Complex Variables
Ample vector bundles without Griffiths-semipositive metrics
Yun-Heng Du and Song-Yan Xie
We construct a rank-two ample holomorphic vector bundle on an abelian surface that admits no smooth Griffiths-semipositive Hermitian metric, thereby giving a counterexample to the Griffiths
conjecture. The proof combines a
curvature obstruction that persists under deformations with the openness of ampleness in an irreducible
moduli space of stable sheaves.
Read the full abstract ▾
Representation Theory
Representation-tame Geiß-Leclerc-Schröer algebras and a revised GLS conjecture on root systems
Qiang Dong, Zengqiang Lin, Ming Lu et al. (+1 more)
Using Galois covering theory and equivariant techniques, we classify all connected representation-tame Geiß–Leclerc–Schröer (GLS)
algebras in terms of their defining triples $(C,D,\Omega)$. We then study the GLS
algebras $H(\widetilde{CD}_n)$ with minimal symmetrizers, where $n\geq 2$ and $\widetilde{CD}_2=\widetilde{B}_2$. By realizing these
algebras as basic
algebras of $\mathbb{Z}_2$-skew group
algebras of the string
algebras $H(\widetilde{C}_{2n-2})$, we introduce extended strings and extended bands to parameterize the connected components of the Auslander–Reiten quivers of $H(\widetilde{CD}_n)$ and determine their shapes. We further classify the indecomposable $\tau$-locally free modules over representation-tame GLS
algebras of affine type and show that their rank vectors form precisely the set of positive roots together with explicitly described non-root vectors in the positive cone of the root
lattice. This yields a precise revision of the GLS
conjecture for representation-tame GLS
algebras of affine type.
Read the full abstract ▾
All authors Qiang Dong, Zengqiang Lin, Ming Lu, Shiquan Ruan
Materials Science
Deep Generative Crystal Structure Prediction: A Benchmark Study and a Controlled Test of Prototype Dependence
Lai Wei, Rongzhi Dong, Ying Feng et al. (+2 more)
Deep generative models are widely reported to enable de novo
crystal structure prediction (CSP), but their capability has not been measured consistently against template-based methods. We evaluate 12 representative generative CSP models, spanning latent-variable,
diffusion, flow-matching, autoregressive, and
manifold random-walk architectures, against TCSP 2.0 on 180 test structures and a leakage-controlled subset of 46. All methods use identical structure-matching,
symmetry, and consensus criteria. Template
retrieval is the strongest single method, reaching 68.3% top-1 success;
symmetry-aware EquiCSP (66.4%) and Uni-3DAR (62.9%) form the next tier. However, comparison with TCSP 2.0 shows that most structures correctly predicted by generative models are also correctly predicted by template substitution. Thus, the set of structures uniquely reachable by generation is small, limiting its practical advantage for discovering structures outside existing prototype libraries. To test the source of this performance, we removed entire stoichiometric prototype families from the training set and retrained the strongest generative model. Accuracy declined by 50-78% across four families, establishing that performance is substantially prototype-dependent. A small minority of structures survived removal of their prototype family, demonstrating a real but limited
retrieval-independent predictive capacity. Present generative CSP models therefore function largely as implicit, softer-edged prototype libraries rather than genuinely de novo predictors. Enlarging this residual capacity, rather than aggregate match rate alone, is the central open problem.
Read the full abstract ▾
All authors Lai Wei, Rongzhi Dong, Ying Feng, Madeline Miklos, Jianjun Hu
Statistics Theory
Inference for Hüsler-Reiss block models
Capel Alexandre, Demangeot Marine, Meyer Nicolas et al. (+1 more)
Estimating the Hüsler-Reiss precision
matrix is a fundamental problem for statistical inference in multivariate extremes. In high-dimensional settings, the number of unknown parameters grows quadratically with the dimension, making regularisation indispensable. Existing approaches regularise the estimation problem by exploiting sparsity. In this paper, we consider an alternative structural assumption, namely that the Hüsler-Reiss precision
matrix is block-structured. To estimate such a
matrix, we introduce a new regularisation framework based on a convex fusion penalty. By encouraging rows and columns to merge, this approach provides a parsimonious representation of the precision
matrix, allowing for the simultaneous estimation of its coefficients and the underlying partition of the variables. The resulting convex optimisation problem is solved by an efficient
algorithm combining gradient-based updates with progressive fusion steps. We establish non-asymptotic concentration bounds for the empirical weights entering the penalty and prove consistency of both block recovery and precision
matrix estimation under suitable regularity conditions. Numerical experiments demonstrate that our methodology accurately recovers the latent block structure while accurately estimating the Hüsler-Reiss precision
matrix across various configurations, illustrating the practical benefits of fusion-based regularisation for multivariate extremes. These benefits are also demonstrated by applying the proposed method to foreign exchange data.
Read the full abstract ▾
All authors Capel Alexandre, Demangeot Marine, Meyer Nicolas, Toulemonde Gwladys
Number Theory
Stable base change from unitary groups in three variables and integral relation of automorphic periods
Tristan Ricoul
Let $E$ be a real quadratic field and let $U_E$ be the quasi-split
unitary group in three variables associated with $E$. We prove a $p$-adic divisibility between the automorphic periods of a cuspidal automorphic representation $\pi_U$ of $U_E$ and the periods of its Rogawski stable base change to $\mathrm{GL}_3(E)$. This generalizes earlier works of Tilouine-Urban and Hida in the case of $\mathrm{GL}_2$. The divisibility we prove involves a new kind of automorphic periods for self-conjugate cuspidal automorphic representations of $\mathrm{GL}_3(E)$, defined within the middle degree of the cuspidal
cohomology rather than the top or bottom degrees. Moreover, we prove an à la Hida adjoint $L$-value formula for $\mathrm{GL}_3(E)$, relating these newly defined middle-degree periods to the usual top and bottom automorphic periods. Finally, we also prove a similar formula for $U_E$, which is the first instance of such a result for quasi-split
unitary groups.
Read the full abstract ▾
Robotics
Generalizing Manipulation Skills with a Local Coding Agent
Raman Talwar, Elias Nijs, Andreas Verleysen et al. (+1 more)
Today, progress in open-weight
language models enables systems capable of writing, executing and debugging code while still running on a single workstation. Most
language-driven
robots give the model a fixed action interface or a trained policy. Generalizing to a new task therefore means more engineering effort or more data collection, both time-consuming. We investigate whether a local open-weight
vision-
language model can control a
robot and one-shot generalize to new variations of a task without new human programming or training. We let a local open-weight VLM, Qwen3.8-27B, drive a UR3e robotic arm from a coding-
agent harness. It writes and runs its own code above a service that implements kinematics, safety limits and classic computer
vision techniques. We investigate if this system is capable of generalizing to unseen tasks. Specifically, we test it on nine tasks built from children's toys designed to probe generalization capability across various object characteristics: color, size, shape, and task variation of those. With five trials for each task, we observe generalization in 30 out of 45 trials with durations ranging from 3.4 to 67.5 minutes depending on task complexity. We further test if there is a speedup when an
agent is asked to redo the task after successful completion. This resulted in a 50% reduction in duration, indicating that there is self-improvement over time. Finally, we expose the limitations of a local coding
agent. We believe that solving those limitations combined with further investigation of self-improvement over time points at a direct path toward real-world deployment of a local coding
agent.
Read the full abstract ▾
All authors Raman Talwar, Elias Nijs, Andreas Verleysen, Francis wyffels
Representation Theory
Brauer webs of finite and affine type
Yaolong Shen, Linliang Song and Weiqiang Wang
We introduce and develop an integral diagrammatic framework for two new monoidal categories, the finite and affine B-web categories. They arise naturally from representation categories of type BCD Lie
algebras, and can be viewed as a Brauer type generalization of the affine web category which we introduced earlier. Integral bases consisting of bubbled elementary diagrams are established for the finite and affine B-webs. A simplified presentation of these categories over a field of characteristic zero is also obtained.
Read the full abstract ▾
Algebraic Topology
Chromatic Purity of Dualizable Categories
Yifan Jin
We develop the chromatic theory of dualizable stable categories, with chromatic purity for continuous $K$-theory at its center. We generalize the chromatic purity
theorem for algebraic $K$-theory to continuous $K$-theory of dualizable stable categories and reformulate it as the purity of $H$-unital rings. Using categorical completion theory, we construct chromatic fracture squares and obtain a refinement of chromatic purity: for every dualizable stable category $\mathcal{C}$, the $T(n)\oplus T(n-1)$-completion map induces an equivalence $$ K_{T(n)}^{\mathrm{cont}}(\mathcal{C})\xrightarrow{\simeq}K_{T(n)}^{\mathrm{cont}}(\mathrm{Nuc}_{T(n)\oplus T(n-1)}(\mathcal{C})). $$ We also establish chromatic descent for continuous $K$-theory of dualizable homotopy fixed points and prove a categorical
nuclear refinement of descent. This framework allows us to lift
redshift bounds, Tate vanishing, and blueshift from spectra to dualizable stable categories. We show that the $T(n)$-completion of a rigid symmetric monoidal stable category is equivalent to the dualizable limit of module categories over a tower of type $n$ generalized Moore spectra. As applications, we prove that both
nuclear solid and
nuclear gaseous module categories satisfy chromatic
redshift.
Read the full abstract ▾
Operator Algebras
Coarse rigidity in von Neumann algebras via derivations
Manish Kumar and Melchior Wirth
Motivated by the interaction between amenability and first
cohomology into tempered representations of groups, we propose a notion of relative coarse
rigidity for $\sigma$-finite von Neumann
algebras through modular derivations and their associated GNS-symmetric
quantum Markov
semigroups. This extends Peterson's framework of $L^2$-
rigidity to the nontracial setting. Our main result shows that the absence of relatively amenable summands forces the associated deformations to converge on relative commutants in ultrapowers. For diffuse von Neumann
algebras with separable predual, this yields a dichotomy between the presence of an amenable summand and
rigidity in terms of central sequences. Combining this principle with mixing and compactness, we provide a novel unified approach for proving indecomposability properties such as fullness, primeness and stable solidity. Applications yield new examples of relatively $\omega$-solid and relatively stably solid type III von Neumann
algebras arising from amalgamated free products, operator-valued free Araki-Woods constructions, cocycle crossed products, and twisted nonsingular discrete measured groupoids.
Read the full abstract ▾
Statistical Methodology
Flexible latent variable models on graphs: Laplace approximated inference for multiview network data
Anna van Es, Eva Cantoni and Davide La Vecchia
We propose a novel and flexible nonlinear approach for dimensionality reduction of large-scale multiview network data and derive its theory. The (linear) predictor incorporates observed
covariates (edge-specific, layer-specific, and global) and Gaussian latent factors. Inference is conducted via the graph Laplace approximated maximum likelihood
estimator. Letting $K$ denote the number of network layers and $n_V$ the number of nodes, we derive asymptotic theory under two regimes: (i) $K \to \infty$ with fixed $n_V$, and (ii) double asymptotics $K, n_V \to \infty$, establishing consistency and asymptotic normality for local and global parameters, with distinct convergence rates. In an application to the
gravity model for commodity trades, we use a zero-adjusted Gamma distribution with latent factors and observable
covariates (e.g.\ distance, tariffs, common
language) to capture excess zeros, skewness, and unobserved heterogeneity. Synthetic and real-data exercises show that our approach outperforms the routinely applied Poisson pseudo-maximum likelihood
estimator with fixed effects. We complement our theoretical and empirical contributions with open-source R/C++ routines and a novel strategy for starting values selection.
Read the full abstract ▾
Signal Processing
A Stochastic Doppler Sensing Framework for Range-Resolved Bidirectional Particle Dynamics in a Bubbling Fluidized Bed at 60 GHz
Tworit Dash, Dingyang Wang, Reddy Madhuri Manila et al. (+2 more)
The problem of observing particle motion inside an optically opaque, dense bubbling fluidized bed is addressed. Differential-pressure measurements identify the bulk fluidization state but do not reveal where particles move or how their velocities vary with height. A commercial 60 GHz frequency-modulated continuous-wave radar is mounted above the facility to obtain depth-resolved measurements without optical access or an intrusive probe. Each measured depth
cell contains an ensemble of 600 micrometre sand grains in the Rayleigh-Mie scattering transition region. Particle displacement,
cell exchange, and changes in shadowing alter echo phases and amplitudes. Their coherent sum is therefore a
stochastic complex signal, whose covariance and Doppler spectrum describe ensemble motion. A statistical spectral analysis retrieves total power, mean line-of-sight radial velocity, and Doppler spectral width over range and time. Eight-minute measurements reveal an expanding radar-detectable layer and alternating motion towards and away from the radar. Positive- and negative-mean regions coexist during developed fluidization and exchange their contribution to particle power. Most detectable spectra contain one dominant Doppler lobe; a small subset supports two distinguishable lobes, including cases with one negative and one positive mean velocity. The 5-s total-power trend has a correlation of 0.734 with measured pressure drop, whereas the local fitted spectral-width trend has a weaker correlation of 0.152. This work develops a range-resolved statistical Doppler measurement that reveals bidirectional particle motion and determines when multiple local velocity contributions are distinguishable.
Read the full abstract ▾
All authors Tworit Dash, Dingyang Wang, Reddy Madhuri Manila, Wiebren de Jong, Johan T. Padding
Signal Processing
Adaptive Predictive Sampling and Communication for Real-time Monitoring
Erfan Delfani and Nikolaos Pappas
We propose an adaptive predictive sampling and communication framework for real-time remote monitoring over packet-erasure channels. Rather than relying on a prescribed dynamical model or continuous sensing, our model-free method uses Taylor-expansion-based prediction to adaptively estimate local signal dynamics and proactively schedule the next sampling times. An analytical compensation factor accounts for packet losses to satisfy a prescribed reconstruction-error level. We develop adaptive sample-wise and episodic policies and evaluate their performance against event-driven and uniform sampling and communication baselines.
Read the full abstract ▾
Computer Vision
Radiomics-Conditioned Modulation of RenalCLIP Features for Clear Cell Renal Cell Carcinoma Classification
Yuan Liang, Sourav Bhattacharjee and Abraham Campbell
Radiomics provides quantitative descriptions of tumour appearance that may complement disease-specific foundation models in small labelled cohorts. We investigate this complementarity for computed
tomography-based classification of clear
cell renal
cell carcinoma. Our framework uses radiomics to modulate RenalCLIP features through feature-wise linear modulation (FiLM), while retaining a direct radiomics contribution. Internal testing and external validation compare it with conventional fusion strategies and reference classifiers. The FiLM model achieves an area under the receiver operating characteristic curve (AUC) of 0.804 internally and 0.854 externally, with the highest mean AUC among the evaluated RenalCLIP fusion strategies in both cohorts. Pathway ablations examine the contributions of conditional modulation and the direct radiomics residual, while feature permutation highlights the role of tumour texture. These findings support radiomics as a useful complement to RenalCLIP in a small labelled cohort and identify FiLM as an effective approach to integrating their representations for robust renal tumour classification.
Read the full abstract ▾
High Energy Physics (Phenomenology)
Nonthermal fixed point for massless scalar field theories
Viktoria Noel, Chuang Huang, Aleksandr N. Mikheev et al. (+1 more)
We compute universal scaling exponents and scaling functions for massless scalar field theories far from
equilibrium. Using a large-$N$ expansion to next-to-leading order, we investigate the nonperturbative infrared behaviour associated with the transport of a) energy or b) effective particle number towards low frequencies $\omega$ and momenta $\bar p$. For massless dispersion $\omega = \bar{p}$, we determine the scaling form of the distribution function whose universal power-law tail $f_S(\bar p) \sim 1/\bar{p}^\kappa$ is found to be a) $\kappa = d+1$ and b) $\kappa = d$ in $d=3$ spatial dimensions. Our results establish that relativistic and nonrelativistic theories do not belong to the same universality class in this case, which opens up new applications to far-from-
equilibrium phenomena in high-energy and (
quantum) many-body physics with linear dispersions.
Read the full abstract ▾
All authors Viktoria Noel, Chuang Huang, Aleksandr N. Mikheev, Jürgen Berges
Robotics
Benchmarking Robots for Everyday Environments: From Lab Experiments to Real-World Operations
Raphael Memmesheimer, Martina Overbeck, Dominik Beyer et al. (+19 more)
This study presents an interdisciplinary process for benchmarking service
robots in public everyday environments, addressing the gap between laboratory metrics and real-world operation at technology readiness levels 6-7. Three
robots were evaluated in three use cases: outdoor park cleaning, pedestrian underpass cleaning, and interactive assistance in a public library. Over three years (2023-2025), seven benchmarking events (a consensus workshop and six on-site evaluations, two per use case) assessed the
robots in increasingly realistic settings, from controlled conditions to public operation. An interdisciplinary panel of experts in robotics, human-
robot interaction, safety, and economic viability consolidated use-case-specific metrics into a shared evaluation concept refined across three phases. Rather than ranking the
robots, the evaluation examined how task fulfillment, interaction quality, safety, and economic viability interact in the transition from laboratory prototype to operational system. Direct quantitative comparability across the heterogeneous use cases remained limited, characteristic of benchmarking at this maturity level. The process yielded cross-case insights: increasing environmental realism introduces challenges that uniform metrics capture only in part; the interaction-quality guideline works best as a modular instrument adapted to each use case; safety was assessed by occupational-safety and product-certification experts who adapted industrial test procedures for public settings; and economic viability proved a cross-cutting dimension. We conclude that benchmarking at higher technology readiness levels is well served by a shared procedure for selecting and documenting metrics across these four categories, with safety as mandatory, rather than by a fixed set of metrics: generalizability lies in the procedure rather than in identical metrics.
Read the full abstract ▾
All authors Raphael Memmesheimer, Martina Overbeck, Dominik Beyer, Björn Kral, Sabine Bellmann, Sven Schneider, Jan Zimmermann, Anna-Maria Meer, Medina Klicic, Simone Roth, Carolin Straßmann, Alexander Arntz, Marlene Wessels, Johannes Kraus, Paul Schweidler, Tristan Schnell, Christoph Zimmermann, Benedikt Pulver, Wilhelm Stork, Martin Gersch, Sven Behnke, Arne Roennau
Computation and Language
Calibration as a First-Class Criterion in LLM Evaluation
Mario Sanz-Guerrero and Katharina von der Wense
Calibration of
language models – the
alignment between expressed or implicit confidence and empirical correctness – is a well-studied subfield within NLP. Methods to measure it already exist. The problem is adoption: outside this subfield, NLP research regularly introduces new models,
datasets, and
benchmarks without checking whether the model's confidence scores are meaningful. We argue that this adoption gap is a major obstacle to trustworthy LLM evaluation. Miscalibration causes problems in two distinct areas: at deployment, where overconfident mistakes cause real harm, and inside the research pipeline, where methods like LLM-as-a-judge, synthetic data generation, and active learning rely on calibrated confidence without verifying it. Standard calibration metrics only require two inputs per example: a confidence score and a correctness judgment. Most
benchmarks in use today already provide both, meaning calibration can be reported immediately. For open-ended generation, however, defining these two inputs is still an open challenge. We argue that each NLP subfield should pair its main performance metric with a calibration score and call for treating calibration as an essential property of every model rather than a niche topic.
Read the full abstract ▾
Computation and Language
Spoken Language Models that Think Aloud
Junyi Ao, Kainan Peng, Mingbo Ma et al. (+10 more)
While Chain-of-Thought (CoT)
reasoning has improved the capability of
language models, directly applying it to Spoken
Language Models (SLMs) may introduce long silent intervals under the serial "think-then-speak" paradigm, disrupting real-time spoken interaction. To address this issue, we propose an asynchronous think-aloud framework for
reasoning-based SLMs within the Thinker-Talker architecture. The framework maintains a primary
reasoning stream for logical deduction and a lightweight think-aloud stream that generates short, task-grounded progress utterances conditioned on the user input and the evolving
reasoning state. A dynamic balance strategy coordinates the two streams at runtime, triggering additional think-aloud
speech to avoid silent gaps and canceling pending utterances when the final response becomes ready. Experiments on spoken
reasoning and question-answering
benchmarks show that our approach substantially reduces user-audible silence during
reasoning while maintaining answer accuracy comparable to that of a serial "think-then-speak" baseline, demonstrating the potential of asynchronous think-aloud for responsive interaction in SLMs.
Read the full abstract ▾
All authors Junyi Ao, Kainan Peng, Mingbo Ma, Shun Zhang, Zhenyu Tang, Xutai Ma, Xiang Li, Yinghao Li, Yuancheng Wang, Zhizheng Wu, Haizhou Li, Qing He, Xubo Liu
Machine Learning
When Recursive Models Finish Computing
Hare Krishna, Shubham Singh, Stephen Ebert et al. (+1 more)
Recursive models can continue updating their latent states beyond their nominal inference budget, so an incorrect output at that budget does not show whether computation is unfinished or has entered a persistently unsuccessful regime. We study the dynamics of completion in
attention- and MLP-based Tiny Recursive Models (TRMs) on 1,000 hard Sudoku puzzles. Extending recurrence from the nominal 16 steps to 512 steps increases cumulative exact-solve accuracy from 59.2% to 87.5% for the
attention model and from 74.4% to 91.9% for the MLP model, solving more than two-thirds of the puzzles unsolved in the nominal budget. Across both architectures, latent-state motion drops sharply after the first exact solution. Completed states are typically locally contractive along the trajectory direction, even though the same local Jacobian retains strongly expanding directions. We characterize this phenomenon as trajectory-conditioned anisotropic stability. Perturbation experiments confirm this directional stability across both models. The multi-step fate of the maximally expanding direction differs: it is absorbed within 16 steps in the
attention model but persists longer in the MLP model. The anisotropic-stability pattern also holds for a second
attention checkpoint. Together, these results distinguish nominal-budget failure from completed computation and identify a common dynamical signature of completion across two recurrent architectures.
Read the full abstract ▾
All authors Hare Krishna, Shubham Singh, Stephen Ebert, Hao-Yu Sun
Audio and Speech Processing
Not Quite My Tempo: Voice Activity-aware Speech Synthesis for Lip-Synchronous Dubbing
Alejandro Pérez-González-de-Martos, Florian Lux, Angelina Elizarova et al. (+3 more)
Automatic lip-synchronous dubbing requires a
speech synthesis model to generate alternating voice and silence patterns in the target
language that match the timing of the source clip precisely to ensure an optimal viewing experience. Prior works address this problem by conditioning the
speech synthesis process on lip movements extracted from the
video signal. In this work, we condition the
speech generation on a binary voice-activity signal, which has a lightweight representation and can be produced in multiple ways. We show that the model follows the voice-activity signal with high accuracy while maintaining natural prosody and semantically appropriate pause placement within sentences, as demonstrated through extensive objective and subjective evaluations. By randomly masking this condition during training, we make the feature entirely optional during inference, allowing editors to enforce or relax lip-sync constraints when desired.
Read the full abstract ▾
All authors Alejandro Pérez-González-de-Martos, Florian Lux, Angelina Elizarova, Milana Shkhanukova, Andreas Kellner, Mattia Antonino Di Gangi
Strongly Correlated Electrons
Extraordinary Lifetime Enhancement of Coherent Phonon-Amplitude Modes in the Excitonic Insulator Phase of Ta2Pd3Te5
Anjan Kumar N M, Shuhan Wang, Snehashish Chatterjee et al. (+9 more)
The excitonic
insulator (EI) is an electronic phase of condensed
excitons. However, in many prototypical materials the presence of a concurrent structural transition complicates the identification of a purely electronic origin of the ordered state. Here, we investigate Ta2Pd3Te5, in which an EI phase develops below TC ~ 100 K in the absence of a detectable structural transition. We characterize the excitonic condensate via its coherent
phonon-amplitude response. Most strikingly, an extraordinarily strong lifetime enhancement of these modes sets in below TC which we establish as a new and robust fingerprint linked to
exciton condensation. We discuss possibilities to capture the enhancement when taking into account a coupling of excitonic and
lattice effects in line with the coupled
phonon-amplitude response. Within an order-parameter polaron picture coupling to the excitonic state may dress the
phonon mode and thereby suppress its relaxation and a phenomenological model combines the anharmonic
phonon response and a reduced electronic scattering due to the gap opening in the EI phase.
Read the full abstract ▾
All authors Anjan Kumar N M, Shuhan Wang, Snehashish Chatterjee, Yan Zhu, MinJae Kim, Tobias Ritschel, Elaheh Sadrollahi, Jochen Geck, Achim Rosch, Chandra Shekhar, Claudia Felser, Stefan Kaiser
Computer Vision
From Token Importance to Conditional Removability: Rethinking Visual Token Pruning in Multimodal Large Language Models
Shengli He, Yongchao Liang, Roumeng He et al. (+5 more)
Training-free visual-
token pruning often uses
token importance, redundancy, or related selection criteria as proxies for safe removal. We show that these signals alone do not fully characterize removability, which is conditioned on both representation depth and the surrounding deletion set. Controlled interventions demonstrate that removing the same
tokens at different depths produces substantially different downstream perturbations, while changing only the deletion context at a fixed depth alters candidate marginals and
pruning-boundary decisions. These findings show that
token importance alone cannot determine when a
token is safely removable or how its removability changes under joint deletion. Motivated by this perspective, we propose CoRePrune, a training-free two-stage framework. Progressive Perturbation-Aware Visual
Pruning refreshes deletion effects as visual representations evolve, while Set-Conditioned Refinement reevaluates candidate rescue benefits under the current deletion set after visual–text interaction. Across five
multimodal large
language model backbones covering standard images, high-resolution inputs, and
video, CoRePrune preserves performance under aggressive
token budgets. On Qwen3.5, with a final budget of 128 visual
tokens, it retains 90.3% of dense-model performance while reducing aggregate prefill time by 51.0%.
Read the full abstract ▾
All authors Shengli He, Yongchao Liang, Roumeng He, Junjie Zeng, Jiyuan He, Xin Fang, Can Wu, Li Zheng
Astrophysics of Galaxies
Exploring the influence of reionisation-era galaxies on their local intergalactic medium using the Sherwood-Relics simulations
Hamish Anderson, Laura C. Keating, Luke Conaboy et al. (+3 more)
Recent surveys have detected excess flux in the Lyman-$\alpha$ forest of
quasar spectra several Mpc from high-
redshift line-emitting
galaxies, indicating a correlation between intergalactic medium (IGM) transmission and reionisation-era
galaxies. We examine the
galaxy-IGM connection using the Sherwood-Relics simulation suite, a set of hybrid radiation-hydrodynamic simulations that model the IGM at high
redshift. We simulate the Lyman-$\alpha$ forest and a population of mock [OIII] emitters using the survey specifications of JWST's ASPIRE programme. We measure the
evolution of the effective optical depth in regions close to
galaxies, and contrast this with a control sample probing the random IGM. We characterise the changing physical properties of the IGM to explain the observed sightline-to-sightline scatter in opacity between 5.4 < $z$ < 6.2. We find that the IGM is more opaque than average at small radii ($r$ $\leq$ 5 cMpc/$h$) close to
galaxies, as these regions are overdense. At larger radii (15 < $r$ $\leq$ 50 cMpc/$h$), the IGM is more ionised than average, due to the higher photoionisation rate near
galaxies. We find qualitative agreement with the observational results from ASPIRE, though we see large realisation-to-realisation scatter driven by
cosmic variance, as also seen in theoretical explorations of the
galaxy-Ly$\alpha$ transmission cross-correlation function. Our results generally agree with observations claiming that there is a large-scale ionisation bias around [OIII] emitters, supporting the idea that their role is significant in shaping the properties of the IGM during reionisation.
Read the full abstract ▾
All authors Hamish Anderson, Laura C. Keating, Luke Conaboy, Xiangyu Jin, James S. Bolton, Ewald Puchwein
Geophysics
Shallow-to-deep velocity model building via diffusion models-Part II: Realistic scenarios
Shijun Cheng, Randy Harsuko and Tariq Alkhalifah
Full-waveform inversion (FWI) requires accurate initial velocity models to avoid cycle-skipping, but constructing such models remains challenging in practice. Building on the depth-progressive
diffusion framework introduced in Part~I, which relied on idealized reflectivity constraints, this work adapts the methodology to realistic exploration scenarios. We replace perfect structural information with migration-derived attributes extracted from seismic images, and introduce smooth background velocity models from
tomography as additional conditioning inputs. The framework jointly leverages background/migration velocity, migrated structural information, and sparse well measurements to synthesize high-resolution velocity models through depth-progressive generation. Validation on synthetic examples demonstrates superior accuracy compared to conventional interpolation and alternative deep learning methods, with generated models successfully initializing FWI and mitigating cycle-skipping even in complex geological structures. Field data confirms practical applicability: despite training on synthetic data, the method generalizes effectively to field conditions, producing velocity models with synthetic data response that nearly match observed seismic data. As a result, this framework establishes a practical pathway to deploy generative
diffusion models for velocity model building under realistic constraints.
Read the full abstract ▾
Multiagent Systems
Behavior is Not Enough: A Mechanism-Based Evaluation of Social Norm Emergence in LLM Societies
Rasika Muralidharan, Haewoon Kwak and Jisun An
Social norms cannot be identified from behavior alone: the same cooperative
equilibrium may reflect shared expectations, strategic incentives, or simple imitation. Yet in multi-
agent large
language model systems, prior work largely treats behavioral convergence as evidence of norm emergence. In this work, we introduce an evaluation framework that measures
agents' reported empirical and normative expectations in addition to behavioral convergence. Through controlled ablations, we test the effect of expectation elicitation and isolate two collective mechanisms central to theories of norm formation—social learning through interaction and social selection through network-based group formation. We further test the stability of these resulting dynamics under
adversarial disruption across four LLM families. We find that eliciting expectations increases cooperative contributions, while social learning stabilizes behavior, and social selection reliably identifies cooperators but provides limited behavioral reinforcement. Following disruption, normative expectations and behavioral coordination recover differently. Together, these results show that similar cooperative outcomes can arise from different underlying social processes. By making expectations observable, our framework allows us to attribute each mechanism's contribution separately, offering designers of multi-
agent systems a principled basis for selecting the social processes that sustain cooperation.
Read the full abstract ▾
Software Engineering
FeatLens: Feature-Guided Dependency Retrieval for Repository-Level Code Generation
Xutian Li, Bo Xiong, Yifeng Zhu et al. (+4 more)
Retrieving task-relevant code dependencies is critical for repository-level code generation. Existing approaches obtain dependency context through similarity-based
retrieval, repository graphs, or LLM-driven exploration. A central challenge is to identify the dispersed dependencies needed to implement a functional requirement, rather than merely retrieve broadly related code. This paper presents FeatLens, a feature-guided dependency
retrieval approach for repository-level code generation. FeatLens first builds an offline index linking natural-
language feature descriptions to function-level code entities. Given a task requirement and target function location, it selects relevant features and constructs dependency seeds. It then expands the seed graph and applies semantic and structural heuristics to select dependencies for inclusion in the generation context. Experiments on DevEval and EvoCodeBench show that FeatLens achieves the highest dependency recall among the evaluated baselines at each fixed
retrieval budget, reaching DR@15 values of 0.527 and 0.529, respectively. On DevEval, the retrieved context supports the highest non-oracle Pass@1 with both DeepSeek-V3.2 and GPT-5-mini. These results are obtained without LLM calls in the online dependency
retrieval stage.
Read the full abstract ▾
All authors Xutian Li, Bo Xiong, Yifeng Zhu, Runbang Yan, Yanzhen Zou, Lu Zhang, Bing Xie
Dynamical Systems
Non-determinism in group actions and topological minimal self-joinings
Nicolás Bitar, Sebastián Donoso and Samuel Petite
We study the topological and geometrical aspects of non-determinism in group actions, generalizing the notion of non-expansive direction for abelian actions. Under suitable conditions on the acting group, we establish restrictions on the collection of non-deterministic horoballs (or half-spaces) for any action. We illustrate our results in the case of free and nilpotent groups, expanding upon known cases for abelian groups. We then apply these findings to standing problems concerning minimal self-joinings in topological dynamics, which are of independent interest.
Read the full abstract ▾
Numerical Analysis
Linear Robin-type Domain Decomposition scheme for Mixed Formulation of Richards' Equation
Åsmund v.B. Synnevåg, Wietse M. Boon, Florin A. Radu et al. (+1 more)
In this work, we present a linear Robin-type domain decomposition scheme for solving the mixed formulation of Richards' equation (mLRDD-scheme), governing flow in variably saturated porous media. Assuming a highly heterogeneous porous medium consisting of multiple blocks/layers of distinct materials, we apply non-overlapping domain decomposition to isolate individual materials, yielding near-homogeneous conditions within each subdomain. Richards' equation is discretized in time by backward Euler and linearized using the L-scheme. A Robin-type interface condition, consistent with the mixed formulation, is then derived to couple the subdomains. For spatial discretization, mixed finite elements are employed, ensuring local mass conservation and providing the flux variables consistent with the Robin condition. The convergence of the proposed scheme is rigorously proved. Heterogeneous, multidomain numerical examples in 2D/3D are presented to demonstrate the efficiency and robustness of the scheme, along with numerical validation of the theoretical convergence results.
Read the full abstract ▾
All authors Åsmund v.B. Synnevåg, Wietse M. Boon, Florin A. Radu, Sarah E. Gasda
Number Theory
A divisibility of automorphic periods for the real quadratic base change of $\mathrm{GL}_3$
Tristan Ricoul
We prove a $p$-adic divisibility between the automorphic periods of a cuspidal automorphic representation $\pi$ of $\mathrm{GL}_3(\mathbb{Q})$ and the periods of its Arthur-Clozel base change to a real quadratic field $E$. As a corollary, we establish a divisibility predicted by the Bloch-Kato
conjecture for the adjoint motive of $\pi$ twisted by an even quadratic character. This generalizes earlier works of Tilouine-Urban and Hida in the case of classical modular forms. The divisibility we prove involves a new kind of automorphic periods for conjugate self-dual cuspidal automorphic representations of $\mathrm{GL}_3(E)$, defined within the middle degree of the cuspidal
cohomology instead of the top or bottom degrees. Moreover, we prove an à la Hida adjoint $L$-value formula for $\mathrm{GL}_3(E)$ that relates these newly defined middle-degree periods to the usual top and bottom automorphic periods.
Read the full abstract ▾
Distributed Computing
Don't let your Memory defy you: Fragmentation-Aware Serverless Allocation with Elastic Memory Locality
Achilleas Tzenetopoulos, Dimosthenis Masouros, Sotirios Xydis et al. (+2 more)
Serverless platforms commonly rely on bundled,
memory-centric configurations, where CPU capacity follows the specified
memory size. Resource decoupling reduces this waste, but can create external fragmentation by producing diverse CPU-
memory shapes that leave residual capacity stranded across nodes.
Memory disaggregation and tiering can turn these resource holes into usable capacity by enabling elastic
memory locality, where instances use different mixes of local and remote
memory. In this paper, we present Memoryless, a fragmentation-aware resource manager that exposes
memory locality as a serverless control-plane primitive. Memoryless jointly selects and places function variants that differ in their compute allocation and
memory-locality ratio. This allows the scheduler to match instance shapes to fragmented node capacity while preserving SLOs. We implement Memoryless on top of Knative and evaluate it with trace-driven workloads. Compared to state-of-the-art serverless placement frameworks, Memoryless reduces node usage by up to 40% in steady state and 46% at peak, while keeping SLO violations within 3%. It also reduces CPU stranding by 44%, showing that elastic
memory locality can convert fragmented node capacity into usable resources.
Read the full abstract ▾
All authors Achilleas Tzenetopoulos, Dimosthenis Masouros, Sotirios Xydis, Francky Catthoor, Dimitrios Soudris
General Relativity and Quantum Cosmology
Stochastic motion of massless particles near event horizons in Schwarzschild and Reissner-Nordström spacetimes
Yinhan Huang, Lezhen Miao and Xiao Zhang
In classical general relativity, it is well known that radially moving massless particles in the exterior region take infinite coordinate time to reach the event horizon in black hole spacetimes. In this short paper, we use two
stochastic versions of ingoing radial null geodesic equation to study the
stochastic radial motion of massless particles in Schwarzschild and Reissner-Nordström spacetimes. We show that the probability that a radially moving massless particle enters the black hole in finite coordinate time remains zero. This indicates the
stochastic perturbations of
quantum effects preserve certain properties of classical black holes.
Read the full abstract ▾