosonic QCMI serves as a finite-size scaling probe analogous to the Binder cumulant.

Post-route FPGA timing closure remains expert-dependent and computationally expensive. We present LATCH (LLM-Assisted Timing Closure Harness), a two-layer framework that separates large language model (LLM) orchestration from deterministic FPGA implementation and evaluation. Across sixteen designs and seven LLMs, a single-iteration study compares three agent control granularities - expert-written recipes, direct EDA operations and LLM-constructed procedures - to ask how much control an agent needs per attempt. Functional-pass rates are 98.2%, 71.4% and 17.9%, respectively; yet direct operations give slightly larger 3.4% median Fmax gains when passing, versus 3.2% for recipes. This exposes a trade-off between execution reliability and timing improvement. Informed by this trade-off, a multi-iteration study fixes the action tier at expert recipes to ask how repeated attempts are best organised, comparing a fitted deterministic heuristic with sequential LLM reasoning, branching exploration and knowledge base-guided exploration within 12 checkpoints per run. Branching raises mean maximum-frequency improvement from 10.5% (sequential) to 16.0%, and the knowledge base to 20.3%, improving 15 of 16 designs. Characterisation across 2,421 recipe attempts profiled their gain and validation rates, showing hard-block re-placement gives the largest cumulative mean gain at 41.4MHz.

Human motion prediction during human-robot collaboration is critical for achieving safe and efficient interactions in shared environments. Unlike previous studies that have primarily focused on humans and objects, we focus on the robot-aware human motion prediction task, which explicitly models the influence of robots on human motion. This task presents new challenges, such as capturing human-robot heterogeneity and modeling complex spatial interactions. To address these issues, we develop an adaptive interaction network (AINet) model that consists of two branches: a primary branch that predicts future human motion and an auxiliary branch that estimates robot trajectories. The two branches are jointly optimized and coupled via a Local-Global Spatial Interaction (LGSI) Module, which effectively captures fine-grained and global contextual dependencies between human and robot motion sequences. Then, we introduce an Adaptive Weighted Aggregation (AWA) Module to dynamically fuse motion features using context-dependent weights, thereby increasing adaptability across diverse scenarios. Furthermore, we adopt a coarse-to-fine prediction strategy, in which a coarse pseudo-trajectory is first predicted, followed by the refinement of detailed human poses based on that trajectory. Extensive experiments on three challenging datasets reveal that our proposed method achieves superior performance, validating its effectiveness in human-robot collaboration scenarios. Our code is available at: https://github.com/LyTingHub/AINet.

Social media heavily influences health decisions, yet few opinion dynamics models account for how information is delivered. This work investigates how signaling source expertise shapes trust and long-term opinion formation. Through an online experiment, we find that adding a professional title to the sender's display name significantly increases a post's perceived credibility. Incorporating these findings to create a data-driven opinion dynamics model, we prove that (1) connecting users through some paths to an expert guarantees consensus with experts' opinion, while (2) higher rates of expertise signaling accelerate reaching consensus. Simulations show that (3) convincing even 20% of users to signal their title reduces consensus time to nearly one-third in Erdos-Renyi networks and (4) higher title adoption brings average opinion much closer to that of experts in scale-free and real-world networks. Altogether, our work highlights expertise signaling as a viable intervention to promote validated medical information online.

With the advancement in language modeling, research on adversarial images has recently shifted from image classifiers to vision-language models (VLMs). Despite the new, text-prompt-driven nature of VLMs, we find that existing VLM adversaries have largely overlooked the dynamics of text prompts. In a typical use case, adversarial images can protect visual privacy by rejecting VLMs from answering questions about the image. However, existing protections are prompt-agnostic, i.e., indiscriminately rejecting any question, including those for potential legitimate use of the image. To address this limitation, we propose Prompt-Adapted (ProAda) adversarial images, which are optimized to balance the rejection of harmful prompts and the correct response of legitimate prompts. In particular, to achieve a more accurate adversarial direction, the text prompt is also iteratively updated with a simple classifier judging whether the harmfulness of the prompt remains. Extensive experiments on visual-privacy-related datasets validate that ProAda can reject over 90% of harmful prompts while allowing over 90% of legitimate ones, demonstrating superior overall performance compared to existing prompt-agnostic VLM adversaries. In addition, our ProAda achieves competitive overall cross-model transferability and improved robustness against countermeasures.

Multi-agent systems (MAS) combine multiple agents to solve complex tasks, and recent MAS scale to more than a thousand agents. At this scale, decentralized MAS remove the central orchestrator that assigns sub-tasks, and the agents, which we call workers, self-organize through their own choices of sub-tasks and collaborators on a shared board. As workers keep adding participation and collaboration records, however, the information explodes: workers spend more and more effort on reading the records of other workers, and self-organization loses efficiency. In the statistical view, we represent the execution history of an MAS by two continually updated networks, a collaboration network among workers and a participation network between workers and task types. We then propose AgentCue, a method built on a new latent space model that places workers and task types in a shared latent space. Each time the networks are updated, the estimated parameters of the latent space model form a compact and interpretable statistical summary, and workers use the statistical summary to choose sub-tasks and collaborators without repeated searches of the shared board. We establish the identifiability of the latent space model and non-asymptotic error bounds for the estimates under network increments, and we show that a fixed number of gradient updates after each network update suffices for the same error rate. Experiments on LongBench v2 Multi-Doc QA show that AgentCue achieves competitive effectiveness, efficiency, and interpretability.

Editing the color of images in a realistic way finds many applications such as changing the perception of an image, data augmentation or film post processing. The design of an automatic tool is a complex and long addressed challenge. In particular, two properties are difficult to meet altogether: generating realistic results without artifacts and the possibility to precisely choose the future color. Conventional methods using segmentation and histogram matching maximize the controllability but also introduce a lack of realism or are complex to automate. On the contrary, GANs that specialize in realism are difficult to control. To overcome these challenges, we propose a novel GAN architecture leveraging any differentiable segmentation model. We demonstrate the genericness of our framework that presents state of the art results on different use cases. It generates images that look realistic while offering a precise color control.

We give a higher unit-torus lift from q-linearized permutation polynomials over Fqn to generally nonlinear sparse permutation polynomials over Fqns. The construction preserves compositional inverses: if a linearized polynomials and its inverse have coefficients (ai) and (Ai), then the lifted polynomial and its inverse have the same coefficients. As a consequence, we obtain nonlinear binomial lifts whose inverses are again binomials in the two-dimensional case, including an explicit inverse for a binomial family over Fq4. The three dimensional specialization yields permutation trinomials whose compositional inverses are again trinomials, together with explicit involution criteria and examples.

Policy representations in reinforcement learning have expanded beyond Gaussian policies to diffusion and flow-matching based generative policies. However, existing RL methods remain representation-specific, lacking a unified optimization interface for emerging generative policies. In particular, extending likelihood-based objectives to multi-step generative policies couples optimization to their internal generation process, where accumulated likelihood ratios can weaken the update signal. Meanwhile, generative-policy-specific methods often rely on multi-candidate generation and critic-based selection, improving overall return while providing limited improvement to the policy itself. We introduce DR-APO (Decoupled-Representation Advantage-Guided Policy Optimization), which decouples policy improvement from policy representation through advantage-guided target improvement and representation-compatible direct action fitting, providing a shared optimization principle for Gaussian, diffusion, and flow policies. Across five MuJoCo-v5 tasks, DR-APO remains competitive with SAC for Gaussian policies and achieves the best aggregate policy-only performance for both diffusion and flow-matching policies. NavDP post-training improves the Success Rate (SR) from 70.33% to 89.55% in simulation and from 25% to 65% in real-world navigation. These results demonstrate a unified policy-improvement principle across Gaussian, diffusion, and flow policies, from continuous-control learning to embodied-policy post-training.

Tests of lepton-flavour universality in semileptonic \(B\)-meson decays provide sensitive probes of physics beyond the Standard Model. The Belle and Belle II experiments, respectively operating on the KEKB and SuperKEKB colliders in Japan, have collected \(e^+e^-\to B\bar{B}\) collisions at a center-of-mass energy corresponding to the \(\Upsilon(4S)\) resonance in samples with a combined integrated luminosity exceeding \(1\,\mathrm{ab}^{-1}\). These data, with low particle multiplicity, constrained initial-state kinematics, and excellent lepton identification, provide an ideal environment to study lepton-flavour universality in semileptonic decays of \(B\) mesons. New results on the ratios of semitauonic \(B\)-decay rates compared to analogous light-lepton rates are presented. These include new Belle II measurements of branching-fraction ratios \(R(D^{(*)})\) involving exclusive \(B\to D^{(*)}\tau\nu\) and \(B\to D^{(*)}\ell\nu\) decays using semileptonic or hadronic reconstruction of the accompanying \(B\)-meson tag. Belle II and Belle measurements of angular observables in \(B\to D^*\ell\nu\) modes that test light-lepton universality and the Lorentz structure of the decay are also presented. The measurements are statistically limited and individually consistent with Standard Model expectations. The new \(R(D^{(*)})\) results are also compatible with experimental world averages, which continue to exhibit tension with Standard Model predictions. Belle II has now accumulated an \(\Upsilon(4S)\) data sample exceeding that of Belle, offering substantial opportunities for new analysis approaches and heightened precision.

We consider quickest change-point detection in generalized linear models when the post-change parameter is unknown. Such problems arise, for example, when monitoring whether the relationship between a learned representation and an observed response changes over time. Since detection must occur soon after a change, the post-change parameter must be estimated from a limited number of observations, making finite-sample estimation accuracy particularly important. We propose a window-limited CUSUM procedure that estimates the post-change parameter using a variational-inequality (VI) formulation. For noncanonical links, VI estimation can offer favorable finite-sample accuracy over short windows despite the asymptotic efficiency advantage of maximum likelihood estimation. We characterize how estimation error reduces the effective post-change information rate and affects detection delay, and derive finite-window guarantees accounting for both statistical error and incomplete numerical optimization. The resulting procedure retains false-alarm control and achieves first-order asymptotic optimality under suitable conditions. Numerical experiments compare detection delay and update cost for fixed-budget estimation and benchmark VI-WLCUSUM against alternative detectors.

Visual token compression has emerged as a critical technique for reducing the inference cost of Vision-Language Models. Existing methods select tokens based on text relevance, visual diversity, or a fixed combination of both—treating the relevance-diversity balance as a universal constant independent of the input. We conduct a systematic analysis to challenge this assumption. At the dataset level, grid-searching the optimal relevance-diversity weight per benchmark reveals substantial cross-task variation. More critically, we further maximize per-sample answer likelihood to recover sample-level oracle weights, finding that their distributions not only differ markedly across benchmarks but also exhibit high within-benchmark variance. Based on these observations, we propose SAVER, a framework for Sample-Aware relevance-diversity balanced Visual tokEn Reduction. We first derive a Geometric Intent Prior (GIP) from the spatial concentration of text-relevant tokens to infer a sample-specific relevance-diversity weight, and then apply this weight to jointly select tokens via a Weighted Text Centroid (WTC) for relevance scoring and a Visual Marginal Coverage Gain (VMCG) for diversity. SAVER requires no additional training and achieves sample-aware balancing entirely at inference time. Extensive experiments demonstrate that SAVER consistently outperforms state-of-the-art baselines across diverse benchmarks, retaining 96.5% of full-model performance on LLaVA-1.5-7B with only 11.1% of the original visual tokens, and further generalizing to LLaVA-NeXT and Mini-Gemini.

To address non-uniqueness in geophysical inversions, Generative Neural Networks (GNNs) are often used to parameterise geological or other prior information in a lower-dimensional space (the so-called latent space), thereby reducing the number of inversion parameters. However, the training samples that represent this information are then always incorrect, because Earth structure is not intrinsically low-dimensional, and all GNNs are subject to training errors. Consequently, even when solutions are mapped back to the original higher-dimensional space, they inherit this imperfection and remain restricted to a low-dimensional manifold. From a mathematical point of view, latent space inversion thus fails fundamentally: generally, it can not even reproduce the best, higher-dimensional approximation to the true Earth. To overcome this limitation, we introduce an `escaping prior' distribution, which provides the necessary freedom to find solutions beyond the manifold of GNN-generated solutions, while still using the GNN as a guide. We use a hierarchical Bayesian approach to adapt the strength of prior information to the observed data and synthetic forward function $\textit{a posteriori}$. We illustrate our method on low-dimensional problems where the escaping prior is defined over a 1D manifold in 2D space. We then compare our approach with standard latent space and model space inversions using a uniform prior distribution in synthetic tomography problems, ranging from simple to more geologically realistic cases. Furthermore, we tested scenarios where the true solution is either far from or close to the manifold. In each case, our escaping prior shows significantly better results than the alternatives. It results in solutions beyond each GNN's generative capabilities, solutions which include the best possible model, and clearly demonstrates the limitations of latent space inversion.

Stable bound states living on a discrete one-dimensional N-point lattice are assumed generated by a certain two-parametric PT symmetric potential $V=V(A,B)$. At any $N$, the coordinates $A=A^{(EP)}$ and $B=B^{(EP)}$ of the Kato's exceptional point (marking the phase-transition instant of the loss of stability) are defined, exactly, via roots of a certain elementary quartic polynomial.

We prove that the singular locus of the geodesic flow on the unit tangent bundle of a nonpositively curved surface enjoys an additional regularity property: its characteristic function belongs to the geometric anisotropic Sobolev space $H^{5/8-}_xH^{-13/16}_\theta$. In particular, this implies that if the geodesic flow is not ergodic with respect to the Liouville measure, then the zero set of the curvature must be thick in the sense that it supports a nonzero, nonnegative function in $H^{5/8-}$. The proof relies on semiclassical analysis and a new radial estimate for parabolic flows.

In this paper we construct a one-parameter family of periodic Airy line ensembles $\mathcal{A}^{\mathrm{per},R}$ for $R>0$ as weak scaling limits of Borodin's periodic Schur processes. We establish several properties of $\mathcal{A}^{\mathrm{per},R}$, including a determinantal point process structure with an explicit correlation kernel. We also show that, after a parabolic shift and normalization, $\mathcal{A}^{\mathrm{per},R}$ satisfies, on its domain $[-R/2,R/2]$, the same Brownian Gibbs property enjoyed by the usual Airy line ensemble. A main technical contribution of the paper is a general tightness criterion for determinantal line ensembles, which extends the argument used by Johansson in [Comm. Math. Phys. 242 (2003), 277-329] to prove a functional limit theorem for the discrete polynuclear growth model.

We study a formulation of regular language membership as a promise problem, which we dub conditional membership: we fix a promise language $P$ and a target language $T \subseteq P$, we are given a word $w$ which is promised to belong to $P$, and we must determine whether $w$ belongs to $T$. We study the deterministic worst-case query complexity of this task, when the length $n$ of $w$ is known in advance and where each query examines the $i$-th position of $w$ for some $1 \leq i \leq n$ and returns the $i$-th letter $w_i$ of $w$. Our goal is to characterize the asymptotic complexity as a function of the word length $n$, depending on the fixed languages $P$ and $T$. Our main result is to show that the complexity of this task admits a trichotomy between three regimes: $O(1)$ queries, $\Theta(\log n)$ queries, and $\Theta(n)$ queries. We further show that the trichotomy criterion is decidable, namely, it is PSPACE-complete to decide which regime applies, given nondeterministic automata $A_P$ and $A_T$ respectively representing $P$ and $T$.