Searcharxiv⌕ Search

arXiv subjects

Zhengdong Wang

Publications and source records attributed to Zhengdong Wang.

17 recordsLinked to original sources
arXiv2026

Comprehensive AI governance requires addressing non-model gains

Frontier AI governance often centres on the model-level governance paradigm, which assumes that a model's capability profile is primarily a function of the compute and data used during training. This position paper argues that model-level governance becomes less effective when capability progress is increasingly driven by "non-model gains"--improvements that are independent from advances in the base model. We formalise the concept of non-model gains and provide a taxonomy of three distinct vectors of capability gain: inference gain (scaling compute at test-time), systems gain (post-training enhancements such as scaffolds), and asset gain (enhancing a model with restricted assets). We demonstrate how these vectors--alongside potential future impacts from embodiment, continual learning, and AI diffusion--may undermine risk management strategies that hinge mostly on pre-deployment evaluation and mitigation. We provide an overview of governance approaches that go beyond the model level: system, entity, agent, and cloud governance. Finally, we emphasise the importance of societal resilience as a complement to these governance layers.

cs.CY↗
arXiv2026

Joint Optimization of Storage and Loading for High-Performance 3D Point Cloud Data Processing

With the rapid development of computer vision and deep learning, significant advancements have been made in 3D vision, partic- ularly in autonomous driving, robotic perception, and augmented reality. 3D point cloud data, as a crucial representation of 3D information, has gained widespread attention. However, the vast scale and complexity of point cloud data present significant chal- lenges for loading and processing and traditional algorithms struggle to handle large-scale datasets.The diversity of storage formats for point cloud datasets (e.g., PLY, XYZ, BIN) adds complexity to data handling and results in inefficiencies in data preparation. Al- though binary formats like BIN and NPY have been used to speed up data access, they still do not fully address the time-consuming data loading and processing phase. To overcome these challenges, we propose the .PcRecord format, a unified data storage solution designed to reduce the storage occupation and accelerate the processing of point cloud data. We also introduce a high-performance data processing pipeline equipped with multiple modules. By leveraging a multi-stage parallel pipeline architecture, our system optimizes the use of computational resources, significantly improving processing speed and efficiency. This paper details the im- plementation of this system and demonstrates its effectiveness in addressing the challenges of handling large-scale point cloud datasets.On average, our system achieves performance improvements of 6.61x (ModelNet40), 2.69x (S3DIS), 2.23x (ShapeNet), 3.09x (Kitti), 8.07x (SUN RGB-D), and 5.67x (ScanNet) with GPU and 6.9x, 1.88x, 1.29x, 2.28x, 25.4x, and 19.3x with Ascend.

cs.CV↗
arXiv2025

SIMA 2: A Generalist Embodied Agent for Virtual Worlds

We introduce SIMA 2, a generalist embodied agent that understands and acts in a wide variety of 3D virtual worlds. Built upon a Gemini foundation model, SIMA 2 represents a significant step toward active, goal-directed interaction within an embodied environment. Unlike prior work (e.g., SIMA 1) limited to simple language commands, SIMA 2 acts as an interactive partner, capable of reasoning about high-level goals, conversing with the user, and handling complex instructions given through language and images. Across a diverse portfolio of games, SIMA 2 substantially closes the gap with human performance and demonstrates robust generalization to previously unseen environments, all while retaining the base model's core reasoning capabilities. Furthermore, we demonstrate a capacity for open-ended self-improvement: by leveraging Gemini to generate tasks and provide rewards, SIMA 2 can autonomously learn new skills from scratch in a new environment. This work validates a path toward creating versatile and continuously learning agents for both virtual and, eventually, physical worlds.

cs.AI↗
arXiv2024

Scaling Instructable Agents Across Many Simulated Worlds

Building embodied AI systems that can follow arbitrary language instructions in any 3D environment is a key challenge for creating general AI. Accomplishing this goal requires learning to ground language in perception and embodied actions, in order to accomplish complex tasks. The Scalable, Instructable, Multiworld Agent (SIMA) project tackles this by training agents to follow free-form instructions across a diverse range of virtual 3D environments, including curated research environments as well as open-ended, commercial video games. Our goal is to develop an instructable agent that can accomplish anything a human can do in any simulated 3D environment. Our approach focuses on language-driven generality while imposing minimal assumptions. Our agents interact with environments in real-time using a generic, human-like interface: the inputs are image observations and language instructions and the outputs are keyboard-and-mouse actions. This general approach is challenging, but it allows agents to ground language across many visually complex and semantically rich environments while also allowing us to readily run agents in new environments. In this paper we describe our motivation and goal, the initial progress we have made, and promising preliminary results on several diverse research environments and a variety of commercial video games.

cs.RO↗
arXiv2023

Reconstructing high-order sequence features of dynamic functional connectivity networks based on diversified covert attention patterns for Alzheimer's disease classification

Recent studies have applied deep learning methods such as convolutional recurrent neural networks (CRNs) and Transformers to brain disease classification based on dynamic functional connectivity networks (dFCNs), such as Alzheimer's disease (AD), achieving better performance than traditional machine learning methods. However, in CRNs, the continuous convolution operations used to obtain high-order aggregation features may overlook the non-linear correlation between different brain regions due to the essence of convolution being the linear weighted sum of local elements. Inspired by modern neuroscience on the research of covert attention in the nervous system, we introduce the self-attention mechanism, a core module of Transformers, to model diversified covert attention patterns and apply these patterns to reconstruct high-order sequence features of dFCNs in order to learn complex dynamic changes in brain information flow. Therefore, we propose a novel CRN method based on diversified covert attention patterns, DCA-CRN, which combines the advantages of CRNs in capturing local spatio-temporal features and sequence change patterns, as well as Transformers in learning global and high-order correlation features. Experimental results on the ADNI and ADHD-200 datasets demonstrate the prediction performance and generalization ability of our proposed method.

eess.IV↗
arXiv2022

BEKG: A Built Environment Knowledge Graph

Practices in the built environment have become more digitalized with the rapid development of modern design and construction technologies. However, the requirement of practitioners or scholars to gather complicated professional knowledge in the built environment has not been satisfied yet. In this paper, more than 80,000 paper abstracts in the built environment field were obtained to build a knowledge graph, a knowledge base storing entities and their connective relations in a graph-structured data model. To ensure the retrieval accuracy of the entities and relations in the knowledge graph, two well-annotated datasets have been created, containing 2,000 instances and 1,450 instances each in 29 relations for the named entity recognition task and relation extraction task respectively. These two tasks were solved by two BERT-based models trained on the proposed dataset. Both models attained an accuracy above 85% on these two tasks. More than 200,000 high-quality relations and entities were obtained using these models to extract all abstract data. Finally, this knowledge graph is presented as a self-developed visualization system to reveal relations between various entities in the domain. Both the source code and the annotated dataset can be found here: https://github.com/HKUST-KnowComp/BEKG.

cs.CL↗
arXiv2020

Galaxy Morphology Network: A Convolutional Neural Network Used to Study Morphology and Quenching in $\sim 100,000$ SDSS and $\sim 20,000$ CANDELS Galaxies

We examine morphology-separated color-mass diagrams to study the quenching of star formation in $\sim 100,000$ ($z\sim0$) Sloan Digital Sky Survey (SDSS) and $\sim 20,000$ ($z\sim1$) Cosmic Assembly Near-Infrared Deep Extragalactic Legacy Survey (CANDELS) galaxies. To classify galaxies morphologically, we developed Galaxy Morphology Network (GaMorNet), a convolutional neural network that classifies galaxies according to their bulge-to-total light ratio. GaMorNet does not need a large training set of real data and can be applied to data sets with a range of signal-to-noise ratios and spatial resolutions. GaMorNet's source code as well as the trained models are made public as part of this work ( http://www.astro.yale.edu/aghosh/gamornet.html ). We first trained GaMorNet on simulations of galaxies with a bulge and a disk component and then transfer learned using $\sim25\%$ of each data set to achieve misclassification rates of $\lesssim5\%$. The misclassified sample of galaxies is dominated by small galaxies with low signal-to-noise ratios. Using the GaMorNet classifications, we find that bulge- and disk-dominated galaxies have distinct color-mass diagrams, in agreement with previous studies. For both SDSS and CANDELS galaxies, disk-dominated galaxies peak in the blue cloud, across a broad range of masses, consistent with the slow exhaustion of star-forming gas with no rapid quenching. A small population of red disks is found at high mass ($\sim14\%$ of disks at $z\sim0$ and $2\%$ of disks at $z \sim 1$). In contrast, bulge-dominated galaxies are mostly red, with much smaller numbers down toward the blue cloud, suggesting rapid quenching and fast evolution across the green valley. This inferred difference in quenching mechanism is in agreement with previous studies that used other morphology classification techniques on much smaller samples at $z\sim0$ and $z\sim1$.

astro-ph.GA↗
arXiv2009

Nonabelian cohomology of compact Lie groups

Given a Lie group $G$ with finitely many components and a compact Lie group A which acts on $G$ by automorphisms, we prove that there always exists an A-invariant maximal compact subgroup K of G, and that for every such K, the natural map $H^1(A,K)\to H^1(A,G)$ is bijective. This generalizes a classical result of Serre [6] and a recent result in [1].

math.GR↗
arXiv2007

Holomorphic harmonic analysis on complex reductive groups

We define the holomorphic Fourier transform of holomorphic functions on complex reductive groups, prove some properties like the Fourier inversion formula, and give some applications. The definition of the holomorphic Fourier transform makes use of the notion of $K$-admissible measures. We prove that $K$-admissible measures are abundant, and the definition of holomorphic Fourier transform is independent of the choice of $K$-admissible measures.

math.GR↗
arXiv2006

On the realization of Riemannian symmetric spaces in Lie groups II

In this paper we generalize a result in [1], showing that an arbitrary Riemannian symmetric space can be realized as a closed submanifold of a covering group of the Lie group defining the symmetric space. Some properties of the subgroups of fixed points of involutions are also proved.

math.GT↗
arXiv2006

Nonabelian cohomology with coefficients in Lie groups

In this paper we prove some properties of the nonabelian cohomology $H^1(A,G)$ of a group $A$ with coefficients in a connected Lie group $G$. When $A$ is finite, we show that for every $A$-submodule $K$ of $G$ which is a maximal compact subgroup of $G$, the canonical map $H^1(A,K)\to H^1(A,G)$ is bijective. In this case we also show that $H^1(A,G)$ is always finite. When $A=\ZZ$ and $G$ is compact, we show that for every maximal torus $T$ of the identity component $G_0^\ZZ$ of the group of invariants $G^\ZZ$, $H^1(\ZZ,T)\to H^1(\ZZ,G)$ is surjective if and only if the $\ZZ$-action on $G$ is 1-semisimple, which is also equivalent to that all fibers of $H^1(\ZZ,T)\to H^1(\ZZ,G)$ are finite. When $A=\Zn$, we show that $H^1(\Zn,T)\to H^1(\Zn,G)$ is always surjective, where $T$ is a maximal compact torus of the identity component $G_0^{\Zn}$ of $G^{\Zn}$. When $A$ is cyclic, we also interpret some properties of $H^1(A,G)$ in terms of twisted conjugate actions of $G$.

math.GR↗
arXiv2005

Curve Selection Lemma for semianalytic sets and conjugacy classes of finite order in Lie groups

Using a strong version of the Curve Selection Lemma for real semianalytic sets, we prove that for an arbitrary connected Lie group $G$, each connected component of the set $E_n(G)$ of all elements of order $n$ in $G$ is a conjugacy class in $G$. In particular, all conjugacy classes of finite order in $G$ are closed. Some properties of connected components of $E_n(G)$ are also given.

math.GR↗
arXiv2005

On the realization of Riemannian symmetric spaces in Lie groups

In this paper we give a realization of some symmetric space G/K as a closed submanifold P of G. We also give several equivalent representations of the submanifold P. Some properties of the set gK\cap P are also discussed, where gK is a coset space in G.

math.GT↗
arXiv2005

The Level Densities of Random Matrix Unitary Ensembles and their Perturbation Invariability

Using operator methods, we generally present the level densities for kinds of random matrix unitary ensembles in weak sense. As a corollary, the limit spectral distributions of random matrices from Gaussian, Laguerre and Jacobi unitary ensembles are recovered. At the same time, we study the perturbation invariability of the level densities of random matrix unitary ensembles. After the weight function associated with the 1-level correlation function is appended a polynomial multiplicative factor, the level density is invariant in the weak sense.

math-ph↗
arXiv2005

A Generalization of Random Matrix Ensemble I: General Theory

We give a generalization of the random matrix ensembles, including all lassical ensembles. Then we derive the joint density function of the generalized ensemble by one simple formula, which give a direct and unified way to compute the density functions for all classical ensembles and various kinds of new ensembles. An integration formula associated with the generalized ensemble is also given. We also give a classification scheme of the generalized ensembles, which will include all classical ensembles and some new ensembles which were not considered before.

math-ph↗
arXiv2005

A Generalization of Random Matrix Ensemble II: Concrete Examples and Integration Formulae

According to the classification scheme of the generalized random matrix ensembles, we present various kinds of concrete examples of the generalized ensemble, and derive their joint density functions in an unified way by one simple formula which was proved in [2]. Particular cases of these examples include Gaussian ensemble, chiral ensemble, new transfer matrix ensembles, circular ensemble, Jacobi ensembles, and so on. The associated integration formulae are also given, which are just many classical integration formulae or their variation forms.

math-ph↗