arXiv subjects

Zhenqi Wang

Publications and source records attributed to Zhenqi Wang.

8 recordsLinked to original sources

arXiv2026

Observe Less, Understand More: Cost-aware Cross-scale Observation for Remote Sensing Understanding

Remote sensing understanding inherently requires multi-resolution observation, since different targets and application tasks demand different levels of spatial detail. While low-resolution (LR) imagery enables efficient global observation, high-resolution (HR) imagery provides critical local details at a much higher acquisition cost and with limited coverage. This motivates a cross-scale sensing strategy that selectively acquires HR imagery guided by LR-based global perception to improve task performance under constrained cost. Existing HR sampling methods typically make selection decisions from isolated LR patches, thereby ignoring fine-grained intra-patch importance and cross-patch contextual interactions, leading to fragmented feature representation and suboptimal scene reasoning under sparse HR observations. To address this issue, we formulate cross-scale remote sensing understanding as a unified cost-aware problem that couples fine-grained HR sampling with cross-patch representation prediction, enabling more effective task reasoning with fewer HR observations. Furthermore, we present GL-10M, a high- and low-resolution dataset with nearly 100,000 scene pairs and 10 million images for large-scale cross-resolution pretraining. Extensive experiments on recognition and retrieval tasks show that our method consistently achieves a superior performance-cost trade-off. The code is publicly available at https://github.com/xzhacc/CrossSO.

cs.CV↗

arXiv2025

Multi-Agents Based on Large Language Models for Knowledge-based Visual Question Answering

Large Language Models (LLMs) have achieved impressive results in knowledge-based Visual Question Answering (VQA). However existing methods still have challenges: the inability to use external tools autonomously, and the inability to work in teams. Humans tend to know whether they need to use external tools when they encounter a new question, e.g., they tend to be able to give a direct answer to a familiar question, whereas they tend to use tools such as search engines when they encounter an unfamiliar question. In addition, humans also tend to collaborate and discuss with others to get better answers. Inspired by this, we propose the multi-agent voting framework. We design three LLM-based agents that simulate different levels of staff in a team, and assign the available tools according to the levels. Each agent provides the corresponding answer, and finally all the answers provided by the agents are voted to get the final answer. Experiments on OK-VQA and A-OKVQA show that our approach outperforms other baselines by 2.2 and 1.0, respectively.

cs.CL↗

arXiv2024

Global smooth rigidity for toral automorphisms

We study regularity of a conjugacy between a hyperbolic or partially hyperbolic toral automorphism $L$ and a $C^\infty$ diffeomorphism $f$ of the torus. For a very weakly irreducible hyperbolic automorphism $L$ we show that any $C^1$ conjugacy is $C^\infty$. For a very weakly irreducible ergodic partially hyperbolic automorphism $L$ we show that any $C^{1+\text{Hölder}}$ conjugacy is $C^\infty$. As a corollary, we improve regularity of the conjugacy to $C^\infty$ in prior local and global rigidity results.

math.DS↗

arXiv2022

Robust N-1 secure HV Grid Flexibility Estimation for TSO-DSO coordinated Congestion Management with Deep Reinforcement Learning

Nowadays, the PQ flexibility from the distributed energy resources (DERs) in the high voltage (HV) grids plays a more critical and significant role in grid congestion management in TSO grids. This work proposed a multi-stage deep reinforcement learning approach to estimate the PQ flexibility (PQ area) at the TSO-DSO interfaces and identifies the DER PQ setpoints for each operating point in a way, that DERs in the meshed HV grid can be coordinated to offer flexibility for the transmission grid. In the estimation process, we consider the steady-state grid limits and the robustness in the resulting voltage profile against uncertainties and the N-1 security criterion regarding thermal line loading, essential for real-life grid operational planning applications. Using deep reinforcement learning (DRL) for PQ flexibility estimation is the first of its kind. Furthermore, our approach of considering N-1 security criterion for meshed grids and robustness against uncertainty directly in the optimization tasks offers a new perspective besides the common relaxation schema in finding a solution with mathematical optimal power flow (OPF). Finally, significant improvements in the computational efficiency in estimation PQ area are the highlights of the proposed method.

eess.SY↗

arXiv2021

Fast Parallel Newton-Raphson Power Flow Solver for Large Number of System Calculations with CPU and GPU

To analyze large sets of grid states, e.g. when evaluating the impact from the uncertainties of the renewable generation with probabilistic Monte Carlo simulation or in stationary time series simulation, large number of power flow calculations have to be performed. For the application in real-time grid operation, grid planning and in further cases when computational time is critical, a novel approach on simultaneous parallelization of many Newton-Raphson power flow calculations on CPU and with GPU-acceleration is proposed. The result shows a speed-up of over x100 comparing to the open-source tool pandapower, when performing repetitive power flows of system with admittance matrix of the same sparsity pattern on both CPU and GPU. The speed-up relies on the algorithm improvement and highly optimized parallelization strategy, which can reduce the repetitive work and saturate the high hardware computational capability of modern CPUs and GPUs well. This is achieved with the proposed batched sparse matrix operation and batched linear solver based on LU-refactorization. The batched linear solver shows a large performance improvement comparing to the state-of-the-art linear system solver KLU library and a better saturation of the GPU performance with small problem scale. Finally, the method of integrating the proposed solver into pandapower is presented, thus the parallel power flow solver with outstanding performance can be easily applied in challenging real-life grid operation and innovative researches e.g. data-driven machine learning studies.

cs.CE↗

arXiv2020

On globally hypoelliptic abelian actions and their existence on homogeneous spaces

We define globally hypoelliptic smooth $\mathbb R^k$ actions as actions whose leafwise Laplacian along the orbit foliation is a globally hypoelliptic differential operator. When $k=1$, strong global rigidity is conjectured for such actions by Greenfield-Wallach and Katok: every such action is smoothly conjugate to a Diophantine flow on the torus. The conjecture has been confirmed for all homogeneous flows on homogeneous spaces \cite{FFRH}. In this paper we conjecture that among homogeneous $\mathbb R^k$ actions ($k\ge 2$) on homogeneous spaces globally hypoelliptic actions exist only on nilmanifolds. We obtain a partial result towards this conjecture: we show non-existence of globally hypoelliptic $\mathbb R^2$ actions on homogeneous spaces $G/Γ$, with at least one quasi-unipotent generator, where $G= SL(n, \mathbb R)$. We also show that the same type of actions on solvmanifolds are smoothly conjugate to homogeneous actions on nilmanifolds.

math.DS↗

arXiv2009

New Cases of Differential Rigidity for Non-Generic Partially Hyperbolic Actions

We prove the locally differentiable rigidity of generic partially hyperbolic abelian algebraic high-rank actions on compact homogeneous spaces obtained from split symplectic Lie groups. We also gave a non-generic action rigidity example on compact homogeneous spaces obtained from SL(2n,R) or SL(2n,C). The conclusions are based on geometric Katok-Damjanovic way and progress towards computations of the generating relations in these groups.

math.DS↗

arXiv2009

Local Rigidity of Partially Hyperbolic Actions

We consider partially hyperbolic abelian algebraic high-rank actions on compact homogeneous spaces obtained from simple indefinite orthogonal and unitary groups. In the first part of the paper, we show local differentiable rigidity for such actions. The conclusions are based on progress towards computations of the Schur multipliers of these non-split groups, which is the main aim of the second part.

math.DS↗