Systems and Control
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Sunday 6 September
The widespread deployment of inverter-based resources (IBRs) is reshaping power system dynamics by their control strategies. Grid-following (GFL) and grid-forming (GFM) inverters are expected to co-exist from now on. However, ensuring the transient stability of such systems remains a critical and underexplored challenge. In this paper, a comprehensive transient stability analysis of the mixed-GFM-GFL inverter system is conducted using a newly derived local energy function. The transient stability analysis model of the system is first derived to reveal the dynamic interactions between GFL and GFM inverters during transient responses. A local energy function is then proposed for analytical transient stability assessment. Instead of ignoring or making harsh assumptions on inverters' coupled damping effect, the proposed local energy function carefully addresses its impact on system transient stability by locally bounding the region where the system's energy dissipation property holds. Such a property fully considers the combined damping effect of GFL and GFM inverters. The derived critical energy and stability index provide a fast and quantitative transient stability assessment. Furthermore, the adverse impact of insufficient damping of GFM inverters on system transient stability is thoroughly investigated. The effectiveness of the proposed method is validated on the modified IEEE 14-bus system with mixed GFL and GFM inverters, and the numerical simulation results further demonstrate the small conservativeness of the proposed method.
Social and Information Networks
Tracking and Predicting Evolution of Social Communities
We develop an algorithmic framework for studying the evolution of communities in social networks. We begin with the theoretical foundation, from which we conclude that an evolution is at most as strong as its weakest link. This allows us to formulate an efficient algorithm to identify all evolutionary sequences in a dynamic social network. We use this algorithm to empirically study community evolution in several large social networks, to identify those features of the early stages of a community that indicate whether a community is going to be shortlived or not. Our results show that it is possible to correlate the lifespan of a community to structural parameters of its early evolution; these conclusions are robust across all the social networks we have investigated.
Read the full abstract
Computer Vision
RoLA: Rotary-Positioned Low-Rank Linear Attention for Efficient Diffusion Transformers
Diffusion Transformers (DiTs) achieve strong video generation quality, but their dense spatiotemporal self-attention scales quadratically with sequence length and quickly becomes the dominant inference bottleneck. Sparse low-rank hybrids alleviate this cost by combining a local sparse branch with a global compressed branch. In video DiTs equipped with 3D Rotary Position Embeddings (RoPE), the global branch faces a structural compatibility issue: when RoPE is applied before a nonlinear feature map, the rotation and nonlinearity generally do not commute, making it difficult to keep a query-independent linear summary while preserving relative rotary geometry. Existing work often sidesteps this issue by replacing genuine cross-token global aggregation with coordinate-conditioned surrogates or learnable absolute positional modules. These compromises can be effective, but they approximate relative decay from absolute coordinates and introduce extra positional parameters. We propose RoLA, a rotary-positioned low-rank linear-attention branch that keeps genuine cross-token aggregation while remaining compatible with a reusable linear summary. The design applies RoPE outside the nonlinear low-rank feature map and reuses a truncated subset of the pre-trained rotary schedule matched to the low-rank bottleneck. This yields a linear-time low-rank global branch with relative positional behavior by design and no additional positional parameters; the full sparse–low-rank module still includes the fixed-sparsity sparse branch. Experiments on open-source video DiTs show that the resulting method remains competitive in generation quality at 90% sparsity while achieving 2.63$\times$ end-to-end inference speedup on Wan2.1-14B (720p, 81 frames, measured on an NVIDIA H100 GPU).
Read the full abstract
High Energy Physics (Theory)
Constructing the Monopole Formula for $A$-Type Good Quivers via Quiver Yangians
Based on the conjectured (and checked for tree-type quivers) quiver Yangian/Coulomb branch algebra correspondence, we give a quiver Yangian interpretation of the monopole formulas for 3D $\mathcal N=4$ good $A$-type quiver gauge theories, also known as the $T_\rho(SU(N))$ theories. Using the algebra action on the $\frac12$-BPS vortices, we reorganize the generators of the truncated shifted quiver Yangian into boundary-adapted generators, and obtain the classical relations in terms of $U(N)$ Casimirs on the Slodowy slice $\mathcal S^{\mathfrak{gl}_N}_\rho$, whose coordinates are given by these boundary-adapted generators. We also show that the boundary-adapted generators and the Casimir relations have the correct fugacities as predicted by the Hall-Littlewood expression of the monopole formula for $T_\rho(SU(N))$, hence reconstructing it as the Hilbert series of the truncated shifted quiver Yangian.
Read the full abstract
Rings and Algebras
Filtered deformations of three-variable polynomial algebras
We classify all filtered deformations of polynomial rings in three variables. We use our characterization to answer a special case of a question of Etingof by showing that if $F$ is an algebraically closed field of positive characteristic and $A$ is an $F$-algebra with a filtration by finite dimensional subspaces whose associated graded ring is a polynomial ring in three variables then $A$ satisfies a polynomial identity.
Read the full abstract
Symplectic Geometry
Lusternik-Schnirelmann category of Lee Forms on locally conformally symplectic manifolds
The Lusternik–Schnirelmann category of a closed symplectic manifold admits various estimates. Locally conformally symplectic (LCS) geometry is a generalization of symplectic geometry involving a closed one-form, called the Lee form. In this paper, we introduce Farber's Lusternik–Schnirelmann category for closed one-forms into LCS geometry by applying it to the Lee class. In particular, this provides a new topological approach to LCS structures of the second kind. We then obtain lower bounds for this invariant under LCS blow-ups.
Read the full abstract
Cryptography and Security
A Trustworthy Watermarking Framework for LLM-Generated Food Safety Content
Large language models are transforming many industries with their text generation abilities. However, their outputs can be easily tampered with, creating serious risks in critical areas such as food safety reporting. To protect the integrity and traceability of AI-generated content, this paper introduces ToSS (Token Oriented Repartitioning and Strategic Selection), a reliable authentication method using adaptive dual watermarking. The key innovation of ToSS is its dual watermark encoding approach that divides vocabulary tokens into black and white sublists, enabling precise bit-level embedding of traceability information. Additionally, an entropy adaptive mechanism dynamically selects text regions with high prediction uncertainty for watermark insertion, maintaining text fluency and factual accuracy while ensuring reliable traceability. Experiments on multiple datasets, including food domain texts, demonstrate that ToSS achieves leading performance in both watermark capacity and decoding accuracy.
Read the full abstract
Strongly Correlated Electrons
Electronic correlations and fluctuating lattice distortions in vanadium dioxide
Metal-insulator transitions in correlated materials are often accompanied by a change of crystal structure. They are commonly described within a coherent lattice approximation, which combines a correlated treatment of the electrons with a lattice represented by one or a few classical distortion coordinates fixed by minimizing an energy. The structural degrees of freedom then carry no entropy of their own, even though the partition of the transition entropy between electrons and lattice is often what decides the transition temperature. Here we develop a stochastic semiclassical extension of dynamical mean-field theory in which correlated electrons and fluctuating lattice distortions are evolved together, extending previous formulations from linear to nonlinear electron-phonon interactions and from a single to several coupled lattice modes. The resulting non-conservative Langevin equations have deterministic forces, damping, and correlated noise generated self-consistently by the interacting electronic subsystem. Applying the approach to a minimal two-orbital model containing the two symmetry-distinct distortions of the monoclinic phase of vanadium dioxide, we find that the two distortions melt at well separated temperatures, giving an insulating, an intermediate metallic, and a high-symmetry phase. The two transitions have different origins: the melting of the dimerization is driven by the coupling to the correlated electrons and is captured already within the coherent lattice approximation, whereas the restoration of the undistorted structure requires the entropy of the fluctuating lattice. Since the electronic subsystem is treated within nonequilibrium dynamical mean-field theory, the framework can be extended in future work to photoexcited systems with nonthermal electronic distributions.
Read the full abstract
Combinatorics
Rainbow paths in directed graphs
An old problem in combinatorial group theory asks, given a group $\Gamma$ and a subset $S \subseteq \Gamma$, when does there exist an ordering $s_1, \cdots, s_k$ of the elements of $S$ such that the partial products $\prod_{i = 1}^j s_i$, $1 \leq j \leq k$, are all distinct. If such an ordering exists, we call $S$ rearrangeable. There have been many conjectures about rearrangeable subsets, the most general being that for every group, every subset not containing the identity element is rearrangeable. We prove an asymptotic version of this: For any Group $\Gamma$ and any subset $S \subseteq \Gamma$, there exists a rearrangeable set $S' \subseteq S$ such that $|S'| = |S| - o(|S|)$. To do this we build upon the work of Bucić, Frederickson, Müyesser, Pokrovskiy, and Yepremyan focusing on the following problem of independent interest in Graph Theory. If $G$ is a d-regular properly colored directed graph does there exist a rainbow path of length $d - 1$? We establish an asymptotic version of this, proving that $G$ contains a rainbow path of length $d - o(d)$. This solves two problems given by Bucić, Frederickson, et al. and proves the above result on rearrangeable subsets of groups by considering the Cayley graph of $\Gamma$ with (not necessarily generating) set $S$.
Read the full abstract
Differential Geometry
Klein bottle and optimal systolic inequality for nonpositively curved surfaces
We show that the systolic area of every nonpositively curved closed surface $M$ other than the torus is at least $1$, with equality if and only if $M$ is isometric to a square flat Klein bottle. The proof focuses on the Klein-double $4\mathbb RP^2$ and exploits Weil's isoperimetric inequality and a comparison theorem involving a new kind of exponential-type map.
Read the full abstract
Computer Vision
Counterfactual Tests for Measuring Chain-of-Thought Faithfulness in Visual Language Models
Chain-of-thought (CoT) may often look plausible, yet it may not faithfully reflect the model's decision-making process. While methods for measuring the faithfulness of CoTs for textual inputs have been increasingly introduced, using these methods for visual inputs is not straightforward. In this work, we adapt the family of counterfactual methods for measuring CoT faithfulness, namely the Counterfactual Test (CT) and Correlational Counterfactual Test (CCT), to visual inputs, and call them vCT and vCCT, respectively. Using vCT and vCCT, we benchmark eight recent open-source Vision Language Models (VLMs) on two datasets. Our analysis shows that CoTs do not reliably track visual evidence that influences model predictions: they may omit the removed object even when its removal causes a large prediction shift, yet mention it when the shift is small. We further find that Predict-then-Explain explanations align more strongly with perturbation-induced probability shifts than pre-answer CoTs, while binary vCT scores are often nearly saturated. We also include a reconstruction control, in which images pass through the same editing pipeline without object removal, and find that the main object-removal intervention induces larger shifts than reconstruction alone. We construct and release Counter-SNLI-VE and Counter-A-OKVQA, two datasets of image pairs that differ by a single object.
Read the full abstract
Computation and Language
DianShi-RxnDB: A Large-Scale, Fine-Grained Organic Reaction Data Platform Built via a Fully Automated Pipeline for Researchers and AI Agents
High-quality structured organic reaction data are essential for developing artificial intelligence for chemistry (AI4Chem), yet much of this knowledge remains dispersed across patent text, images, and reaction schemes. We present DianShi-RxnDB, a large-scale, fine-grained organic reaction data platform built via a fully automated extraction and normalization pipeline integrating patent text, images, and reaction schemes. Its corpus covers organic synthesis patents from the USPTO and EPO published between 1976 and 2025, yielding approximately 24 million reaction instances, of which approximately 14.8 million (61.7%) pass automated qualification checks. Each instance represents a specific single-step experiment recording participants, roles, quantities, temperatures, reaction times, yields, experimental procedures, and provenance links to source patents. In a manual evaluation of 1,300 sampled qualified instances, the micro-averaged field-level accuracy was 92.95%. A matched comparison with Pistachio further indicated advantages in deduplicated record counts, representation granularity, and field-level exact agreement. The platform provides a Web research workbench for searching, filtering, comparing, and source-verifying records, and a Model Context Protocol (MCP) service offering AI agents composable structured retrieval tools. DianShi-RxnDB is available at https://dianshi.opendatalab.org.cn/ .
Read the full abstract
Computation and Language
PARSER: Read in Parallel, Reason in Depth for Long-Context LLM Agents
Sequential memory agents process long documents by reading chunks one after another while maintaining a compact memory state, coupling document traversal to reasoning depth. This coupling introduces sensitivity to evidence placement and ties inference latency linearly to document length. We introduce PARSER, which decouples reading from reasoning. A bank of lightweight subagents each bound to a single chunk read the entire document in parallel, while a lead agent reasons in depth through iterative scatter–gather rounds: at each round it broadcasts a query to all subagents, aggregates the returned evidence, and formulates a deeper follow-up query conditioned on what has been found so far. This decoupled design concentrates all learnable behavior in the lead agent, which is optimized with reinforcement learning, while the subagents remain frozen off-the-shelf models. On multi-hop QA with contexts ranging from 7K to 896K tokens, PARSER with a 4B backbone outperforms the strongest sequential memory baseline by 5.7 points on average and by 12.0 points at 896K tokens. Scaling to a 9B backbone, PARSER surpasses DeepSeek-V4-Pro by 6.3 points. Controlled experiments confirm that PARSER is robust to perturbations in evidence position, order, and distance, conditions that cause large accuracy swings in sequential methods, while reducing inference latency by up to 11x.
Read the full abstract
Combinatorics
Computing and Bounding the Number of Eulerian Orientations for Certain Classes of $4$-Regular Graphs
An Eulerian orientation of a $4$-regular undirected graph (simple or multigraph) $G=(V,E)$ with $n=|V|$ vertices is an assignment of directions to its edges such that every vertex $v \in V$ has the same indegree and outdegree. In the present article, we improve the bounds on the number of Eulerian orientations for certain classes of connected, loopless $4$-regular graphs. The previous bound is due to M. Las Vergnas (1983) and is exactly $9\cdot 2^{n-3}$, which is a sharp bound for a certain family of multigraphs with $n\geq 4$. Here, we show that the number of Eulerian orientations for all biconnected $4$-regular multigraphs is at most $2^n+2$, which is also sharp. We exhibit families of graphs that attain this maximum value. For simple graphs, we prove an upper bound of $\mathcal{O}(3^{n/2})$ in the biconnected case and $\mathcal{O}(6^{n/3})$ for the separable case. Additionally, we provide a divide-and-conquer algorithm that leverages structural properties to compute the exact number of Eulerian orientations for separable graphs without exhaustive enumeration. Finally, we analyze the effect of standard inductive construction operations, used to generate $4$-regular graphs from smaller ones, as shown by F.Bories et.al. (1983) for simple graphs and by G.Ding et.al. (2003) for multigraphs, on the number of Eulerian orientations.
Read the full abstract
Metric Geometry
Embeddability and rectifiability of Lipschitz differentiability spaces
We prove that Lipschitz differentiability spaces which bi-Lipschitz embed into an RNP-space are countably rectifiable. In contrast to earlier methods of Cheeger and Kleiner, our approach does not rely on differentiating RNP-targets, and uses instead decomposability bundles and a careful blow-up analysis. We also present decomposability bundles in a way which avoids the mention of Alberti representations and generalizes the approach of Alberti–Marchese to measures in RNP-spaces. We moreover study fragment-wise differentiability into RNP-targets, give a new ${\rm Lip}-{\rm lip}$-type characterization of RNP-differentiability spaces, and address a question of Le Donne asking for a characterization of spaces $(X,\mu)\subset\ell^2$ whose Gromov–Hausdorff tangents are Hausdorff limits of $r^{-1} (X-x)$ in $\ell^2$ as $r\to 0$.
Read the full abstract
Computer Vision
Attention-Enhanced Deep Features with Heterogeneous Ensemble Learning for Glaucoma Detection
Glaucoma is a progressive optic neuropathy characterized by irreversible damage to the optic nerve, making timely diagnosis critical to prevent permanent vision loss. Although deep learning has demonstrated promising performance in automated glaucoma detection, existing approaches often overlook feature refinement, suffer from class imbalance, and rely on individual classifiers that limit prediction robustness. To address these challenges, this paper proposes a hybrid glaucoma detection framework that integrates attention-enhanced deep feature extraction with heterogeneous ensemble learning. Specifically, deep representations are extracted using InceptionV3 and subsequently refined by incorporating the Convolutional Block Attention Module (CBAM) to enhance discriminative retinal features. To improve classification robustness, the extracted features are classified using multiple machine learning models together with Single-Level Ensemble (SLE) and Double-Level Ensemble (DLE) strategies, while SMOTE combined with Tomek Links (SMOTE+TL) is employed to alleviate class imbalance. Furthermore, a systematic comparison of handcrafted, deep, and attention-enhanced deep feature representations is conducted. Experimental evaluation on two public retinal fundus datasets demonstrates that deep feature-based methods consistently outperform handcrafted feature-based methods, while the proposed attention-enhanced framework achieves the best overall performance. Furthermore, Grad-CAM visualizations confirm that the proposed model focuses on clinically relevant retinal regions, providing interpretable evidence on the model's prediction process.
Read the full abstract
Systems and Control
User-specified Kelvin-hour budgets within model predictive control for energy-efficient buildings: Simulation and field demonstration
For energy-efficient building control, model predictive control (MPC) has been widely proposed to reduce energy use while maintaining prescribed indoor temperature bounds. In practice, MPC implementations may still produce temperature-bound violations because of model mismatch, weather-forecast errors, and softened constraints introduced to preserve feasibility. Existing soft- constrained MPC usually handles this trade-off through slack-weight tuning, so users cannot prescribe a violation-severity budget before operation. This paper develops a Kelvin-hour (Kh)-budget MPC that allows users to specify a running-average budget for temperature-bound violation severity. During operation, the controller uses the realized Kh budget surplus or overspend to update the temperature bounds, providing step-by-step budget feedback. We evaluate the method in two high-fidelity BOPTEST simulation cases and in an occupied residential deployment. In the first one-zone case, prescribed budgets produce clear running-average Kh violation responses across predictors and disturbance settings. At a prescribed budget of 0.005 Kh/step, the controller reduces energy use by 31.9% and Kh violation by 36.2% relative to the built-in default controller. In the second coupled two-zone case, separate zone budgets produce zone-level responses, with 18.9%-20.5% energy reduction and 9.4%-64.3% zone-level Kh violation reduction. In the occupied residential deployment, the running-average Kh violations remain close to or below their prescribed budgets under real sensing, actuation, weather, and occupancy conditions.
Read the full abstract
Signal Processing
A GNN-Enhanced Reinforcement Learning Framework for Emergency Communications in ORAN-based Non-Terrestrial Networks
During disaster scenarios and periods of extreme data demand in next-generation wireless communications, conventional terrestrial networks (TNs) often become unreliable or fail entirely, leading to critical service disruptions. In such contexts, non-terrestrial networks (NTNs) emerge as a promising solution to ubiquitous and resilient connectivity. Furthermore, the open radio access network (ORAN) paradigm facilitates network disaggregation and flexible functional splitting among its key components, namely the central unit (CU), distributed unit (DU), and radio unit (RU), which can be deployed across heterogeneous NTN platforms according to service requirements. However, this flexibility introduces significant challenges in terms of network complexity and real-time control. To address these challenges, this paper proposes an intelligent ORAN-enabled NTN framework for emergency communication scenarios. The proposed system leverages graph neural networks (GNNs) to model the dynamic network topology and employs a reinforcement learning (RL)-based Q-learning algorithm, formulated as a Markov decision process (MDP), to enable adaptive and real-time network control. In this framework, network nodes are treated as states, and optimal decisions are learned based on system dynamics. The spatial distribution of user equipment (UE) is modeled using an inhomogeneous Poisson point process (IPPP) with a rejection sampling technique, capturing realistic user density variations. Simulation results demonstrate that the proposed GNN-enhanced RL approach significantly improves network performance in terms of latency and service reliability, thereby enabling efficient and robust operation under emergency conditions.
Read the full abstract
Statistical Methodology
Bayes Estimators with Performance Comparable to Empirical Bayes Estimators and Improved Local Robustness
Bayes estimation has been extensively studied and widely used in statistics, decision theory, signal processing, machine learning, and system identification. Among its variants, empirical Bayes (EB) estimation has attracted considerable attention due to its favorable estimation performance and computational tractability. However, the direct plug-in dependence of an EB estimator on hyperparameters can make it locally sensitive to hyper-parameter perturbations. This paper considers the linear regression model and focuses on the EB estimator by employing the marginal maximum likelihood hyper-parameter estimator. For conciseness, this estimator is simply referred to as the EB estimator. Given a family of EB weighting functions, a generalized Bayes estimator is constructed with the same excess mean squared error (XMSE) as the corresponding EB estimator. Here, the XMSE is a second-order asymptotic measure of the mean squared error difference between the estimator of interest and the maximum likelihood estimator. Furthermore, the EB estimator is shown to be at most firstorder sensitive to hyper-parameter perturbations, whereas the constructed Bayes estimator is at most second-order sensitive, making it locally more robust. The computational complexities of these two estimators are also analyzed. In some cases, the constructed Bayes estimator can be computationally comparable to, or more efficient than, the EB estimator. These theoretical results are further supported by numerical simulations.
Read the full abstract
Quantum Physics
Thermodynamic performance of a measurement-driven quantum engine with a two-parameter $(p,q)$-deformed harmonic oscillator
We address a single-bath measurement-driven engine with a two-parameter $(p,q)$-deformed oscillator spectrum. A nonselective Gaussian quadrature measurement supplies energy between two population-preserving work strokes. We evaluate its finite-dimensional transition matrix without outcome quadrature and derive the leading weak-measurement heat. Uniform spectral scaling gives the Otto efficiency exactly; deformation changes the relative gaps and the conversion of measurement energy into work. The infinite quadrature is essentially self-adjoint for $\max(p,q)\le1$; above this range a self-adjoint extension must be specified. We instead choose an explicit $N=40$ effective matrix model. At $\beta\hbar\omega_i=1$, fixed physical resolution $\sigma=4$, and $\omega_f/\omega_i=2$, the largest reported efficiency in $0.7\le p_f,q_f\le1.2$ is approximately $0.782$, compared with the undeformed value $0.5$. Independent dimension checks reproduce these values but do not establish an infinite-dimensional limit in the enhanced regime. Comparisons at selected matched resources and at fixed mean deformation reveal dependence on spectral shape beyond first order. The working-medium efficiency excludes measurement-apparatus and control implementation costs.
Read the full abstract
Distributed Computing
ForgeStencil: Automating Per-Case Stencil Specialization from Kernels to 100+ Real Applications
Industrial and scientific computing rests on a few core kernels, and the stencil is among the most widely used: weather and climate models, seismic imaging, fluid dynamics, and image processing all run on it. No single stencil implementation is fastest: the optimal kernel changes qualitatively with stencil shape, grid shape, precision, and host application. For two decades the field has answered with general methods (DSLs, code generators, autotuners), because specialized solutions were too expensive to build per case, so all reuse one human-authored recipe. That reuse costs performance; we call the cost the generality tax. This premise no longer holds: code-synthesis agents now build a correct, specialized solution per case at acceptable cost. ForgeStencil automates this. A Kernel Agent synthesizes CUDA and forges a per-configuration map of specialized operators, removing the tax case by case. On an A100 the map beats the strongest public baseline in 37 of 37 cases: geometric mean 2.35x against same-precision f32 baselines and 1.95x for fp16, each reported under its own precision. The same change reaches end-to-end application performance. A generic operator library is tuned once for its own general case and reused across applications, so its shapes, layouts, and launch boundaries are optimal for none of them: using it is the application-level form of the tax. An App Agent instead forges a specialized solution per application, locating hotspots, rewriting application structure, and validating and integrating each change. Across 100 real industrial and scientific codes the end-to-end median speedup is 1.41x against each application's own GPU baseline. To our knowledge this is the first demonstration that per-case synthesis carries from a kernel library to complete applications at this breadth, and evidence that reuse is no longer the default in a domain built on it for two decades.
Read the full abstract
Instrumentation and Methods for Astrophysics
Chromatic Anisoplanatism in Adaptive Optics: A detailed comparison of theory with numerical simulations
Context. Chromatic Anisoplanatism arises in Adaptive Optics because atmospheric refraction makes rays of different wavelengths follow different paths through the atmosphere resulting in different wavefronts at the sensing wavelength and the science wavelength, introducing a correction error that grows rapidly with zenith distance. Aims. Using numerical simulations based purely on geometric optical propagation we aim to validate the theoretical framework describing wavefront errors caused by Chromatic Anisoplanatism. We characterize the resulting phase variance and Strehl ratio loss for telescope apertures up to 39 m, retrieving their second-order statistics. Methods. We compare the analytical expressions for the Chromatic Anisoplanatism wavefront phase variance against independent Monte Carlo simulations. The framework models an 8-layer approximation of the median Cerro Armazones turbulence profile using Kolmogorov and von Karman turbulence models. We simulate pupil diameters from 4.2 to 39 m with lambda_WFS = 589.16 nm for the wavefront sensing and lambda_Sci = 1.25 microns as the science wavelength. A phase-screen recombination strategy increases the number of quasi-independent realizations up to ~ 5 x 10^5, optimizing statistical accuracy. Results. Simulated phase variances match theoretical predictions well across all apertures, though theory slightly overestimates variance in the von Karman case. The Strehl Ratio loss exceeds 30% for zenith angles >= 70 deg. The residual wavefront is dominated by high spatial frequencies; removing tip-tilt reduces phase variance by <= 1% for apertures >= 10 m. The massive dataset reveals highly non-Gaussian probability density functions for the phase variance and Strehl Ratio values, both characterized by non-zero skewness and kurtosis, and long tails toward rare, high wavefront phase variance and/or low Strehl Ratio error events.
Read the full abstract
Analysis of PDEs
Gradient estimates for the Lagrangian mean curvature equation with Lipschitz critical phase
Let $n\ge3$. We prove a priori interior gradient estimate for the Lagrangian mean curvature equation \[ \sum_{i=1}^n\arctan\lambda_i(D^2u)=\theta(x) \] under the critical and supercritical phase condition $\theta\ge (n-2)\pi/2$, assuming only that the Lipschitz norm of the prescribed phase is bounded.
Read the full abstract
Distributed Computing
Gutenberg: Taming Latency-Critical Cloud Services with Near-Data-Processing
Latency-critical cloud services place growing pressure on memory while requiring isolation, fairness, and predictable QoS. Near-data processing (NDP) reduces data movement by executing requests close to memory, and prior systems further improve locality through caching and replication. However, writes make replica maintenance expensive, while uneven compute and memory traffic can overload a few NDP units and increase tail latency. Existing throughput-oriented schedulers do not fully address these challenges for co-located cloud services. We present Gutenberg, a CPU+NDP for mutable, latency-critical cloud services. Gutenberg stages subpage updates in a CPU-resident delta buffer, allowing hot writable pages to remain replicated without eager full-page synchronization. It also adopts CPU helper cores to assist request execution when NDP execution or replica maintenance becomes costly. An online controller jointly decides page placement, replication, CPU/NDP execution, and routing using access patterns, queue pressure, and feedback from prior decisions. The system further enforces isolation and fair resource allocation across services. We also model-check CPU–NDP coordination protocol for correctness. We evaluate on TailBench using ZSim with Ramulator-calibrated memory timing. Across evaluated services, Gutenberg outperforms prior systems, reducing average and p99 latency by up to 80.4% and 85.8%. It also improves isolation and fairness while adapting to changing workload behaviors.
Read the full abstract
Computation and Language
A Grapheme-Aware Indic Tokenizer for Tamil: Large-Scale Training and Intrinsic Evaluation
Tokenization forms the foundation of modern Natural Language Processing (NLP) systems by transforming raw text into discrete units that neural language models can process. The effectiveness of this process directly influences vocabulary efficiency, sequence length, computational cost, and downstream model performance. Although multilingual tokenizers such as Byte Pair Encoding (BPE), WordPiece, and SentencePiece have performed well across numerous languages, they often segment morphologically rich Indic languages inefficiently. Tamil, in particular, poses unique challenges because its grapheme-based writing system can represent a single visible character with multiple Unicode code points. In this work, we present a grapheme-aware Indic tokenizer for Tamil that preserves complete grapheme clusters through a reversible Unicode mapping strategy prior to WordPiece vocabulary learning. By operating on grapheme-level representations instead of individual Unicode code points, the tokenizer produces linguistically meaningful token boundaries while remaining fully compatible with transformer-based language models. The tokenizer is trained on a large-scale Tamil corpus and evaluated using a comprehensive intrinsic evaluation framework that measures compression efficiency, token fragmentation, information density, and vocabulary utilization. Experimental evaluation compares the proposed tokenizer against five widely used multilingual tokenizers: GPT-2, mBERT, mT5, mBART, and NLLB. The proposed tokenizer achieves the strongest performance among the evaluated tokenizers on fragmentation- and sequence-efficiency-oriented intrinsic metrics, while matching the highest observed compression ratio. These results demonstrate the effectiveness of grapheme-aware preprocessing for Tamil tokenization.
Read the full abstract
Functional Analysis
Order closure, order adherence and Fatou norms
We record two results in the theory of vector and Banach lattices related to order adherence. First, we give a negative answer to Gao and Leung's question on whether the $uo$-adherence of sublattices must be order closed. Second, we present a self-contained counterexample showing that a Banach lattice with a weakly Fatou norm need not admit any equivalent lattice norm with the Fatou property.
Read the full abstract
Quantum Physics
Conditional Error Exponents of Probabilistic Quantum Resource Distillation
Quantum resource distillation converts imperfect resource states into highly resourceful targets and is a central task in quantum information processing. When distillation is probabilistic, unfavorable outcomes may be discarded and the quality of the output is evaluated conditioned on success, leading to a notion of reliability that differs from the deterministic setting. Here we study the conditional error exponents of probabilistic quantum resource distillation under approximately resource-nongenerating instruments and derive its general connection to the postselected composite quantum hypothesis testing against the corresponding set of free states. This reduction leads naturally to scale-invariant projective distinguishability measures and provides finite-blocklength bounds for a broad class of resource theories. We apply the framework to entanglement, coherence, and magic. For entanglement and magic, we obtain exact zero-rate asymptotic characterizations of the conditional error exponent in terms of regularized projective distinguishability, while for coherence we derive general finite-blocklength and asymptotic bounds. For representative entangled and magic states, the resulting quantities can be evaluated analytically and reveal regimes in which postselection yields a strictly larger conditional error exponent than deterministic distillation. Furthermore, we also show that the conditional error exponents for probabilistic entanglement distillation under a strictly positive success probability constraint are smaller than those with no constraints on the success probability.
Read the full abstract
Populations and Evolution
Modeling Medea gene-drive population replacement: thresholds and release strategies
Mosquito-borne diseases such as dengue, Zika, and yellow fever impose a substantial global health burden, motivating genetic control strategies that replace wild mosquito populations with disease-refractory ones. Maternal-effect dominant embryonic arrest (Medea) is a gene drive in which the offspring of a Medea-carrying mother die unless they inherit the Medea allele, producing biased inheritance capable of driving a linked refractory trait to high prevalence. We develop and analyze a continuous-time compartmental model of Medea dynamics in Aedes aegypti that tracks mosquito abundance by life stage and genotype, and that generalizes the drive mechanism to allow both imperfect Medea-killing and imperfect rescue. We characterize the biologically relevant equilibria and derive their local stability conditions, together with genotype-specific reproduction numbers and the basic reproduction number (R0) governing invasion from low frequency. Bifurcation analysis reveals bistability, and hence a release threshold that must be exceeded for Medea to take over, even though R0<1 at baseline. Imperfect killing and imperfect rescue reshape this bifurcation structure in distinct ways: killing leakage governs the invasion threshold, while rescue efficiency governs the composition of the resulting population. Sensitivity analysis identifies the fitness coefficient and killing leakage as the dominant drivers of both threshold and coverage outcomes, with mosquito demographic parameters affecting absolute abundances but not genotype proportions. Simulations of batched release programs inform efficient field deployment and show that reversing an established drive is substantially more costly than achieving forward invasion, though co-releasing both sexes accelerates reversal considerably more than it accelerates invasion.
Read the full abstract
High Energy Physics (Experiment)
Likelihood-Based Unsupervised Anomaly Detection in CMS Dijet Events
We present an unsupervised search for anomalous dijet events in proton–proton collision data using neural spline flow density estimation. A normalizing flow model is trained on a high-dimensional feature space comprising jet, dijet, and event-level observables to learn the dominant Standard Model background directly from data, without assuming a specific signal hypothesis. Events assigned low likelihood under the learned density are identified as potential anomalous events. Using this approach on a CMS Open Data dijet sample, we investigate extreme events in the tail of the anomaly-score distribution and perform an extensive validation and robustness study. This study includes feature-level statistical comparisons, mass decorrelation tests, permutation-based null tests, and evaluations of training stability. The selected anomalous events exhibit notable departures from the background-only expectation, primarily in jet-substructure observables, while remaining stable under several known sources of bias in unsupervised learning. The identified anomalies are distributed across the kinematic phase space and do not exhibit a narrow structure in the dijet invariant-mass spectrum. Instead, they show correlated deviations across multiple observables, consistent with a multivariate difference in jet substructure and event topology rather than a localized resonance. While no claim of new physics is made, this study demonstrates that neural spline flow-based density estimation can be sensitive to rare, structured deviations in collider data and may provide a model-independent exploratory tool for searches for physics beyond the Standard Model at the LHC.
Read the full abstract
Machine Learning (Statistics)
RoPE attention is an exact forward-pass gradient step with softmax intact
We derive an exact gradient-step representation of the RoPE-softmax forward pass. For every deterministic RoPE-softmax attention head with arbitrary affine projection weights, we construct a query-dependent effective matrix $\Delta M_i$ satisfying $y_i = \mu_i + u_i^\top \Delta M_i$, where $\mu_i$ is the uniform mean of the attended values and $u_i$ is the augmented query input. The construction applies the classical exponential divided difference $\rho = \phi_1$ to retain the softmax exactly. Its positive coefficients give a unit gradient-step representation on a query-conditioned quadratic objective. The same function connects the RoPE generator to exact positional finite differences. We derive a tokenwise formula for the error of reusing one query's matrix and prove that a nonconstant finite-cache head cannot admit a globally exact affine query readout. Reconstruction checks and frozen-reuse calibration on one pretrained Qwen2.5-0.5B layer verify the representation and quantify the correction required when one query's matrix is reused.