arXiv subjects

Yuwen Wang

Publications and source records attributed to Yuwen Wang.

17 recordsLinked to original sources

arXiv2026

Hear, Invoke, and Understand: A Skill-Calling Multimodal Agent for Large Audio Language Models

Complex acoustic problems may require models to perform acoustic operations, interact with external tools and reason over the resulting textual or processed-audio observations rather than answer directly from a fixed audio input. We study such problems as tool-interactive audio reasoning and develop SpeechAgent-R, an audio agent that coordinates its intrinsic multimodal understanding with external skills and tools. To support this capability, we construct HIU-Corpus, comprising 65,492 interaction trajectories and 507.6 hours of audio across 24 tasks, 8 skills and 9 tools. SpeechAgent-R first learns structured interaction behaviors through trajectory-based supervised fine-tuning and then improves its decisions through multi-turn reinforcement learning. We further introduce HIU-Bench to jointly evaluate task performance, interaction quality and generalization to diverse task settings. It contains 1,395 samples across 56 tasks, including in-distribution (ID) and out-of-distribution (OOD) splits with substantial shifts in tool usage and workflow composition. SpeechAgent-R achieves 84.17 on ID tasks and 70.94 on OOD tasks, improving over the base model under the same agent harness by 15.40 and 14.23 points. These results demonstrate that learning skill and tool coordination improves audio agents' ability to handle diverse task settings and adaptive tool interactions.

cs.MM↗

arXiv2026

PALM-Bench: A Comprehensive Benchmark for Personalized Audio-Language Models

Large Audio-Language Models (LALMs) have demonstrated strong performance in audio understanding and generation. Yet, our extensive benchmarking reveals that their behavior is largely generic (e.g., summarizing spoken content) and fails to adequately support personalized question answering (e.g., summarizing what my best friend says). In contrast, human conditions their interpretation and decision-making on each individual's personal context. To bridge this gap, we formalize the task of Personalized LALMs (PALM) for recognizing personal concepts and reasoning within personal context. Moreover, we create the first benchmark (PALM-Bench) to foster the methodological advances in PALM and enable structured evaluation on several tasks across multi-speaker scenarios. Our extensive experiments on representative open-source LALMs, show that existing training-free prompting and supervised fine-tuning strategies, while yield improvements, remains limited in modeling personalized knowledge and transferring them across tasks robustly. Data and code will be released.

cs.CL↗

arXiv2025

Community detection in heterogeneous signed networks

Network data has attracted growing interest across scientific domains, prompting the development of various network models. Existing network analysis methods mainly focus on unsigned networks, whereas signed networks, consisting of both positive and negative edges, have been frequently encountered in practice but much less investigated. In this paper, we formally define strong and weak balance in signed networks, and propose a signed block $\beta$-model, which is capable of modeling strong- and weak-balanced signed networks simultaneously. We establish the identifiability of the proposed model by leveraging properties of bipartite graphs, and develop an efficient alternating updating algorithm to optimize the resulting log-likelihood function. More importantly, we establish the asymptotic consistencies of the proposed model in terms of both probability estimation and community detection. Its advantages are also demonstrated through extensive numerical experiments and the application to a real-world international relationship network.

stat.ME↗

arXiv2025

FAITH: A Framework for Assessing Intrinsic Tabular Hallucinations in Finance

Hallucination remains a critical challenge for deploying Large Language Models (LLMs) in finance. Accurate extraction and precise calculation from tabular data are essential for reliable financial analysis, since even minor numerical errors can undermine decision-making and regulatory compliance. Financial applications have unique requirements, often relying on context-dependent, numerical, and proprietary tabular data that existing hallucination benchmarks rarely capture. In this study, we develop a rigorous and scalable framework for evaluating intrinsic hallucinations in financial LLMs, conceptualized as a context-aware masked span prediction task over real-world financial documents. Our main contributions are: (1) a novel, automated dataset creation paradigm using a masking strategy; (2) a new hallucination evaluation dataset derived from S&P 500 annual reports; and (3) a comprehensive evaluation of intrinsic hallucination patterns in state-of-the-art LLMs on financial tabular data. Our work provides a robust methodology for in-house LLM evaluation and serves as a critical step toward building more trustworthy and reliable financial Generative AI systems.

cs.LG↗

arXiv2025

Nonparametric learning of heterogeneous graphical model on network-linked data

Graphical models have been popularly used for capturing conditional independence structure in multivariate data, which are often built upon independent and identically distributed observations, limiting their applicability to complex datasets such as network-linked data. This paper proposes a nonparametric graphical model that addresses these limitations by accommodating heterogeneous graph structures without imposing any specific distributional assumptions. The proposed estimation method effectively integrates network embedding with nonparametric graphical model estimation. It further transforms the graph learning task into solving a finite-dimensional linear equation system by leveraging the properties of vector-valued reproducing kernel Hilbert space. Moreover, theoretical guarantees are established for the proposed method in terms of the estimation consistency and exact recovery of the heterogeneous graph structures. Its effectiveness is also demonstrated through a variety of simulated examples and a real application to the statistician coauthorship dataset.

stat.ME↗

arXiv2025

From GNNs to Trees: Multi-Granular Interpretability for Graph Neural Networks

Interpretable Graph Neural Networks (GNNs) aim to reveal the underlying reasoning behind model predictions, attributing their decisions to specific subgraphs that are informative. However, existing subgraph-based interpretable methods suffer from an overemphasis on local structure, potentially overlooking long-range dependencies within the entire graphs. Although recent efforts that rely on graph coarsening have proven beneficial for global interpretability, they inevitably reduce the graphs to a fixed granularity. Such an inflexible way can only capture graph connectivity at a specific level, whereas real-world graph tasks often exhibit relationships at varying granularities (e.g., relevant interactions in proteins span from functional groups, to amino acids, and up to protein domains). In this paper, we introduce a novel Tree-like Interpretable Framework (TIF) for graph classification, where plain GNNs are transformed into hierarchical trees, with each level featuring coarsened graphs of different granularity as tree nodes. Specifically, TIF iteratively adopts a graph coarsening module to compress original graphs (i.e., root nodes of trees) into increasingly coarser ones (i.e., child nodes of trees), while preserving diversity among tree nodes within different branches through a dedicated graph perturbation module. Finally, we propose an adaptive routing module to identify the most informative root-to-leaf paths, providing not only the final prediction but also the multi-granular interpretability for the decision-making process. Extensive experiments on the graph classification benchmarks with both synthetic and real-world datasets demonstrate the superiority of TIF in interpretability, while also delivering a competitive prediction performance akin to the state-of-the-art counterparts.

cs.LG↗

arXiv2024

SPHERE: Unveiling Spatial Blind Spots in Vision-Language Models Through Hierarchical Evaluation

Current vision-language models may grasp basic spatial cues and simple directions (e.g. left, right, front, back), but struggle with the multi-dimensional spatial reasoning necessary for human-like understanding and real-world applications. To address this gap, we develop SPHERE (Spatial Perception and Hierarchical Evaluation of REasoning), a hierarchical evaluation framework supported by a new human-annotated dataset. SPHERE systematically probes models across increasing levels of complexity, from fundamental skills to multi-skill integration and high-level reasoning that combines spatial, visual, and logical understanding. Benchmark evaluation of state-of-the-art models reveals significant deficiencies, especially in reasoning about distance and proximity, understanding both egocentric and allocentric perspectives, and applying spatial logic in physical contexts. These findings expose critical blind spots in existing models and underscore the need for more advanced spatial reasoning techniques, driving the development of vision-language models that align more closely with human spatial cognition. The SPHERE benchmark is available at https://github.com/zwenyu/SPHERE-VLM.

cs.CV↗

arXiv2024

Unveiling Global Interactive Patterns across Graphs: Towards Interpretable Graph Neural Networks

Graph Neural Networks (GNNs) have emerged as a prominent framework for graph mining, leading to significant advances across various domains. Stemmed from the node-wise representations of GNNs, existing explanation studies have embraced the subgraph-specific viewpoint that attributes the decision results to the salient features and local structures of nodes. However, graph-level tasks necessitate long-range dependencies and global interactions for advanced GNNs, deviating significantly from subgraph-specific explanations. To bridge this gap, this paper proposes a novel intrinsically interpretable scheme for graph classification, termed as Global Interactive Pattern (GIP) learning, which introduces learnable global interactive patterns to explicitly interpret decisions. GIP first tackles the complexity of interpretation by clustering numerous nodes using a constrained graph clustering module. Then, it matches the coarsened global interactive instance with a batch of self-interpretable graph prototypes, thereby facilitating a transparent graph-level reasoning process. Extensive experiments conducted on both synthetic and real-world benchmarks demonstrate that the proposed GIP yields significantly superior interpretability and competitive performance to~the state-of-the-art counterparts. Our code will be made publicly available.

cs.LG↗

arXiv2024

Temporal Prototype-Aware Learning for Active Voltage Control on Power Distribution Networks

Active Voltage Control (AVC) on the Power Distribution Networks (PDNs) aims to stabilize the voltage levels to ensure efficient and reliable operation of power systems. With the increasing integration of distributed energy resources, recent efforts have explored employing multi-agent reinforcement learning (MARL) techniques to realize effective AVC. Existing methods mainly focus on the acquisition of short-term AVC strategies, i.e., only learning AVC within the short-term training trajectories of a singular diurnal cycle. However, due to the dynamic nature of load demands and renewable energy, the operation states of real-world PDNs may exhibit significant distribution shifts across varying timescales (e.g., daily and seasonal changes). This can render those short-term strategies suboptimal or even obsolete when performing continuous AVC over extended periods. In this paper, we propose a novel temporal prototype-aware learning method, abbreviated as TPA, to learn time-adaptive AVC under short-term training trajectories. At the heart of TPA are two complementary components, namely multi-scale dynamic encoder and temporal prototype-aware policy, that can be readily incorporated into various MARL methods. The former component integrates a stacked transformer network to learn underlying temporal dependencies at different timescales of the PDNs, while the latter implements a learnable prototype matching mechanism to construct a dedicated AVC policy that can dynamically adapt to the evolving operation states. Experimental results on the AVC benchmark with different PDN sizes demonstrate that the proposed TPA surpasses the state-of-the-art counterparts not only in terms of control performance but also by offering model transferability. Our code is available at https://github.com/Canyizl/TPA-for-AVC.

cs.LG↗

arXiv2024

Hidden Markov Models and the Bayes Filter in Categorical Probability

We use Markov categories to generalize the basic theory of Markov chains and hidden Markov models to an abstract setting. This comprises characterizations of hidden Markov models in terms of conditional independences and algorithms for Bayesian filtering and smoothing applicable in all Markov categories with conditionals. When instantiated in appropriate Markov categories, these algorithms specialize to existing ones such as the Kalman filter, forward-backward algorithm, and the Rauch-Tung-Striebel smoother. We also prove that the sequence of outputs of our abstract Bayes filter is itself a Markov chain with a concrete formula for its transition maps. There are two main features of this categorical framework. The first is its abstract generality, as manifested in our unified account of hidden Markov models and algorithms for filtering and smoothing in discrete probability, Gaussian probability, measure-theoretic probability, possibilistic nondeterminism and others at the same time. The second feature is the intuitive visual representation of information flow in terms of string diagrams.

math.ST↗

arXiv2023

Expected hitting time estimates on finite graphs

The expected hitting time from vertex $a$ to vertex $b$, $H(a,b)$, is the expected value of the time it takes a random walk starting at $a$ to reach $b$. In this paper, we give estimates for $H(a,b)$ when the distance between $a$ and $b$ is comparable to the diameter of the graph, and the graph satisfies a Harnack condition. We show that, in such cases, $H(a,b)$ can be estimated in terms of the volumes of balls around $b$. Using our results, we estimate $H(a,b)$ on various graphs, such as rectangular tori, some convex traces in $\mathbb{Z}^d$, and fractal graphs. Our proofs use heat kernel estimates.

math.PR↗

arXiv2023

Adversarial Erasing with Pruned Elements: Towards Better Graph Lottery Ticket

Graph Lottery Ticket (GLT), a combination of core subgraph and sparse subnetwork, has been proposed to mitigate the computational cost of deep Graph Neural Networks (GNNs) on large input graphs while preserving original performance. However, the winning GLTs in exisiting studies are obtained by applying iterative magnitude-based pruning (IMP) without re-evaluating and re-considering the pruned information, which disregards the dynamic changes in the significance of edges/weights during graph/model structure pruning, and thus limits the appeal of the winning tickets. In this paper, we formulate a conjecture, i.e., existing overlooked valuable information in the pruned graph connections and model parameters which can be re-grouped into GLT to enhance the final performance. Specifically, we propose an adversarial complementary erasing (ACE) framework to explore the valuable information from the pruned components, thereby developing a more powerful GLT, referred to as the ACE-GLT. The main idea is to mine valuable information from pruned edges/weights after each round of IMP, and employ the ACE technique to refine the GLT processing. Finally, experimental results demonstrate that our ACE-GLT outperforms existing methods for searching GLT in diverse tasks. Our code will be made publicly available.

cs.LG↗

arXiv2022

Central Limit Theorems in Deterministic Systems

This is a note on some results of the central limit theorem for deterministic dynamical systems. First, we give the central limit theorem for martingales, which is a main tool. Then we give the main results on the central limit theorem in dynamic system in the cases of martingale and backward martingale

math.PR↗

arXiv2022

Abelian sandpiles on Sierpinski gasket graphs

The aim of the current work is to investigate structural properties of the sandpile group of a special class of self-similar graphs. More precisely, we consider Abelian sandpiles on Sierpinski gasket graphs and for the choice of normal boundary conditions, we give a characterization of the identity element and a recursive description of the sandpile group. Finally, we consider Abelian sandpile Markov chains on the aforementioned graphs and we improve the existing bounds on the speed of convergence to stationarity.

math.CO↗

arXiv2022

Random walks on finite nilpotent groups driven by long-jump measures

We consider a variant of random walks on finite groups. At each step, we choose an element from a set of generators ("directions") uniformly, and an integer from a power law ("speed") distribution associated with the chosen direction. We show that if the finite group is nilpotent, the mixing time of this walk is of the same order of magnitude as the diameter of a suitable pseudo-metric, D(S, a), on the group, which depends only on the generators and speeds. Additionally, we give sharp bounds on the $\ell^2$-distance between the distribution of the position of the walker and the stationary distribution, and compute D(S,a) for some examples.

math.PR↗

arXiv2019

The Impact of Tribalism on Social Welfare

We explore the impact of mutual altruism among the players belonging to the same set -- their tribe -- in a partition of all players in arbitrary strategic games upon the quality of equilibria attained. To this end, we introduce the notion of a τ-tribal extension of an arbitrary strategic game, in which players' subjective cost functions are updated to reflect this, and the associated Price of Tribalism, which is the ratio of the social welfare of the worst Nash equilibrium of the tribal extension to that of the optimum of social welfare. We show that in a well-known game of friendship cliques, network contribution games as well as atomic linear congestion games, the Price of Tribalism is higher than the Price of Anarchy of either the purely selfish players or fully altruistic players (i.e. ones who seek to maximise the social welfare). This phenomenon is observed under a variety of equilibrium concepts. In each instance, we present upper bounds on the Price of Tribalism that match the lower bounds established by our example.

cs.GT↗

arXiv2013

On curves and polygons with the equiangular chord property

Let $C$ be a smooth, convex curve on either the sphere $\mathbb{S}^{2}$, the hyperbolic plane $\mathbb{H}^{2}$ or the Euclidean plane $\mathbb{E}^{2}$, with the following property: there exists $α$, and parameterizations $x(t), y(t)$ of $C$ such that for each $t$, the angle between the chord connecting $x(t)$ to $y(t)$ and $C$ is $α$ at both ends. Assuming that $C$ is not a circle, E. Gutkin completely characterized the angles $α$ for which such a curve exists in the Euclidean case. We study the infinitesimal version of this problem in the context of the other two constant curvature geometries, and in particular we provide a complete characterization of the angles $α$ for which there exists a non-trivial infinitesimal deformation of a circle through such curves with corresponding angle $α$. We also consider a discrete version of this property for Euclidean polygons, and in this case we give a complete description of all non-trivial solutions.

math.DG↗