Computer Vision
AGILE-GS: Anchor-Guided Fast Next-Best-View Selection for Active 3D Gaussian Splatting
Amirhossein Mollaei Khass and Nader Motee
Radiance fields need hundreds of views, and their placement matters as much as their number. Next-best-view (NBV) selection for 3D Gaussian Splatting (3DGS) usually scores every candidate in the pool and keeps one. Searching for information and choosing a camera, however, are separable problems. We present AGILE-GS, an anchor-guided NBV method that separates the two. A virtual anchor pose is optimized on SE(3) by Riemannian gradient ascent on expected information gain. It need not be reachable or in the pool; it marks where the model is most uncertain. Candidates are scored against the anchor's viewing geometry, and a greedy ridge-leverage step distills the pool into a small, non-redundant shortlist without rendering any candidate. The shortlist can be used in two ways. AGILE-GS takes the first view on it as the next view, so no Fisher information is computed for any candidate. AGILE-GS+ computes the Fisher information gain of each shortlisted view and picks the best, so the expensive evaluation runs on a handful of views rather than the whole pool. On standard
benchmarks and in closed-loop
embodied acquisition, both match or exceed existing baselines while cutting selection latency by one to two orders of magnitude.
Read the full abstract ▾
Robotics
FailPatch: Failure Residual Patching for Vision-Language-Action Models
Peng Yu, Jiacheng Wang, Ziheng Zhang et al. (+6 more)
Vision-
Language-Action (VLA) policies are typically adapted using successful demonstrations, which provide direct action supervision but rarely cover failure-prone states. Deployment failures expose these states, yet lack the corrective actions needed for conventional supervised learning. We propose FailPatch, a failure-driven residual patching framework that decouples action supervision from execution-reliability supervision. Successful demonstrations ground how the policy should act, while deployment trajectories indicate when its behavior becomes unreliable. We further observe that action hidden representations exhibit clear linear separability between reliable and failure-associated states while directly conditioning action generation. Building on these insights, FailPatch introduces a Null-gated Residual Expert Bank into the action hidden space of a frozen VLA policy. A unified Preserve–Redirect–Trust objective retains the original policy in reliable states, selects residual experts in failure-associated states and redirects representations from failure regions toward success-associated regions under bounded intervention. With only 0.52% trainable parameters, FailPatch improves success rates by 11.0 percentage points on four long-horizon RoboTwin tasks under clean evaluation, 9.5 percentage points under clean-to-random generalization, and 16.7 percentage points over the baseline across three real-world tasks. Project and code: https://github.com/yupeng-2003/FailPatch.
Read the full abstract ▾
All authors Peng Yu, Jiacheng Wang, Ziheng Zhang, Xuchong Zhang, Baoting Li, Zhuoyuan Yu, Yuxiang Chen, Tiancai Wang, Hongbin Sun
Machine Learning
GradLev: Token-Parallel Test-Time Training Via Costate Prediction
Bo Liu and Qiang Liu
Test-time training (TTT) allows a model to improve its predictions at inference time by updating weights after every observed
token. However, sequential gra- dient writes make parallel training difficult. We observe that, given layer inputs and activation gradients (costates), online gradient descent admits exact parallel scans for both forward evaluation and reverse backpropagation. GradLev lever- ages this duality: a causal auxiliary network predicts costates across all
tokens in parallel; associative scans compute the adapted weights and forward activations and propagate gradients backward; and the resulting gradient targets supervise the predictor via a consistency loss. Exact consistency guarantees exact recovery of the sequential online learner. At deployment, the auxiliary predictor is discarded, and the model updates natively via
token-by-
token forward and backward passes.
Read the full abstract ▾
Systems and Control
A Weak Notion of Symmetry for Control Systems
Jake Welde, Riley Link and Pieter van Goor
Symmetry (or invariance) is a powerful structural property that enables efficient, effective solutions for estimation and control. However, the constraints imposed on a system's dynamics by classical invariance make
symmetry a very rigid property, which may be broken by external forces or confined to only a portion of the overall system. Seeking greater flexibility, this work introduces a novel relaxed notion of
symmetry, termed “weak invariance”, in which the non-symmetric part of the dynamics (the “residual”) can be captured entirely by another control system evolving on the
symmetry group. Weakly invariant systems are strictly more general than classical invariant systems, but they nonetheless enjoy many similar favorable properties. In particular, we prove that any weakly invariant system admits a cascade decomposition in which the driven subsystem is group affine, showing that weak
symmetry generalizes not only classical
symmetry, but also the (thus far distinct) class of group affine systems. We also show that a weak
symmetry with autonomous residual can be factored out of the system's error dynamics, enabling yet a greater reduction of dimensionality as compared to classical
symmetries. Finally, we study the example of an aerial vehicle under the influence of
gravity, for which we propose a nine-dimensional weak
symmetry (strictly containing the system's familiar four-dimensional classical
symmetry). Weak invariance thus generalizes classical
symmetry while also preserving key structural properties, thereby laying a foundation for more flexible methods of
symmetry-informed control.
Read the full abstract ▾
Machine Learning
GPARA: Graph-Posterior-Aligned Refinement and Active Acquisition for Grounding Diffusion Priors
Wangqian Chen, Hao Wang, Yumeng Zhang et al. (+3 more)
Active grounding of a frozen
diffusion prior requires jointly determining where new measurements should be taken and how they should be used to refine the current reconstruction. Posterior-ensemble-based methods can estimate acquisition utility from generated samples, but require repeated ensemble generation as observations accumulate and capture posterior geometry only through empirical statistics. This paper proposes GPARA, which learns a context-dependent graph surrogate over
diffusion prediction residuals, inducing an explicitly reusable posterior response operator that propagates measurement innovations to unobserved variables and evaluates candidate measurements through weighted posterior-risk reduction. Under the matched surrogate, we show that the same response operator also determines expected one-step acquisition benefit and yields an analytic ranking consistent with expected reconstruction improvement. A bounded learned residual calibrates the analytic utility to account for surrogate mismatch, while a small prior ensemble is generated once and reconditioned to update risk weights without repeated
diffusion posterior sampling during acquisition. Experiments on two reconstruction tasks spanning physical field and computer
vision show consistent improvements in refinement and active acquisition over the evaluated baselines. Ablations further support the complementary roles of step-wise graph refinement, adaptive risk weighting, and analytically anchored calibration.
Read the full abstract ▾
All authors Wangqian Chen, Hao Wang, Yumeng Zhang, Jiajia Guo, Junting Chen, Jun Zhang
Optics
Monolithically integrated photonic neural network at driven-dissipative criticality
Yuming Zhang, Ruoran Wang, Jingcheng Li et al. (+4 more)
Artificial intelligence increasingly demands computing architectures capable of adaptive representation, temporal information processing, and robust operation under uncertainty. However, conventional photonic neural architectures typically rely on predefined computational operations and separate functional modules, limiting the ability of physical systems to exploit their intrinsic dynamics for intelligence. Here, we demonstrate a monolithically integrated active photonic neural network (APNN) that harnesses driven-dissipative criticality as a physical computational resource. By balancing optical excitation and dissipation, the system operates near a critical regime where weak input-dependent variations are amplified into separable representations while stable attractor dynamics are preserved. The APNN integrates an optoelectronic nonlinear array and a reconfigurable symmetric optical coupling
matrix on a silicon photonic chip, enabling recurrent physical computation within a compact closed loop. We demonstrate nonlinear classification with 95.2% accuracy using only 100 training samples and 44 trainable parameters. And it remains robust at an ultralow signal-to-noise ratio of -30 dB. Beyond static recognition, the same driven-dissipative dynamics enable long-range temporal feature extraction from electrocardiogram signals and associative recovery of corrupted patterns through attractor-based relaxation. These results establish driven-dissipative criticality as a unified mechanism for representation enhancement, temporal information retention, and error-resilient computation, providing a pathway toward scalable photonic systems in which intelligence emerges from intrinsic physical dynamics.
Read the full abstract ▾
All authors Yuming Zhang, Ruoran Wang, Jingcheng Li, Hailong Zhou, Jianji Dong, Zhipei Sun, Xinliang Zhang
Robotics
RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Models
Yuhan Chen, Ke Yu, Pengfei Liu et al. (+3 more)
Flow-based
vision-
language-action (VLA) models are highly effective for generalist
robot manipulation, yet their reliance on computationally expensive VLM encoding and multi-step iterative action generation imposes a significant latency bottleneck. The resulting inference latency makes it difficult for
robots to respond quickly, especially in dynamic environments. We address this limitation with RAVEL (Rolling Asynchronous VLA Enabling Low-Latency Control), an asynchronous inference framework that addresses the computational bottlenecks of both the VLM backbone and the action expert. To reduce the delay from multi-step action denoising, RAVEL allows near-term actions to be executed after a single denoising step by carrying partially denoised future actions forward in a rolling buffer. To avoid blocking on slow VLM encoding, RAVEL decouples VLM encoding from rolling action generation, allowing the action expert to operate continuously using the latest available VLM context, while a lightweight Fast Observation Pathway (FOP) directly conditions the action expert on current observations. Across simulated and real-world manipulation tasks, RAVEL consistently achieves substantially lower response latency while maintaining the task capability of the underlying VLA, enabling high-frequency and responsive closed-loop control.
Read the full abstract ▾
All authors Yuhan Chen, Ke Yu, Pengfei Liu, Shuxun Wang, Yi Yang, Linchao Zhu
Computational Engineering
From Word Counts to Context: Topic Models for Asset Pricing
Kevin Foley, Jonathan Hartadi, Shivesh Prakash et al. (+1 more)
News may reveal systematic risk, but whether its context enhances the construction of systematic risk factors is still unclear. We seek to test whether utilizing a sentence
transformer represents an improvement over techniques such as Latent Dirichlet Allocation (LDA) in the coherence of topic term lists generated from unstructured text data. To test this, the same collection of unstructured text data comprising of 394,661 articles and the same downstream financial portfolio construction pipeline were applied with the text layer differing, including the length of article text each model used and how topic terms were ranked: we
benchmark LDA against a frozen sentence
transformer with k-means clustering. We find that the sentence
transformer branch had higher observed scores both in terms of coherence (measured by NPMI) as well as financial performance (measured by Sharpe), although the available tests do not establish outperformance. Further exploratory specifications such as utilizing spherical clustering and multi-horizon exposures had an observed excess-return Sharpe of 1.03 for the combined model. We believe that there is some promise in applying context-aware techniques on unstructured news text, but stricter tests using only information available at each date and broader
datasets may be required to enhance the confidence in the observed performance.
Read the full abstract ▾
All authors Kevin Foley, Jonathan Hartadi, Shivesh Prakash, Swapnil Vatsal
Dynamical Systems
Geometric and functional mixing by 2D stationary incompressible flows
Weiwei Hu, Ziqian Li and Yubiao Zhang
We study quantitative mixing and deformation of sets and curves for a class of two-dimensional autonomous
Hamiltonian flows with finitely many critical points satisfying local conditions that allow finite-order degeneracy. Variation of the period across neighboring trajectories generates transverse shear, providing a common mechanism for scalar mixing, set deformation, and curve stretching. First, for $H^1$ initial data supported away from
equilibria and infinite-period trajectories, in regions where the period gradient is bounded away from zero, we establish sharp $(1+t)^{-1}$ decay in $H^{-1}$ towards the time average of the initial data along each periodic trajectory. Second, under the same
geometric conditions, we prove matching upper and lower bounds of order $(1+t)^{-1}$ for an orbit-relative
geometric mixing scale of transported Lipschitz subdomains whose closures are not invariant under the flow. This scale measures how closely the transported subdomain covers the union of trajectories meeting its initial position. Third, for Lipschitz curves separated from infinite-period trajectories, we derive an explicit first-order large-time expansion of their length with a remainder bounded uniformly in time. In particular, their length grows at most linearly. Counterexamples illustrate how the stated conclusions can fail when selected nondegeneracy or separation assumptions are removed. The analysis combines coordinates adapted to the periodic trajectories with quantitative estimates and asymptotic expansions for the flow Jacobian. Numerical simulations for cellular and radial flows illustrate the functional and
geometric mixing rates and the
evolution of curve length.
Read the full abstract ▾
Computer Vision
Natural Image Autoencoder-Based fMRI Representations for Trait and State Prediction
Juhyeon Park, Yeonwoo Kim, Peter Yongho Kim et al. (+5 more)
Foundation models pre-trained on large-scale fMRI
datasets have shown strong downstream performance, but at substantial data and computation cost. To investigate how much fMRI-specific pre-training is actually needed for such performance, we introduce FReD, which derives fMRI representations from a frozen Deep Compression AutoEncoder (DCAE) pre-trained exclusively on natural images and pairs them with a task specific readout. For trait prediction, FReD summarizes frame-wise representations by their temporal mean and log-standard deviation and applies linear probing, with late fusion across two normalization schemes. For state prediction, it represents each frame as a single
token and models temporal dependencies with a shallow
Transformer. Across four resting-state
datasets spanning six trait-prediction targets, linear probes on frozen DCAE features generally outperform those on fMRI foundation model representations and remain competitive with fully fine-tuned fMRI foundation models. On three task-fMRI state-prediction tasks, a temporal readout on DCAE features performs comparably to the strongest foundation models evaluated. A Gaussian injection analysis further shows that localized signal changes are recovered more accurately from the frozen DCAE features than from the evaluated foundation-model representations. Together, these results show that strong performance on current fMRI
benchmarks is possible without fMRI-specific representation pre-training, making frozen natural-image features as a useful baseline for assessing its added value.
Read the full abstract ▾
All authors Juhyeon Park, Yeonwoo Kim, Peter Yongho Kim, Yansen Wang, Mingqing Xiao, Dongqi Han, Dongsheng Li, Taesup Moon
Machine Learning
Hidden Activations are not Enough I: Knowledge Matrices as Higher Representations
Marco Armenta
We study the knowledge
matrix of a trained feedforward network as a higher representation of its inputs. A network is a pair $(W,f)$, a thin representation $W$ of its quiver and an activation $f$; its function factorizes through the space of quiver representations, each input $x$ inducing a representation, and the knowledge
matrix $M(x)\in\mathbb{R}^{C\times(d+1)}$ is the contraction of that representation to one
matrix whose rows sum exactly to the logits. At one trained network we ask what determines it, what it is invariant to, what it determines, and what its geometry measures. Under (LCS), a locally constant slope diagonal, as for ReLU, the
matrix at a regular input is a function of the realized germ; its stabilizer among encodings regular there is exactly the germ stabilizer at inputs with no vanishing coordinate,
neuron permutation a special case; and it recovers the germ, whereas hidden activations, gauge-covariant and germ-incomplete, are not enough. Under (LCS) it equals per-class gradient$\times$input plus an exact aggregate bias attribution, grounding it in attribution theory and computing it by $C$ vector-Jacobian products instead of probing. The fixed shape gives an
alignment-free per-sample distance between ResNet-152, DenseNet-121 and GoogLeNet; the row-sum identity gives an exact visible/invisible displacement decomposition whose unit-free coherence $A=(d_\Psi/d_M)^2$ puts
adversarial germ motion at median $A\le 0.23$, with an attack-family ordering concordant across six architectures (Kendall $W=0.921$; $0.97$ on the three networks at full scale). Two honest negatives: on AlexNet/CIFAR-10 penultimate features win 5 of 6 detectors and all 16 attacks, and a
matrix-direction counterfactual fails 0/54.
Read the full abstract ▾
Number Theory
Exceptional poles of archimedean Asai $L$-functions
Akash Yadav
Let $\pi$ be an irreducible generic representation of $\operatorname{GL}_n(\mathbb{C})$. We classify the exceptional poles of the associated Asai $L$-function $L(s,\pi,\mathrm{As})$ and give a representation-theoretic characterization of them. When $\pi$ is in general position, we further express $L(s,\pi,\mathrm{As})$ in terms of the exceptional $L$-factors attached to the irreducible constituents occurring in the derivatives of $\pi$.
Read the full abstract ▾
Chemical Physics
Basis Functions for Time-Dependent Kohn-Sham Inversion
Paul M. Zimmerman, Joshua Kammeraad, Duy-Khoi Dang et al. (+1 more)
Floquet theory provides insight into the inversion of time-dependent Kohn-Sham density functional theory. Specifically, mathematical derivations show that the fundamental frequencies of a time-dependent wavefunction solution are the leading-order harmonics for the TD-KS state. Numerical tests of the resulting ansatz in 1D and 3D for atomic and molecular cases demonstrate its utility. In particular, low $L_{2}$ errors in the time-dependent density and longitudinal current were found, even though currents were not an explicit optimization objective. In all, the proposed inversion ansatz provides exchange-correlation potentials from time-dependent wavefunctions, is highly interpretable, and may significantly help in the development of nonadiabatic density functionals.
Read the full abstract ▾
All authors Paul M. Zimmerman, Joshua Kammeraad, Duy-Khoi Dang, Harish S. Bhat
Robotics
Bojun Long, Lingfan Bao, Tianhu Peng et al. (+2 more)
Vision-
language navigation (VLN) is typically evaluated as a one-way task, although deployed
robots may need to return after reaching a goal. We study continuous round-trip VLN and diagnose failures in directional observability, deviation recovery, and termination stability. We propose a reliability-aware sparse route
memory that records the executed Outbound trajectory as ordered
geometric anchors and queries them in reverse through a structured hint, action-level arbitration, and terminal verification. On 50 reverse-paired episodes using NaVILA and a simulated Unitree Go2,
language-only Return succeeds in 22.0% of episodes, while our online system reaches 55.1%. With exact route information, the same interfaces achieve 86.0%, showing that effective Return requires both accurate information and consistent action on that information. The remaining online gap arises mainly from
geometric evidence that is too unreliable to authorise intervention. These results distinguish information quality, behavioural consistency, and online reliability as separate limits in long-horizon
navigation.
Read the full abstract ▾
All authors Bojun Long, Lingfan Bao, Tianhu Peng, Jingcheng Sun, Chengxu Zhou
Distributed Computing
SlideDP: Scaling Host-Resident LLM Fine-Tuning Across Multiple GPUs
Ruijia Yang, Shiyuan Lin, Yulong Ao et al. (+5 more)
Host-resident layer streaming enables full-parameter LLM fine-tuning beyond GPU
memory, but data-parallel ranks compete for shared host resources. Replicated transfers amplify traffic, while strong scaling can expose host work as computation windows shrink. We present SlideDP, a synchronous data-parallel runtime for shared-host multi-GPU systems. It maintains one authoritative host state, decouples communication routes from state layout, and pipelines parameter delivery, gradient aggregation, and CPU updates across ranks and chunks. An analytical step-time model characterizes resource bottlenecks and pipeline exposure; runtime measurements guide communication, chunking, and activation policies under a GPU
memory budget. In matched-batch sweeps, SlideDP achieves
geometric-mean throughput ratios of 1.46-2.64$\times$ over SlideFormer, MegaTrain, and ZeRO-Offload. On four H100s, SlideDP approaches GPU-resident FSDP2 throughput for Qwen3-14B at a smaller batch size. With a larger batch, it processes over 1M
tokens per step and exceeds FSDP2's measured peak throughput by 11.2%. Separately, it supports 256K-
token sequences for the same model and fine-tunes Qwen2.5-72B on four RTX 4090 GPUs. Project page: https://github.com/RegiaYoung/SlideDP.
Read the full abstract ▾
All authors Ruijia Yang, Shiyuan Lin, Yulong Ao, Zhiyu Li, Yingli Zhao, Xianduo Li, Yonghua Lin, Zeyi Wen
Machine Learning (Statistics)
GT-PSSM: Unified Probabilistic Framework for Stochastic Dynamics Modeling and Dependency Learning in Multivariate Time Series Anomaly Detection
Wonmo Koo, Jaeyeong Lee, Taeseong Yoon et al. (+1 more)
Multivariate time series anomaly detection (MTAD) is crucial for ensuring the safe and reliable operation of complex systems. Many existing methods learn normal patterns by training reconstruction or
forecasting models on predominantly normal data. However, a large portion of these approaches rely on deterministic models and their associated point-wise output errors for anomaly scoring. Since real-world multivariate time series are inherently
stochastic due to measurement noise and intrinsic system randomness, purely error-based scores can be unreliable, as large errors may arise from benign fluctuations rather than true anomalies. Probabilistic approaches address this limitation by quantifying uncertainty in model outputs. In particular, probabilistic state-space models (PSSMs) provide a principled framework by modeling
stochastic system dynamics through latent state transitions and measurement noise via emission models. Despite this advantage, existing PSSM-based MTAD methods often struggle to capture long-range temporal dependencies and inter-variable dependencies, as they typically rely on noise-sensitive recurrent architectures and lack explicit cross-variable structure modeling. To address these limitations, we propose Graph-
Transformer-Enhanced Probabilistic State-Space Model (GT-PSSM), a novel PSSM-based MTAD method that tightly integrates PSSM-based probabilistic modeling of
stochastic dynamics with Graph
Transformer-based learning of temporal and inter-variable dependencies. By jointly modeling stochasticity, long-range temporal dependence, and variable interactions within a unified probabilistic framework, GT-PSSM enables more robust anomaly detection.
Read the full abstract ▾
All authors Wonmo Koo, Jaeyeong Lee, Taeseong Yoon, Heeyoung Kim
Artificial Intelligence
RoutePrism: Tracing Construction Order Effects in Agent Memory
Dong Xu, Zhangfan Yang, Jiantao Wu et al. (+5 more)
Processing the same records in a different order can discard different evidence, yet endpoint accuracy alone cannot reveal what changed or whether it mattered. We introduce RoutePrism, a diagnostic protocol that builds
memory twice from the same source pool in two processing orders, then traces which sources, compiled contexts, and answers differ. Because record content, timestamps, policy, and the answer model all stay fixed, any observed difference is localized to the
memory construction step. A matched four-condition intervention tests whether a record displaced by reordering actually carried task-relevant evidence: restoring that single record recovers over 60 percentage points of lost accuracy, while substituting a non-supporting record of equal length does not. We evaluate the protocol on PersonaMem-32K (63 primary queries, 29 users) and 470 LongMemEval-S questions with histories spanning 38 to 62 sessions, replicating the core intervention across five answer models. Survivor selection, defined as the choice of which record a cluster retains, drives most source-level changes, while different
memory policies (compaction, bounded recency, MemoChat-style summarization, A-MEM) produce distinct failure signatures at the source, context, and metadata layers.
Read the full abstract ▾
All authors Dong Xu, Zhangfan Yang, Jiantao Wu, Shipeng Zhang, Zexuan Zhu, Jiangqiang Li, Jun Zhang, Junkai Ji
Machine Learning
WorldGraph: Graph-Native World Modeling
Zezhong Ding, Yipeng Li and Xike Xie
World models infer latent states of an environment to capture its underlying dynamics and predict future
evolution. Many real-world environments, however, are inherently relational and observed as evolving graphs, where entities, relations, and their properties change over time. Prior graph-related world models use graph structures to organize internal states or support task-specific
reasoning, rather than treating an evolving graph itself as the modeled world. We instead study graph world modeling (GWM), where graph
evolution itself constitutes the world dynamics. We formulate graph world modeling over observed graph
evolution, latent graph states, and heterogeneous graph-transition predictions. Based on this formulation, we construct GWM-Zero, a
benchmark covering node-, edge-, and graph-level transitions over eight temporal graph
datasets. We propose WorldGraph, which combines a state-aware graph
transformer for multi-granularity structural and transition-conditioned
evolution modeling with transition-aware GRPO using dynamic grouping and structure-aware verifiable rewards. Extensive experiments on GWM-Zero show that WorldGraph consistently outperforms representative graph representation, temporal graph learning, graph pretraining, and graph world-model baselines across all three transition granularities.
Read the full abstract ▾
Computation and Language
Toward a Graded Measure of Belief Stability in Large Language Models
Samantha Dies, Branden Fitelson and Tina Eliassi-Rad
Large
language models (LLMs) increasingly mediate how people access and reason with information, yet factual reliability is usually evaluated one judgment at a time. We introduce graded belief stability, a relational measure of how well a belief persists within an LLM's broader belief system. Unlike individual belief probability, it asks whether support for a claim persists when that claim is considered alongside the model's other epistemic commitments. We operationalize this idea with a Direct Conditional
estimator that uses internal model representations to estimate conditional belief probabilities. Across 12 LLMs and three domains, lower-stability beliefs exhibit greater mean behavioral movement under conversational challenge in 83.3% of model-domain settings after matching on individual belief probability. Graded belief stability therefore extends reliability assessment beyond how strongly an LLM supports a claim to how robustly that belief is supported within its broader system of beliefs.
Read the full abstract ▾
Artificial Intelligence
TableSeek: Structure-Preserving Agentic Evidence Seeking over Heterogeneous Table Corpora
Jiaming Tian, Liyao Li, Wentao Ye et al. (+5 more)
Open-domain table
retrieval seeks tables that contain sufficient evidence for answering a question or verifying a claim. Yet semantic relevance is often misleading: topically similar tables may lack the required facts, while answer-bearing evidence is often confined to a few
cells whose meaning depends on surrounding schema and table context. Heterogeneous schemas, value formats, and serializations further weaken one-shot matching. We present TableSeek, a structure-preserving agentic search framework for heterogeneous table corpora. Instead of ranking tables once, an LLM
agent iteratively follows sparse clues, inspects schema-preserving previews, identifies schema- and value-level mismatches, and refines its investigation. TableSeek uses
cells and schemas as evidence anchors while retaining complete tables as evidence units, enabling fine-grained localization without losing the context required for interpretation and answerability checking. Without relying on retriever training or a precomputed semantic index, TableSeek produces transparent evidence-seeking trajectories and achieves competitive end-to-end performance against strong
retrieval-and-reranking pipelines on heterogeneous table
benchmarks. These results suggest that active, structure-preserving evidence seeking is a promising paradigm for open-domain table
retrieval.
Read the full abstract ▾
All authors Jiaming Tian, Liyao Li, Wentao Ye, Haobo Wang, Lihua Yu, Zujie Ren, Gang Chen, Junbo Zhao
Machine Learning
Transfer Calibrated Prediction Powered Inference
Aditya T. Vadlamani, Jae Ho Chang, Srinivasan Parthasarathy et al. (+1 more)
Prediction-powered inference (PPI) and its power-tuned extension (PPI++) improve confidence intervals by combining a small gold-standard labeled sample with a large AI model's predictions. Its efficiency gain relies on low residual variance, which may not hold if the predictor is pre-trained on a different source domain. We propose Transfer Calibrated Prediction-Powered Inference (TC-PPI), adapting the source-domain predictor to the target domain using gold-standard samples through cross-fitting. This approach supports various adaptation methods, such as sparse linear calibration, LoRA, and fine-tuning. Our jointly tuned cross-fit
estimator, Joint-TC-Cross-PPI++, maintains unbiasedness and is simultaneously at least as efficient as classical inference, PPI, and PPI++, thereby protecting against negative transfer. We provide high-dimensional MSE bounds for calibration and show empirical improvements over baseline methods across various real-world applications.
Read the full abstract ▾
All authors Aditya T. Vadlamani, Jae Ho Chang, Srinivasan Parthasarathy, Subhadeep Paul
High Energy Astrophysical Phenomena
Comment on "Fermi-LAT Galactic Center Excess Morphology of Dark Matter in Simulations of the Milky Way Galaxy"
Kevork N. Abazajian, Jason Kumar and Oscar Macias
Muru et al. [Phys. Rev. Lett. 135, 161005 (2025)] conclude from the HESTIA constrained simulations that the $\gamma$-ray morphologies of the
Galactic Center Excess predicted by dark matter annihilation and by an old
stellar population are essentially indistinguishable. Their "quadratic dark matter" maps, however, show the square of the mass projected along the line of sight, $\bigl[\int\rho\,s^{2}\,ds\bigr]^{2}$, rather than the annihilation integral $\int\rho^{2}\,ds$. Squaring a nonnegative map preserves its isophote shapes at corresponding intensity levels, as confirmed by the reported contour axis ratios. The corrected calculation on the same simulations does not support their conclusion.
Read the full abstract ▾
Differential Geometry
Properties and Perturbations of Extremally Ricci-Pinched $G_2$-Structures
Aaron Kennon
We study properties of Extremally
Ricci-Pinched (ERP) $G_2$-Structures on compact $7$-
manifolds. We provide characterizations of the ERP condition in terms of the traceless $\ast$-
Ricci tensor, the Hodge
Laplacian of the torsion, the
Ricci eigenvalues, the
Laplacian flow, and the $27$-dimensional Weyl
curvature component. We also prove that the
automorphism group of a compact ERP $G_2$-Structure is finite. We then identify perturbations of ERP $G_2$-Structures, both in general and for specific examples, that preserve the ERP condition.
Read the full abstract ▾
Machine Learning
SPINET: Sheaf Protein Inverse Folding Network
Jens Lundsgaard, Colin Mikulski, Zhixuan Yan et al. (+1 more)
Proteins change shape as they function, yet most inverse folding models predict amino acid sequences from a single, fixed backbone. A central challenge in
protein engineering is to design
proteins that undergo specific motions, which requires accounting for how their structures change over time. This motivates inverse
protein folding conditioned on
protein motion. We introduce SPINET, which predicts sequences from molecular dynamics trajectories. It uses cellular sheaves to represent residue interactions within each frame and recurrent units to integrate information across frames, then predicts all amino acids in a single pass. We evaluate SPINET on mdCATH and ATLAS, where it outperforms all evaluated static and ensemble baselines in sequence recovery. On mdCATH, it achieves 56.7% top-1 recovery, compared with 44.5% for the strongest static baseline and 40.7% for the strongest ensemble baseline. We also evaluate whether the predicted sequences are compatible with conformations sampled along the target trajectory. On mdCATH, they achieve a median TM-score of 0.760, and structural recovery favors target conformations over unrelated decoys for 99.5% of test domains.
Read the full abstract ▾
All authors Jens Lundsgaard, Colin Mikulski, Zhixuan Yan, Dhananjay Bhaskar
Computation and Language
Quantitative Measurement of Language Distance among Closely Related Indo-European Languages Using Pretrained Language Models: A Case Study on the North Germanic Branch
Yiping Bai
Among closely related North Germanic
languages, the quantification of
language distance has traditionally relied on qualitative methods, lacking a unified multi-dimensional computational framework.
Multilingual pretrained models based on the
Transformer architecture can map texts from different
languages into a shared vector space, enabling quantitative measurement of
language distance. This paper focuses on the three North Germanic
languages—Danish, Norwegian (Bokmål), and Swedish—and proposes a three-metric quantitative framework based on pretrained
language models: (1)~sentence-level semantic distance, computed as cosine similarity between LaBSE and mBERT encodings of parallel sentences; (2)~orthographic fragmentation rate, measuring subword tokenization efficiency when cross-applying monolingual BERT vocabularies to parallel texts; (3)~MLM predictability, comparing prediction confidence and
entropy in masked
language modeling using mBERT across
languages. Using 150 trilingual parallel sentence triplets from the Tatoeba corpus as controlled samples, we obtain consistent distance rankings on two independent models: LaBSE: da–no $0.012 < $ no–sv $0.016 < $ da–sv $0.020$; mBERT: da–no $0.016 < $ no–sv $0.045 \approx $ da–sv $0.046$. This ranking is consistent with the historical linguistic conclusion that “400 years of Danish rule over Norway (1380–1814) led to highly cognate written
languages.” The three metrics—semantic, orthographic, and predictability—converge on the same conclusion, providing a reproducible computational framework for the quantitative study of distance among closely related
languages, extensible in principle to more branches of the Indo-European
language family, pending validation on additional
language groups.
Read the full abstract ▾
Artificial Intelligence
Self-Evolving Agents via Likelihood-Guided Tool-Space Optimization
Xuanqi Zhang, Ruinan Jin, Running Yang et al. (+4 more)
Self-evolving
agents can continually improve their behavior, while tools define the executable action space through which they interact with the environment. However, exposing the full tool library to model introduces substantial irrelevant context and can impair tool-use decisions. We study tool-space self-
evolution, where each recurring task type maintains a persistent tool space which is constructed from accumulated output experience. We identify three limitations of existing methods: (1) output-unaware selection: they rely primarily on tool descriptions or model priors rather than observed tool outputs; (2) statelessness across request: they select tools independently for each request without consolidating prior output experience into persistent task-specific state; (3) inference cost: they repeatedly search, rank, or reason over candidate tools for subsequent requests of the same task. We address these limitations through output-aware tool scoring, persistent task-specific tool spaces, amortized tool selection, and reusable configurations across models. We introduce LOTS (Likelihood-Only Tool Scoring), which evolves an
agent's tool space from accumulated output experience while keeping model parameters fixed. After each request, LOTS holds the model's generated answer and estimates each tool's contribution by measuring how much the answer likelihood changes when its observed output is removed. These contributions are aggregated within each recurring task to rank tools and update its persistent space. Across three
benchmarks, LOTS improves task performance while substantially reducing tool context. More importantly, sequential experiments demonstrate that task-specific spaces persist and continue to improve over time, while cross-model experiments show that learned configurations transfer across different models.
Read the full abstract ▾
All authors Xuanqi Zhang, Ruinan Jin, Running Yang, Yuxuan Zhang, Minghui Chen, Wenlong Deng, Xiaoxiao Li
Astrophysics of Galaxies
PEGASUS (Pfs Emission-line GAlaxy SUrvey with Subaru): II. Metal Poor AGNs in High Equivalent-Width Narrow-Band Emitters
Hisakazu Uchiyama, Tohru Nagao, Mariko Kubo et al. (+19 more)
We present Subaru 'Ōnohi'ula
Prime Focus Spectrograph (PFS) optical–NIR spectroscopic follow-up of high-equivalent-width (EW) narrow-band (NB) emission-line
galaxies selected from Hyper Suprime-Cam Subaru Strategic Program, to assess whether high-EW sources preferentially host low-
metallicity AGNs. PFS spectroscopy yielded secure identifications for 583 targets, including 581 sources with reliable PFS
redshift solutions and two manually identified $z\sim4$ interlopers. We construct BPT diagnostics for the 70 objects with significant detections in all four BPT lines. Among them, $\sim34\%$ are located in the AGN region of the BPT diagram, suggesting that AGNs constitute a significant fraction of the sample. A non-negligible fraction does not populate the classical AGN branch; instead, it lies in the intermediate “valley” between the
star-forming and AGN sequences. Comparison to Cloudy models suggests that the valley population is consistent with sub-
solar metallicities. Three He II/[Ne V]-detected sources are also found in the
star-forming region of the BPT diagram and extend to lower-
metallicity solutions, consistent with $Z/Z_{\odot}\approx0.25-0.5$ ($12+\log({\rm O/H})\approx8.1-8.4$). One object shows a clear broad H$\alpha$ component; a multi-component fit yields $M_{\rm BH}\approx8.8\times10^{6} M_{\odot}$ and $\lambda_{\rm Edd}\approx0.035$, consistent with near-
solar narrow-line region
metallicity. The fraction of point-like sources in the parent target sample ($22\%\pm2\%$) is intermediate between those of
star-forming
galaxies ($12\%\pm5\%$) and AGN ($33\%\pm10\%$), consistent with a contribution from compact components. Stacked [O III] line profiles show no clear evidence for strong
outflows. Our findings suggest that high-EW NB selection efficiently identifies candidates for metal-poor AGNs.
Read the full abstract ▾
All authors Hisakazu Uchiyama, Tohru Nagao, Mariko Kubo, Tadayuki Kodama, Rhythm Shimakawa, Akio Inoue, Masayuki Tanaka, Jialai Wang, Ronaldo Laishram, Yuma Sugahara, Masato Onodera, Kazuki Daikuhara, Ko Ishida, Ryo Albert Sutanto, Haruka Kusakabe, Yoshiki Matsuoka, Alberto Rodriguez-Ardila, Yuta Suzuki, Yoshiki Toba, Xinfeng Xu, Yongquan Xue, Yongming Liang
Computer Vision
Functional Hand Type Prior for 3D Hand Pose Estimation and Action Recognition from Egocentric View Monocular Videos
Wonseok Roh, Seung Hyun Lee, Won Jeong Ryoo et al. (+5 more)
Current methods for
egocentric view action recognition often face challenges in perceiving dynamic hand movements relying solely on geometrical or physical information. In this work, we effectively address this problem by gaining insights into the correlation between functional hand configurations and objects, which improves the detailed interpretation of real-world scenarios. To this end, we introduce a practical taxonomy of hand types based on the functioning perspective and utilize it for per-frame hand type labeling on existing
datasets. We also propose a novel hand action recognition framework considering semantic details of the hand type as prior. This approach boosts the network's understanding of the continuous hand interaction throughout the action sequence. Our whole pipeline consists of three main modules: (1) Feature Extraction, (2)
Egocentric Knowledge Module, which estimates 3D hand pose, object category, and hand type leveraging short-term cues, and (2)
Egocentric Action Module, which aggregates per-frame knowledge, including text
embeddings of hand type, over a longer time. In our extensive experiments with large-scale
benchmarks, FPHA and H2O, our model outperforms current state-of-the-art methods, demonstrating its superior performance.
Read the full abstract ▾
All authors Wonseok Roh, Seung Hyun Lee, Won Jeong Ryoo, Jakyung Lee, Gyeongrok Oh, Sooyeon Hwang, Hyung-gun Chi, Sangpil Kim
Computer Vision
Antonio Jun, Haoshui Yu, Zhengyi Lu et al. (+2 more)
Vision-
Language Models (VLMs) are far more reliable at recognizing what appears in a
video than at
reasoning about its spatial and temporal properties, such as metric distances, object dimensions, and consistent object identities across frames. We present
Geometric Code, a perception-to-geometry pipeline that computes explicit spatial structure from
video and supplies it to VLMs as context to augment
reasoning. A perception layer segments and classifies objects and recovers depth, camera pose, and intrinsics from
monocular RGB
video. A deterministic
geometric engine then back-projects, merges, and cleans these outputs into a spatial code, including per-object positions, dimensions, counts, inter-object distances, appearance order, and room geometry. The code is serialized into VLMs'
prompts, either alongside the
video or replacing it entirely. Specifically, there is no component trained or fine-tuned in our approach. On VSI-Bench, augmenting 2B and 4B open models with the spatial code improves average accuracy by +4.1 points over the frames-only baseline, with the largest gains on numeric estimation tasks such as absolute distance (+24.1 points). The results suggest that explicitly computed geometry, delivered through the
language channel, recovers spatial competence that small VLMs cannot extract from pixels alone.
Read the full abstract ▾
All authors Antonio Jun, Haoshui Yu, Zhengyi Lu, Huirong Fu, Yao Qiang
Audio and Speech Processing
SPEAR-Gen: Generation-Aware Pre-training for Unified Speech Representations
Xiaoyu Yang, Arthur Hinsvark, Antonios Alexos et al. (+3 more)
Speech understanding and generation place different demands on
speech representations, and existing models are typically optimised towards one capability or the other. To reduce this gap, we introduce SPEAR-Gen, a
speech representation model that learns a single representation for both capabilities. Task-aligned feature aggregation consolidates complementary linguistic and paralinguistic information across a frozen encoder into discrete targets for masked prediction, while a coarse-to-fine objective combines log-Mel reconstruction with residual flow matching to preserve spectral structure and fine-grained acoustic variation. Experiments on SUPERB and
speech resynthesis show that SPEAR-Gen maintains strong understanding performance while substantially improving resynthesis quality and speaker preservation. These results demonstrate that a single
speech representation can effectively support both understanding and generation.
Read the full abstract ▾
All authors Xiaoyu Yang, Arthur Hinsvark, Antonios Alexos, Osama Hanna, Philip C. Woodland, Yiting Lu