Dynamical Systems
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Thursday 3 September
Let $g$ be a $C^\infty$ diffeomorphism in the isotopy class of a pseudo-Anosov homeomorphism $f$ such that $g$ and $f$ have the same topological entropy. In 1988, Handel proved that this implies the existence of a semiconjugacy $\pi$ from $g$ to $f$. He stated that, in general, there is at least one point $x$ such that $\pi^{-1}(x)$ is disconnected. We show that this is not the case: for every $x$, the set $\pi^{-1}(x)$ is the intersection of a nested sequence of closed topological disks, and hence is connected. We also prove that there is a unique $g$-invariant probability measure projecting to the measure of maximal entropy of $f$. This measure is entropy-maximizing, hyperbolic, and Bernoulli, and the semiconjugacy induces a metric isomorphism between the corresponding measure-preserving systems.
Soft Condensed Matter
Computing stable configurations of confined smectic liquid crystals with a deep variational framework
Smectic liquid crystals are layered liquid-crystalline phases characterized by orientational order and periodic density modulation. Although their structures can be modeled using continuum theories, computing stable configurations remains challenging in complex geometries, particularly when the high-frequency density modulations associated with smectic layering should be resolved. We propose a deep variational framework (DVF) for computing these configurations within the modified Landau–de Gennes model, in which the coupled orientational and positional order parameters are represented on a regular reference domain while physical confinement is incorporated through coordinate mappings. A warmup penalty mitigates the spectral bias of neural networks toward smooth, nonlayered fields, enabling robust recovery of oscillatory smectic states. Comparisons with a neural-network baseline and finite-difference relaxation demonstrate the essential role of this penalty and the numerical stability of the resulting layered states. The DVF reproduces experimentally established smectic-A defect structures and layer morphologies across diverse confinement geometries and further predicts a chevron-like smectic-C state in a tangent-anchored sphere. Together, these results demonstrate the applicability of the DVF to computing stable smectic configurations across experimentally relevant confinement geometries and anchoring conditions.
Read the full abstract
Superconductivity
Doping dependence of local moments in infinite layer nickelates
The infinite layer nickelates are notable for their lack of long-range antiferromagnetic ordering, in contrast to the parent compounds of the superconducting cuprates. Instead, the nickelates show evidence of short-range glassy behavior in both the undoped and optimally-doped regimes, implying that local electronic moments exist independent of superconductivity. However, the systematic doping-dependent magnetic behavior is not yet fully resolved, and characterizing it could uncover the relationship between local moments and the superconducting dome. In this work, we use muon spin rotation ($\mu$SR) on a (La,Sr)NiO$_2$ doping series from the undoped parent compound, through the superconducting dome, to the over-doped normal state (Sr substitution 0% $\leq$ x $\leq$ 25%) to probe the magnetic ground state and the temperature-dependent static and dynamic behavior. We find that local moments experience spin freezing into a glassy state at temperatures on the order of a few tens of kelvin regardless of the doping level. We also observe a subtle destabilization of the glassy state with increased hole doping. These observations suggest that magnetism and superconductivity are largely decoupled phenomena with indirect interactions described in a multi-orbital framework.
Read the full abstract
Fluid Dynamics
Cascade coalescence dynamically sustains bubble retention near gas-evolving surfaces
Bubble detachment from solid surfaces governs heat, mass, and charge transport across technologies vital to clean energy, including high-current-density water electrolysis and boiling thermal management. At high gas fluxes, however, bubbles remain trapped at active surfaces despite immense buoyancy, severely restricting mass transfer and increasing energy losses. Here, we show that this unexpected surface retention originates from cascade coalescence between unequal-sized bubbles. High-speed observations around microelectrodes demonstrate that when a rising bubble merges with a smaller surface-attached successor, its trajectory abruptly reverses, accelerating toward the substrate at nearly two orders of magnitude above its rising speed. Direct numerical simulations and scaling analysis reveal that asymmetric interfacial retraction during merging generates non-canceling viscous stresses, producing a net downward impulse toward the smaller bubble. Repeated cascade coalescence events accumulate these transient impulses into a steady, time-averaged retaining force capable of opposing buoyancy three to four orders of magnitude beyond quasistatic limits. Our findings establish bubble coalescence as a previously unrecognized mechanism that dynamically sustains bubble retention under high gas flux.
Read the full abstract
Analysis of PDEs
Twisted Gradient Flow Approach to Vegetation-Rainfall-Bushfire Interactions Model
We consider a vegetation-rainfall-bushfire interaction model consisting of two partial differential equations describing the evolutions of bushfire intensity and water availability, together with an ordinary differential equation governing the vegetation density. The main feature of the model is that the ODE contains a nonlocal coefficient multiplying the vegetation density, which makes the mathematical analysis highly nontrivial. In this paper, we regard this nonlocal coefficient as the generator of a deformation of the metric structure of the underlying real Hilbert space. This interpretation enables us to formulate the ODE as a twisted gradient flow on state-dependent Hilbert spaces and to establish the existence of strong solutions to the associated initial-boundary value problem.
Read the full abstract
High Energy Physics (Phenomenology)
Meson mass spectrum in isospin QCD medium from the $N_f=2+1$ quark-meson model
We study phase structures, meson mass spectra, and sound velocities at finite temperature and density in QCD with isospin chemical potential ($\mu_I$). We employ the quark-meson model with $N_f=2+1$, incorporating the Kobayashi-Maskawa-'t Hooft type coupling to capture dynamical effects of the $U(1)_A$ axial anomaly. Within the mean-field approximation at quark one-loop, we analyze the onset of pion condensation, which manifests as a second-order phase transition at low temperatures and may exhibit a first-order behavior at higher temperatures within this approximation. We examine the corresponding mass spectra of scalar and pseudoscalar singlet-octet mesons, in which the $\pi_+$ mass is exactly massless in the superfluid phase due to its Nambu-Goldstone boson nature. The neutral pion mass is, meanwhile, found to exhibit a strictly linear growth with $\mu_I$ in the pion condensed phase. We also investigate the isothermal squared sound velocity and identify characteristic structures associated with the phase transitions. Furthermore, we highlight how enhanced $U(1)_A$ anomaly effects facilitate the pion condensate to generate a less-pronounced sound velocity peak in cold medium. Our findings are expected to provide future lattice simulations with useful information on meson mass spectra from symmetry aspects.
Read the full abstract
Computer Vision
FoRIS: Progressive Foreground Refinement for Training-Free In-Context Segmentation
In-Context Segmentation (ICS) aims to precisely segment arbitrary semantic concepts, such as objects or parts, given one or a few annotated visual exemplars. In this paper, we revisit ICS from a more classical segmentation perspective, viewing it as a coarse-to-fine progressive refinement process. Rather than directly predicting the final mask through reference-query matching, we progressively refine the segmentation from coarse and ambiguous foreground responses to precise and complete foreground structures. Building upon this perspective, we propose a training-free in-context segmentation framework, termed FoRIS. Specifically, FoRIS consists of three key stages: Foreground Purification, Foreground Localization, and Foreground Consolidation, which progressively suppress background distractions, localize discriminative target regions, and recover complete foreground structures through semantic aggregation. Experimental results demonstrate that FoRIS achieves SOTA performance across semantic and part segmentation tasks, with average improvements of 4.5 and 4.8 mIoU points over existing approaches in the 1-shot and 5-shot settings, respectively. Code: https://github.com/Xi-Mu-Yu/FoRIS.
Read the full abstract
Machine Learning
TIGPO: Temporal Instance-Graph Policy Optimization for Long-Horizon LLM Agents
Graph-based policy optimization improves credit assignment for long-horizon LLM agents by organizing rollout trajectories into state-transition graphs. However, existing methods construct graphs independently within each policy update, discarding transitions discovered by earlier policies and limiting advantage estimation to small, batch-local rollout groups. We propose Temporal Instance-Graph Policy Optimization (TIGPO), which extends graph-based credit assignment across policy updates. TIGPO maintains a persistent transition graph for each task, allowing valid transitions discovered by different policy versions to jointly determine credit for current rollouts. To actively reconnect current exploration with historical experience, TIGPO allocates a fixed rollout budget between Exploration slots for ordinary task sampling and Revisit slots for delayed reattempts of previously explored tasks. For each revisit, TIGPO pairs the current rollout group with its corresponding earlier Exploration group to construct a cross-temporal reference. The enlarged reference is designed to stabilize relative advantage estimation under small rollout groups, while comparison on the same task directly captures policy improvement across training stages. Historical transitions and scores serve only as structural and detached statistical references and are never replayed in the policy loss. Experiments on ALFWorld and WebShop demonstrate that TIGPO consistently outperforms prior group-based and graph-based policy optimization methods.
Read the full abstract
Dynamical Systems
SurgeGen: A Hybrid Generative Diffusion Framework for Storm Surge Scenario Synthesis
Predicting storm surge induced by landfalling tropical cyclones is crucial for flood mitigation and coastal risk management. Traditionally, physics-based numerical models simulate storm surge by solving the Navier–Stokes equations using numerical methods, but these simulations are computationally expensive. Generative models are promising for storm surge emulation because they can generate diverse realizations rather than producing a single deterministic prediction. However, their use for storm surge emulation remains largely unexplored. In this paper, we leverage diffusion models for storm surge surrogate modeling, combining a baseline prediction stage with conditional generation to provide a more interpretable modeling framework. We develop SurgeGen, a two-stage generative framework for generating storm surge scenarios conditioned on hypothetical storms with parameters defined in a continuous space. First, a baseline model produces a coarse estimate of the storm surge height. This estimate then conditions a diffusion model, which generates refined storm surge scenarios that better capture spatial patterns and variability. We demonstrate that our approach can generate realistic and diverse storm surge scenarios under conditions both within and outside the training distribution.
Read the full abstract
Audio and Speech Processing
StreamWSR: Streamable and Lightweight Waveform-Domain Neural Speech Super-Resolution
This paper proposes StreamWSR, a Streamable neural Waveform-domain model for speech Super-Resolution (SR). By adopting a fully causal architecture with compact frame-level waveform representation, the proposed StreamWSR supports zero-look-ahead streaming inference while avoiding vocoder-based reconstruction and explicit phase prediction. Specifically, StreamWSR downsamples the input waveform into a compact frame-level representation using strided causal convolutions. Then, a lightweight causal long-short-term modeling backbone is employed to capture both local waveform structures and long-range historical dependencies under causal constraints. Finally, the modeled output is converted back to the waveform domain through a causal transposed-convolution and combined with the input waveform via a residual connection to generate the final high-resolution speech. Experimental results on 16 kHz speech SR show that StreamWSR achieves competitive or superior speech quality and intelligibility compared with representative waveform- and spectrum-based baselines, while maintaining a zero-look-ahead streaming advantage with only 9M parameters and 2G FLOPs.
Read the full abstract
Geophysics
PSDWII: Physical-Structure-Driven Waveform Inversion Imaging
This paper proposes a Physical-Structure-Driven Waveform Inversion Imaging framework (PSDWII), grounded in the recognition that surface-recorded seismic data are directly related to subsurface virtual sources rather than to the model parameters themselves. Based on this recognition, we decompose seismic wave propagation into a three-step physical process: source excitation and incident wave propagation; interaction of the incident wave with heterogeneities, which generates virtual sources with distinct radiation patterns and excites secondary waves; and propagation and reception of the secondary waves. The adjoint of the propagation operator (reverse-time extrapolation) is then used to formulate a linear inversion projection for the virtual sources. By distinguishing different virtual-source expressions and their underlying physical mechanisms, we establish a unified mathematical representation for three inversion tasks: Full Waveform Inversion(FWI), stratigraphic physical properties imaging, and stratigraphic structures imaging. Deconvolution and linear inversion are respectively adopted to achieve multi-parameter decoupling in FWI and in angle-domain common-image gather inversion for stratigraphic physical properties imaging. The PSDWII framework shifts the conventional “mathematical optimization driven” paradigm of waveform inversion to a “physical-structure-driven” one. It neither constructs objective functions nor computes their gradients or Hessian inverses. Within this framework, we develop three specific methods: physical-structure-driven FWI (PSDFWI), physical-structure-driven stratigraphic physical properties imaging with angle-domain common-image gather inversion (PSDSI), and physical-structure-driven stratigraphic structures imaging (PSDMig).
Read the full abstract
Machine Learning
RecurTrace: Adaptive Latent Reasoning with Loop-Time Memory
Repeating a small block of middle layers increases a language model's effective inference depth without adding parameters or generating extra tokens, and recent work shows that this latent recurrence improves reasoning. However, two design choices limit these gains. Each iteration sees only the previous output and cannot directly access earlier computations. Moreover, a fixed loop count wastes depth on easy inputs while leaving hard ones with too little computation. We introduce RecurTrace, which addresses both limitations using the loop's own trajectory. Specifically, Loop Memory Attention lets each looped layer attend to its own states from previous iterations along the loop-time axis, so the model can revisit earlier computations instead of relying on the latest state alone. A halting head then reads the loop state and predicts whether to continue, with supervision from an oracle that identifies when additional depth still reduces loss. In a controlled MathQA comparison on the same looped backbone, RecurTrace achieves 56.9% accuracy with an average of 2.0 loops, exceeding the best fixed loop depth by 2.2 points at matched compute. By comparison, ACT and PonderNet collapse to one loop, and CALM reaches only 54.1% with 5.6 loops, while the stronger LoopUS-Conf and TaH-Mismatch baselines reach 55.3% at 3.2 loops and 55.7% at 2.1 loops. Finally, RecurTrace improves generation accuracy over same-budget fine-tuned baselines at 0.6B, 1.7B, 4B, and 8B, with the gain growing with model size from 0.6 to 3.4 points.
Read the full abstract
Computer Vision
When Depth Hurts: Reliability-Aware Geometry Distillation for Depth-Free RGB-D Salient Object Detection
Depth can resolve appearance ambiguity in RGB-D salient object detection (SOD), yet sensor depth is not uniformly reliable. Missing regions, blurred boundaries, and structural artifacts can propagate through multimodal fusion and make an RGB-D detector less accurate than its RGB-only counterpart. Existing quality-aware approaches regulate observed depth but remain dependent on the same potentially defective modality. We propose \method, a reliability-aware geometry distillation framework developed for RGB-D SOD benchmarks without using dataset-provided depth during training or inference. A frozen Depth Anything V2 model serves only as a training-time teacher, transferring dense relative geometry, hierarchical spatial attention, and boundary structure to a compact edge-aware geometry branch. Pooled bidirectional interaction aligns geometry with appearance, and a pixel-wise reliability estimator selectively injects geometry that is compatible with the current RGB representation. The teacher is removed after training, leaving an RGB-only inference network. Trained on 2,985 RGB-mask pairs, \method{} achieves the best or tied-best result in 26 of 36 metric-dataset comparisons against ten recent RGB-D SOD methods, including a 13.4% relative MAE reduction on ReDWeb-S. When retrained on DUTS-TR, it also improves the strongest prior $F$-measure by 4.2% on PASCAL-S, showing that the distilled geometry transfers beyond a particular sensor or dataset domain. Code will be released upon publication.
Read the full abstract
Machine Learning
SimpleDesign: A Joint Model for Protein Sequence and Structure Codesign
Proteins are fundamental to biological processes, with their function determined by the complex interplay between the amino acid sequence and the three-dimensional structure. Developing generative models capable of understanding this intrinsically multi-modal relationship is crucial for fields like drug discovery and protein engineering. Existing models often rely on a multi-stage training process where autoencoders that tokenize data into latent representations are trained in a first stage. Secondly, a generative model is trained on the latent representation of the autoencoder(s), i.e., generative modeling in a latent space. We hypothesize that this multi-stage training is not necessary to obtain performant co-design models and thus present SimpleDesign, an effective multi-modal protein design model trained directly in the data space. SimpleDesign leverages a single-stage end-to-end objective that combines discrete cross-entropy for sequences and a regression objective for structures. In order to effectively model the difference in sequence and structure modalities, we develop a Mixture-of-Transformer architecture that allows modality-specific processing while keeping global self-attention over both modalities. We train SimpleDesign on over 2M sequence-structure pairs achieving strong performance across co-design and unconditional sequence/structure generation benchmarks.
Read the full abstract
Cryptography and Security
Spruce: Scalable Private Outsourced Retrieval Using Compact Embeddings
Retrieval-Augmented Generation (RAG) has made dense retrieval over large document collections a standard building block. Organizations increasingly outsource vector indexes to untrusted clouds, exposing proprietary corpora and user queries. Cryptographic protection is challenging because each query searches corpus-scale state, causing computation, correlated randomness, and communication to grow with the corpus. At million-document scale, a naive secure implementation takes minutes and about 90 GB of communication per query. Even recent optimized systems require 10–22 seconds. We propose Spruce (Scalable Private Outsourced Retrieval Using Compact Embeddings), which co-designs representations with the cryptographic protocol. Spruce learns compact binary codes that preserve candidates for full-precision reranking, replacing corpus-wide embedding scoring with efficient Hamming-distance computation under two-server multi-party computation (MPC). A corpus-calibrated fixed-radius protocol avoids multi-round candidate selection while preserving retrieval quality. Spruce also provides private cluster pruning, which trades minor quality loss for substantially less computation, and a one-core owner-operated dealer that removes cloud OT preprocessing bottlenecks. Across four corpora containing 383K–5.42M documents, Spruce preserves the original search quality with median candidate sets of only 382–1,952. At 10 Gbps inter-server bandwidth, full scans take 0.21–2.97 seconds, $4.8$–$6.7\times$ faster than the closest measured prior work. Private pruning takes 0.06–1.09 seconds, achieves $13.1$–$22.9\times$ speedups, and retains $93.9\%$–$97.3\%$ of full-float NDCG. On the largest corpus, pruning and the dealer jointly improve sustained throughput by $31.5\times$ at 1 Gbps per link.
Read the full abstract
Algebraic Topology
Generalized Telescope Conjecture
We extend Balmer's tensor-triangular geometry to arbitrary presentably symmetric monoidal $\infty$-categories $\mathcal V$ by introducing the atomic smashing frame. This construction is functorial in $\mathcal V$, recovers the classical Balmer spectrum in the stable compactly-rigidly generated case, and yields a formulation of the telescope conjecture for arbitrary $\mathcal V$. Using dualizable and locally rigid $\infty$-categories, we identify smashing ideals with locally rigid localizations and establish recollement results for smashing frames and atomic smashing frames. As a consequence, we obtain a recollement principle for the telescope conjecture. We also establish descent for the telescope conjecture along several natural classes of coverings. As a principal application to chromatic homotopy theory, we analyze the structure of the smashing frame of $\mathrm{Sp}$. We prove a prime decomposition in terms of the smashing frames of the $p$-local $\infty$-categories $\mathrm{Sp}_{(p)}$ and, at each prime, an embedding into the product of the smashing frames of the monochromatic layers $\mathrm{Sp}_{T(n)}$. In particular, the spatiality of the smashing frame of $\mathrm{Sp}$ reduces entirely to that of $\mathrm{Sp}_{T(n)}$. Beyond the stable setting, we characterize the telescope conjecture for $\infty$-topoi in terms of smashing fields, and for connective module categories and hypercomplete connective sheaves in terms of Pierce-type conditions. Finally, we introduce the Serre smashing frame. Over a connective $\mathbb{E}_\infty$-ring $R$, this frame sits between the atomic and usual smashing frames, providing an intermediate structural layer in the study of the telescope conjecture for the connective $R$-module $\infty$-category.
Read the full abstract
Instrumentation and Methods for Astrophysics
Night-sky emission correction techniques for high-resolution spectroscopy: Demonstration with NIRPS
Ground-based spectrographs operating in the near-infrared (NIR) regime are hampered by various absorption and emission features of Earth's atmosphere. While considerable attention has been paid to mitigating telluric absorption, correcting telluric emission features remains non-trivial and can significantly affect the observation of faint targets. We aim to develop and implement automated algorithms for sky background emission correction in the context of high-resolution spectroscopy. These empirical-based algorithms have been officially integrated into both NIRPS data reduction pipelines: NIRPS DRS and APERO DRS. Designed for flexibility, these techniques can be incorporated into the reduction workflow of any high-resolution spectrograph to improve the radial velocity (RV) performance. In our approach, a reference sky spectrum is first created by deep-stacking NIRPS sky calibration frames on a common wavelength grid and calculating the weighted median flux per pixel, separately for both the high-accuracy and high-efficiency instrument modes. This process is repeated for all spectral orders and for both the object and sky-calibration fibres: fibre A and fibre B, respectively. In this reference sky spectrum, the sky emission lines can be identified and used to construct a static library. During the data reduction process, the emission features in the library are individually scaled in terms of flux using two distinct techniques, each specific to the two DRS pipelines. Finally, they are locally subtracted from the science observations to minimise their noise contribution to the final spectrum. We find that the correction algorithms significantly improve the RV measurements obtained using both the cross-correlation function and line-by-line techniques, enabling NIRPS to achieve submetre-per-second precision in the NIR. The techniques have been successfully validated and demonstrated.
Read the full abstract
Instrumentation and Methods for Astrophysics
3.5-meter Segmented-Mirror Robotic Space Telescope Mission White Paper V. Key Scientific Mission: Compact-Object Time-Domain Science
An isolated compact object retains the point-source resolving power of the space-based slitless spectrograph. The baseline wavelength range is 0.2–1.5 $\mu$m. The planning baseline uses $R \simeq 1000$ for broad and faint transient spectra and reserves selectable bands at $R \simeq 5000$ for accretion-disk profiles, velocity structure, and precision line ratios. Broad features can be measured after binning the native $R \simeq 5000$ data to lower resolution. Rapid-response spectroscopy follows gravitational-wave counterparts and kilonovae from hours to days. Repeated spectra of dwarf novae and compact binaries trace accretion state and orbital phase, while uninterrupted imaging of white dwarfs measures pulsation frequencies. The program combines mission-based monitoring with external alerts, including KGMT transient detections. The instrument study must preserve calibrated throughput to 2.70 $\mu$m and evaluate a 3.0 $\mu$m operational band edge, with 2.5 $\mu$m retained as the formal engineering off-ramp. Mid-infrared imaging is not part of the adopted compact-object baseline.
Read the full abstract
General Topology
On fuzzy contractions in fuzzy metric spaces
The classical Banach principle relies on contractive sequences being Cauchy. Gregori and Sapena (2002) established a fuzzy analogue but required this as an extra hypothesis, leaving it as an open problem. Gregori et al. (2020) later asked whether strictly fuzzy $\psi$-contractive sequences are necessarily Cauchy. We settle both problems negatively via a single counterexample: in a constructed GV-fuzzy metric space, there exists a sequence which is both GS-contractive and strictly fuzzy $\psi$-contractive, yet not Cauchy. This shows that contractivity never implies Cauchyness in general GV-fuzzy metric spaces, and additional structural conditions in fuzzy fixed point theorems are indispensable.
Read the full abstract
Instrumentation and Methods for Astrophysics
3.5-meter Segmented-Mirror Robotic Space Telescope Mission White Paper IV. Key Scientific Mission: Solar-System Small Bodies and Planetary Defense
The baseline 0.2–1.5 $\mu$m observatory provides rapid-response astrometry, visible and near-infrared taxonomy, rotation and phase curves, recovery, and long-arc orbit improvement for near-Earth objects and other small bodies. The instrument study also evaluates calibrated throughput to 2.70 $\mu$m with a 3.0 $\mu$m operational band-edge goal. A reduction to 2.5 $\mu$m remains the formal engineering off-ramp if thermal, detector, cooling, mass, power, or cost constraints require it. The 3.5-meter Segmented-Mirror Robotic Space Telescope does not carry a mid-infrared channel. Coordinated ground-based mid-infrared telescopes provide the thermal fluxes required to infer diameter and albedo, while the space mission supplies contemporaneous reflected-light measurements and observing geometry. The program combines recovery, physical characterization, orbit refinement, and covariance-based hazard assessment. Its CODES dynamics system and OGFinder-to-OpenOrb processing path connect measured astrometry to reproducible orbit solutions and close-approach predictions.
Read the full abstract
Computation and Language
FrameBench:A Language Understanding Benchmark Based on Frame Semantics
In frame semantics, sentence comprehension is assumed to proceed by relating lexical meaning to background knowledge called semantic frames, thereby enabling readers to implicitly enrich the text with unstated information. Recent large language models (LLMs) have achieved strong performance across a wide range of downstream tasks. However, it remains unclear whether they can reproduce the kinds of implicit enrichment that humans naturally make during comprehension. To address this question, we introduce FrameBench, a benchmark grounded in frame semantics. FrameBench consists of multiple-choice questions that test whether models distinguish the frames evoked by the same verb across contexts. We construct the benchmark for English and Japanese using FrameNet-style resources and a generation-and-verification pipeline with native-speaker judgments. Our experiments on a diverse set of models reveal challenges for small models, while several large models surpass the human reference scores. We release the constructed FrameBench dataset and the code for dataset construction and evaluation at https://github.com/SasanoLab/FrameBench.
Read the full abstract
Information Retrieval
HypRQ-VAE: Hyperbolic Item Indexing for Long-Tail-Aware Generative Recommender Systems
Sequential recommender systems model user behavior as item ID sequences, while recent generative methods cast recommendation as a language modeling task using large language models (LLMs). While this paradigm incorporates rich textual semantics, it introduces a fundamental mismatch: LLMs operate on text tokens, whereas recommender systems depend on discrete item indices. This misalignment often leads to hallucinations in generative recommendations. Existing methods attempt to bridge this gap by learning item vocabularies in Euclidean space, but they struggle to model the inherent long-tail distribution of real-world catalogs, where a small number of head items dominate, and a vast number of tail items reflect users' niche preferences. To address this issue, we introduce Hyperbolic Residual-Quantized Variational AutoEncoder (HypRQ-VAE), the first framework to learn item indexing in hyperbolic space. HypRQ-VAE leverages the unique properties of hyperbolic geometry, whose exponential volume expansion naturally accommodates the power law structure of user-item interactions. This allows the model to encode rich textual semantics while preserving the representational fidelity of sparse, long-tail items. Experiments on three benchmark datasets show that HypRQ-VAE significantly improves the performance of recommendation, particularly in recommending tail items. Our analysis attributes these gains to the superior capacity of hyperbolic space to model item hierarchies and sparsity in generative recommendation. Our code and data are available at: https://github.com/wulongfeng/HypRQ-VAE.
Read the full abstract
High Energy Physics (Phenomenology)
Leading-Neutron Electroproduction at HERA and the EIC: Sullivan Process, Target Fragmentation, and Pion PDFs
Leading-neutron electroproduction measurements from the H1 and ZEUS experiments at HERA have been used to constrain pion parton distribution functions (PDFs) at small momentum fractions within the Sullivan one-pion-exchange (OPE) framework, complementing large-$x_\pi$ constraints from pion-induced Drell–Yan measurements. Previous analyses have focused primarily on the region of large neutron longitudinal momentum fraction, $x_L$, where contributions from deep-inelastic scattering (DIS) target fragmentation are suppressed. In this work, we use the Pythia event generator to model the target-fragmentation contribution and show that its combination with the OPE contribution reproduces the main features of the HERA leading-neutron data over the full measured $x_L$ range without introducing additional ad hoc normalization factors. This result demonstrates the potential of incorporating a broader range of leading-neutron data into future global analyses, thereby extending sensitivity to smaller pion momentum fractions $x_\pi$. We investigate the model dependence associated with the pion–nucleon vertex form factor and show that the HERA data are sensitive to different target-fragmentation treatments implemented in Pythia. Finally, we present projections for leading-neutron production at the future U.S. Electron-Ion Collider (EIC), identifying beam-energy configurations and kinematic regions that provide enhanced sensitivity to pion structure while suppressing DIS target-fragmentation contributions.
Read the full abstract
Astrophysics of Galaxies
Evolution of Virial Clouds - II: From the Formation of First Stars up to their Explosion
The existence of cold gas and dust clouds close to the cosmic microwave background (CMB) temperature was proposed as a potential repository for a significant fraction of the missing baryons in galactic halos. While the evolution of the virial clouds from the last scattering surface (LSS) ($z=1100$) up to the formation of Population III (Pop III) stars ($z=48$) was studied in [N. Tahir, A. Qadir, M. Sakhi, & F. De Paolis, The Euro. Phys. Jour. C 81, 827 (2021)] (Paper I), the subsequent evolution during the epoch of the first stars remained unexplored. In the present work, we investigate the second evolutionary phase of the virial clouds, covering the redshift range $48 \gtrsim z \gtrsim 10$. We develop a comprehensive model incorporating the coupled thermal, chemical, and dynamical evolution of the cloud, including CMB heating, adiabatic compression due to gravitational contraction, H$_2$ and metal-line cooling, and feedback from Pop III supernovae. Solving the coupled ordinary differential equations (ODE) numerically, we find that the cloud cools monotonically from $137$~K to $41.5$~K, contracts from $120$~pc to $89.8$~pc, and becomes enriched in metals and molecular hydrogen, while the cooling rate dominates the heating rate throughout the evolution.
Read the full abstract
Computation and Language
Accountable AI with Grounded, Faithful, Consistent, Actionable Rationales: A Case Study in Clinical Trial Matching with VERDICT
Accountability means a decision can be examined, justified, and contested. LLMs make this hard: fluent output may be ungrounded, incomplete, or unfaithful to the decision process. Achieving accountability requires verified rationales (how was the decision reached), assumptions (what was assumed rather than known), policy consistency (the same treatment for the same facts), and pivotal conditions (what would change the outcome). We introduce self-faithfulness as an automatic test of accountability: changing the pivotal conditions should change the decision. We examine accountable AI through clinical trial matching, a high-stakes task central to evidence-based medicine. Although LLM-based matchers match patients to trials reasonably accurately, they apply decision policies inconsistently and produce rationales that are unfaithful to their own decisions. We introduce VERDICT, an LLM-based agent that translates a decision task, its constraints, and its policy into Satisfiability Modulo Theories (SMT), then derives the decision with SMT and MaxSMT solvers – so policies are applied consistently and decisions are accountable by construction. Across a SIGIR 2016-derived dataset and TREC 2021, VERDICT achieves the strongest decision accuracy among LLM-only and neurosymbolic baselines, applies policies with perfect consistency, and produces clinician-preferred rationales grounded in explicit assumptions and pivotal conditions, with improved counterfactual self-faithfulness.
Read the full abstract
Representation Theory
Singular equivalences of $n$-adjoint type and standard eventually homological isomorphisms
We refine the definition of eventually homological isomorphisms so that it becomes more natural and fruitful. We introduce singular equivalences of $n$-adjoint type which can be viewed as standard singular equivalences of different levels and are more convenient for reducing various homological conjectures and transferring homological properties. We develop the theories of singular equivalences of $n$-adjoint type for $n\ge 2$ and standard eventually homological isomorphisms between derived categories by giving their characterizations in terms of $\otimes$-(co)perfect complexes of bimodules, two constructions via opposite algebras and tensor product algebras, examples induced by homomorphisms of algebras and idempotents, and applications to the reduction of homological conjectures such as the finitistic dimension conjecture, Gorenstein symmetry conjecture and Auslander-Reiten conjecture, as well as to the preservation of homological properties such as Gorensteinness, the Fg condition, Happel's property and Han's property. Moreover, we show that essentially surjective standard eventually homological isomorphisms between derived categories correspond to standard singular equivalences with canonical left adjoint.
Read the full abstract
Atmospheric and Oceanic Physics
Added value of dynamical downscaling in sub-seasonal tropical cyclone forecast
Improving the sub-seasonal forecast of tropical cyclones (TCs) remains a significant challenge for climate models. This study evaluated the characteristics of tropical cyclones (TCs) in sub-seasonal forecasts using the Global Seasonal Forecast System 6 (GloSea6), an operational seasonal-to-sub-seasonal forecasting model operated by the Korea Meteorological Administration (KMA) during June-September (JJAS) from 1993 to 2016 over the western North Pacific (WNP). GloSea6 was found to underestimate TC frequency, tracks, particularly in mid-latitudes, lifetime, and intensity across all months, with the most significant errors occurring in August. To address these deficiencies, we examined whether applying dynamical downscaling to GloSea6 during August 2016, a period characterized by the lowest TC forecast skill in GloSea6, could improve sub-seasonal TC forecasts. The application of dynamical downscaling demonstrated added value by directly improving the simulation of TCs in terms of frequency, structure, intensity, and lifetime, although slight overestimations were observed. Furthermore, improved reproductions of the Indian monsoon and circumglobal teleconnection (CGT), both of which strongly influence the western North Pacific subtropical high (WNPSH), contributed to more accurate forecasts of WNPSH variability and, consequently, better predictions of TC activity in the mid-latitudes and East Asia. Therefore, dynamical downscaling can significantly advance sub-seasonal TC forecasts by not only directly improving TC characteristics but also indirectly enhancing the environmental fields (e.g., WNPSH, Indian monsoon, and CGT) that are critical to TC activity.
Read the full abstract
Sound
PACodec: A Low-bitrate Neural Speech Codec with Parallel Additive Vector Quantization
This paper proposes PACodec, a novel low-bitrate neural speech codec based on parallel additive vector quantization (PAVQ). Unlike the mainstream residual vector quantization (RVQ) used in most neural speech codecs, where vector quantizers (VQs) are sequentially dependent, the PAVQ strategy adopted in PACodec aggregates parallel quantization results to optimize bitrate usage. Specifically, the PAVQ adopts a "global-local-global" (GLG) design: the global encoded features are quantized in parallel by multiple independent VQs, each attending to a local component of the representation, and their outputs are aggregated through addition to yield the final global quantization result for decoding. Experimental results show that PACodec, as each VQ focuses only on local information, supports smaller codebooks and reduces bitrate by 30% compared with baselines at the same decoding quality, with only minor model complexity. Further analysis shows that, owing to the GLG framework of PAVQ, the proposed PACodec is disentanglement-friendly, and each independent VQ captures different aspects of speech, e.g., content, timbre, and acoustic details, suggesting potential for application to downstream tasks such as voice conversion.
Read the full abstract
Robotics
ARTiS: An Adaptive Robotic Gripper for Enhanced Tool Manipulation in Disassembly Applications
Grasping and holding tools while using them presents a considerable challenge not only for robots but also for humans. Such a challenge is particularly noticeable in processes involving assembly and disassembly, where efficiency and consistency depend on performing rapidly adaptive tasks. Nonetheless, contemporary robotic grasping technologies that can securely manipulate tools during operation frequently have significant constraints. In this paper, introduce ARTiS (Adaptive Robotic Tool Gripper in Disassembly Systems), a novel gripper that combines the adaptability of soft grippers, the dexterity of anthropomorphic hands, and the robustness of rigid mechanisms with a soft palm and fingertips. This unique combination makes it possible to hold tools securely in a variety of situations through using active jamming in the palm and fin-ray adaptation in fingertips. Furthermore, high finger dexterity is achieved through the seven degrees of freedom design, which enables the fingertips to orient to any surface, both for automated solutions and collaborative tasks. A comprehensive evaluation was conducted using a range of conventional disassembly tools to assess the gripper's compliance, durability, and functional versatility. More information, hardware instructions, and videos at https://romanmykhailyshyn.github.io/artis/
Read the full abstract
Algebraic Geometry
Grassmann–Plücker Parametrization of Convolutional Filter Subspaces: Regularity and Closed Embeddings
We propose a geometric parametrization of the filters in a single convolutional layer: the parameter is no longer an ordered family of filter vectors, but a fixed-dimensional subspace of the filter space. For one-dimensional finite-stride convolution, the filter-to-convolution-operator correspondence gives an injective linear map $\mathcal{C}:\mathcal{K}\to H$. This map sends filter subspaces in $\mathrm{Gr}(q,\mathcal{K})$ to operator subspaces in $\mathrm{Gr}(q,H)$; composing it with the Plücker embedding yields a projective parametrization $\Phi:\mathrm{Gr}(q,\mathcal{K})\to\mathbb{P}(\bigwedge^q H)$. Using $T_U\mathrm{Gr}(q,\mathcal{K})\cong\mathrm{Hom}(U,\mathcal{K}/U)$, we compute the differential of the induced Grassmannian map and show that the differential of $\Phi$ is injective at every point. We then use the vanishing equations for Plücker coordinates and standard affine coordinates on a Grassmannian to prove that $\mathrm{Gr}(q,\mathcal{C}(\mathcal{K}))\hookrightarrow\mathrm{Gr}(q,H)$ is a closed embedding, and hence that $\Phi$ is a closed embedding. Consequently, the parameter space is isomorphic to its projective image, the parametrization is finite and birational onto its image, every fiber is a singleton, and the resulting projective neural variety is smooth. For $k=4$ and $q=2$, we also use Singular to recover the image ideal and check its dimension, degree, chart rank, and smoothness. This computation illustrates, rather than replaces, the general proof. Finally, we discuss possible connections with filter redundancy and low-rank convolution, while distinguishing the proved geometric results from application proposals requiring numerical validation.