arXiv subjects

Hanwen Zhang

Publications and source records attributed to Hanwen Zhang.

At least 73 records · Page 4Linked to original sources

arXiv2023

Automated interpretation of congenital heart disease from multi-view echocardiograms

Congenital heart disease (CHD) is the most common birth defect and the leading cause of neonate death in China. Clinical diagnosis can be based on the selected 2D key-frames from five views. Limited by the availability of multi-view data, most methods have to rely on the insufficient single view analysis. This study proposes to automatically analyze the multi-view echocardiograms with a practical end-to-end framework. We collect the five-view echocardiograms video records of 1308 subjects (including normal controls, ventricular septal defect (VSD) patients and atrial septal defect (ASD) patients) with both disease labels and standard-view key-frame labels. Depthwise separable convolution-based multi-channel networks are adopted to largely reduce the network parameters. We also approach the imbalanced class problem by augmenting the positive training samples. Our 2D key-frame model can diagnose CHD or negative samples with an accuracy of 95.4\%, and in negative, VSD or ASD classification with an accuracy of 92.3\%. To further alleviate the work of key-frame selection in real-world implementation, we propose an adaptive soft attention scheme to directly explore the raw video data. Four kinds of neural aggregation methods are systematically investigated to fuse the information of an arbitrary number of frames in a video. Moreover, with a view detection module, the system can work without the view records. Our video-based model can diagnose with an accuracy of 93.9\% (binary classification), and 92.1\% (3-class classification) in a collected 2D video testing set, which does not need key-frame selection and view annotation in testing. The detailed ablation study and the interpretability analysis are provided.

eess.IV↗

arXiv2023

Filling the Missing: Exploring Generative AI for Enhanced Federated Learning over Heterogeneous Mobile Edge Devices

Distributed Artificial Intelligence (AI) model training over mobile edge networks encounters significant challenges due to the data and resource heterogeneity of edge devices. The former hampers the convergence rate of the global model, while the latter diminishes the devices' resource utilization efficiency. In this paper, we propose a generative AI-empowered federated learning to address these challenges by leveraging the idea of FIlling the MIssing (FIMI) portion of local data. Specifically, FIMI can be considered as a resource-aware data augmentation method that effectively mitigates the data heterogeneity while ensuring efficient FL training. We first quantify the relationship between the training data amount and the learning performance. We then study the FIMI optimization problem with the objective of minimizing the device-side overall energy consumption subject to required learning performance constraints. The decomposition-based analysis and the cross-entropy searching method are leveraged to derive the solution, where each device is assigned suitable AI-synthesized data and resource utilization policy. Experiment results demonstrate that FIMI can save up to 50% of the device-side energy to achieve the target global test accuracy in comparison with the existing methods. Meanwhile, FIMI can significantly enhance the converged global accuracy under the non-independently-and-identically distribution (non-IID) data.

cs.LG↗

arXiv2023

Energy Efficient Robust Beamforming for Vehicular ISAC with Imperfect Channel Estimation

This paper investigates robust beamforming for system-centric energy efficiency (EE) optimization in the vehicular integrated sensing and communication (ISAC) system, where the mobility of vehicles poses significant challenges to channel estimation. To obtain the optimal beamforming under channel uncertainty, we first formulate an optimization problem for maximizing the system EE under bounded channel estimation errors. Next, fractional programming and semidefinite relaxation (SDR) are utilized to relax the rank-1 constraints. We further use Schur complement and S-Procedure to transform Cramer-Rao bound (CRB) and channel estimation error constraints into convex forms, respectively. Based on the Lagrangian dual function and Karush-Kuhn-Tucker (KKT) conditions, it is proved that the optimal beamforming solution is rank-1. Finally, we present comprehensive simulation results to demonstrate two key findings: 1) the proposed algorithm exhibits a favorable convergence rate, and 2) the approach effectively mitigates the impact of channel estimation errors.

cs.IT↗

arXiv2023

NP-Hardness of Tensor Network Contraction Ordering

We study the optimal order (or sequence) of contracting a tensor network with a minimal computational cost. We conclude 2 different versions of this optimal sequence: that minimize the operation number (OMS) and that minimize the time complexity (CMS). Existing results only shows that OMS is NP-hard, but no conclusion on CMS problem. In this work, we firstly reduce CMS to CMS-0, which is a sub-problem of CMS with no free indices. Then we prove that CMS is easier than OMS, both in general and in tree cases. Last but not least, we prove that CMS is still NP-hard. Based on our results, we have built up relationships of hardness of different tensor network contraction problems.

cs.CC↗

arXiv2023

A monotone numerical integration method for mean-variance portfolio optimization under jump-diffusion models

We develop a efficient, easy-to-implement, and strictly monotone numerical integration method for Mean-Variance (MV) portfolio optimization in realistic contexts, which involve jump-diffusion dynamics of the underlying controlled processes, discrete rebalancing, and the application of investment constraints, namely no-bankruptcy and leverage. A crucial element of the MV portfolio optimization formulation over each rebalancing interval is a convolution integral, which involves a conditional density of the logarithm of the amount invested in the risky asset. Using a known closed-form expression for the Fourier transform of this density, we derive an infinite series representation for the conditional density where each term is strictly positive and explicitly computable. As a result, the convolution integral can be readily approximated through a monotone integration scheme, such as a composite quadrature rule typically available in most programming languages. The computational complexity of our method is an order of magnitude lower than that of existing monotone finite difference methods. To further enhance efficiency, we propose an implementation of the scheme via Fast Fourier Transforms, exploiting the Toeplitz matrix structure. The proposed monotone scheme is proven to be both $\ell_{\infty}$-stable and pointwise consistent, and we rigorously establish its pointwise convergence to the unique solution of the MV portfolio optimization problem. We also intuitively demonstrate that, as the rebalancing time interval approaches zero, the proposed scheme converges to a continuously observed impulse control formulation for MV optimization expressed as a Hamilton-Jacobi-Bellman Quasi-Variational Inequality. Numerical results show remarkable agreement with benchmark solutions obtained through finite differences and Monte Carlo simulation, underscoring the effectiveness of our approach.

q-fin.CP↗

arXiv2023

H2-Mapping: Real-time Dense Mapping Using Hierarchical Hybrid Representation

Constructing a high-quality dense map in real-time is essential for robotics, AR/VR, and digital twins applications. As Neural Radiance Field (NeRF) greatly improves the mapping performance, in this paper, we propose a NeRF-based mapping method that enables higher-quality reconstruction and real-time capability even on edge computers. Specifically, we propose a novel hierarchical hybrid representation that leverages implicit multiresolution hash encoding aided by explicit octree SDF priors, describing the scene at different levels of detail. This representation allows for fast scene geometry initialization and makes scene geometry easier to learn. Besides, we present a coverage-maximizing keyframe selection strategy to address the forgetting issue and enhance mapping quality, particularly in marginal areas. To the best of our knowledge, our method is the first to achieve high-quality NeRF-based mapping on edge computers of handheld devices and quadrotors in real-time. Experiments demonstrate that our method outperforms existing NeRF-based mapping methods in geometry accuracy, texture realism, and time consumption. The code will be released at: https://github.com/SYSU-STAR/H2-Mapping

cs.RO↗

arXiv2023

Fullwave design of cm-scale cylindrical metasurfaces via fast direct solvers

Large-scale metasurfaces promise nanophotonic performance improvements to macroscopic optics functionality, for applications from imaging to analog computing. Yet the size scale mismatch of centimeter-scale chips versus micron-scale wavelengths prohibits use of conventional full-wave simulation techniques, and has necessitated dramatic approximations. Here, we show that tailoring "fast direct" integral-equation simulation techniques to the form factor of metasurfaces offers the possibility for accurate and efficient full-wave, large-scale metasurface simulations. For cylindrical (two-dimensional) metasurfaces, we demonstrate accurate simulations whose solution time scales \emph{linearly} with the metasurface diameter. Moreover, the solver stores compressed information about the simulation domain that is reusable over many design iterations. We demonstrate the capabilities of our solver through two designs: first, a high-efficiency, high-numerical-aperture metalens that is 20,000 wavelengths in diameter. Second, a high-efficiency, large-beam-width grating coupler. The latter corresponds to millimeter-scale beam design at standard telecommunications wavelengths, while the former, at a visible wavelength of 500 nm, corresponds to a design diameter of 1 cm, created through full simulations of Maxwell's equations.

physics.comp-ph↗

arXiv2023

Small area estimation for composite indicators: the case of multidimensional poverty incidence

This paper proposes a methodology to obtain estimates in small domains when the target is a composite indicator. These indicators are of utmost importance for studying multidimensional phenomena, but little research has been done on how to obtain estimates of these indicators under the small area context. Composite indicators are particularly complex for this purpose since their construction requires different data sources, aggregation procedures, and weighting which makes challenging not only the estimation for small domains but also obtaining uncertainty measures. As case study of our proposal, we estimate the incidence of multidimensional poverty at the municipality level in Colombia. Furthermore, we provide uncertainty measures based on a parametric bootstrap algorithm.

stat.ME↗

arXiv2023

Almost Abelian Lie groups, subgroups and quotients

An almost Abelian Lie group is a non-Abelian Lie group with a codimension 1 Abelian normal subgroup. The majority of 3-dimensional real Lie groups are almost Abelian, and they appear in all parts of physics that deal with anisotropic media - cosmology, crystallography etc. In theoretical physics and differential geometry, almost Abelian Lie groups and their homogeneous spaces provide some of the simplest solvmanifolds on which a variety of geometric structures such as symplectic, Kähler, spin etc., are currently studied in explicit terms. Recently, almost Abelian Lie algebras were classified and studied in details. However, a systematic investigation of almost Abelian Lie groups has not been carried out yet, and the present paper is devoted to an explicit description of properties of this wide and diverse class of groups. The subject of investigation are real almost Abelian Lie groups with their Lie group theoretical aspects, such as the exponential map, faithful matrix representations, discrete and connected subgroups, quotients and automorphisms. The emphasis is put on explicit description of all technical details.

math.GR↗

arXiv2022

3D Cross-Pseudo Supervision (3D-CPS): A semi-supervised nnU-Net architecture for abdominal organ segmentation

Large curated datasets are necessary, but annotating medical images is a time-consuming, laborious, and expensive process. Therefore, recent supervised methods are focusing on utilizing a large amount of unlabeled data. However, to do so, is a challenging task. To address this problem, we propose a new 3D Cross-Pseudo Supervision (3D-CPS) method, a semi-supervised network architecture based on nnU-Net with the Cross-Pseudo Supervision method. We design a new nnU-Net based preprocessing. In addition, we set the semi-supervised loss weights to expand linearity with each epoch to prevent the model from low-quality pseudo-labels in the early training process. Our proposed method achieves an average dice similarity coefficient (DSC) of 0.881 and an average normalized surface distance (NSD) of 0.913 on the MICCAI FLARE2022 validation set (20 cases).

eess.IV↗

arXiv2022

Deterministic 3-Server on a Circle and the Limitation of Canonical Potentials

The deterministic $k$-server conjecture states that there is a $k$-competitive deterministic algorithm for the $k$-server problem for any metric space. We show that the work function algorithm is $3$-competitive for the $3$-server problem on circle metrics, a case left open by Coester and Koutsoupias (2021). Our analysis follows the existing framework but introduces a new potential function which may be viewed as a relaxation of the counterpart by Coester and Koutsoupias (2021). We further notice that the new potential function and many existing ones can be rewritten in a canonical form. Through a computer-aided verification, however, we find that no such canonical potential function can resolve the deterministic $3$-server conjecture for general metric spaces under the current analysis framework.

cs.DS↗

arXiv2022

Basal hydrofractures near sticky patches

Basal crevasses are macroscopic structural discontinuities at the base of ice sheets and glaciers. Motivated by observations and the mechanics of elastic fracture, we hypothesise that in the presence of basal water pressure, spatial variations in basal stress can promote and localise basal crevassing. We quantify this process in the theoretical context of linear elastic fracture mechanics. We develop a model evaluating the effect of shear stress variation on the growth of basal crevasses. Our results indicate that sticky patches promote the initiation of basal crevasses, increase their length of propagation into the ice and, under some conditions, give them curved trajectories that incline upstream. A detailed exploration of the parameter space is conducted to gain a better understanding of the conditions under which sticky-patch-induced basal crevassing likely occurs beneath ice sheets and glaciers.

physics.geo-ph↗

arXiv2021

Conservation-law-based global bounds to quantum optimal control

Active control of quantum systems enables diverse applications ranging from quantum computation to manipulation of molecular processes. Maximum speeds and related bounds have been identified from uncertainty principles and related inequalities, but such bounds utilize only coarse system information, and loosen significantly in the presence of constraints and complex interaction dynamics. We show that an integral-equation-based formulation of conservation laws in quantum dynamics leads to a systematic framework for identifying fundamental limits to any quantum control scenario. We demonstrate the utility of our bounds in three scenarios -- three-level driving, decoherence suppression, and maximum-fidelity gate implementations -- and show that in each case our bounds are tight or nearly so. Global bounds complement local-optimization-based designs, illuminating performance levels that may be possible as well as those that cannot be surpassed.

quant-ph↗

arXiv2021

Magmatic channelisation by reactive and shear-driven instabilities at mid-ocean ridges: a combined analysis

It is generally accepted that melt extraction from the mantle at mid-ocean ridges (MORs) is concentrated in narrow regions of elevated melt fraction called channels. Two feedback mechanisms have been proposed to explain why these channels grow by linear instability: shear flow of partially molten mantle and reactive flow of the ascending magma. These two mechanisms have been studied extensively, in isolation from each other, through theory and laboratory experiments as well as field and geophysical observations. Here, we develop a consistent theory that accounts for both proposed mechanisms and allows us to weigh their relative contributions. We show that interaction of the two feedback mechanisms is insignificant and that the total linear growth rate of channels is well-approximated by summing their independent growth rates. Furthermore, we explain how their competition is governed by the orientation of channels with respect to gravity and mantle shear. By itself, analysis of the reaction-infiltration instability predicts the formation of tube-shaped channels. We show that with the addition of even a small amount of extension in the horizontal, the combined instability favours tabular channels, consistent with the observed morphology of dunite bodies in ophiolites. We apply the new theory to MORs by calculating the accumulated growth and rotation of channels along streamlines of the solid flow. We show that reactive flow is the dominant mechanism deep beneath the ridge axis, where the most unstable orientation of high-porosity channels is sub-vertical. Channels are then rotated by the solid flow away from the vertical. The contribution of the shear-driven instability is confined to the margins of the melting region. Within the limitations of our study, the shear-driven feedback is not responsible for significant melt focusing or for shallowly dipping seismic anisotropy [abridged].

physics.geo-ph↗

arXiv2021

Embedding API Dependency Graph for Neural Code Generation

The problem of code generation from textual program descriptions has long been viewed as a grand challenge in software engineering. In recent years, many deep learning based approaches have been proposed, which can generate a sequence of code from a sequence of textual program description. However, the existing approaches ignore the global relationships among API methods, which are important for understanding the usage of APIs. In this paper, we propose to model the dependencies among API methods as an API dependency graph (ADG) and incorporate the graph embedding into a sequence-to-sequence (Seq2Seq) model. In addition to the existing encoder-decoder structure, a new module named ``embedder" is introduced. In this way, the decoder can utilize both global structural dependencies and textual program description to predict the target code. We conduct extensive code generation experiments on three public datasets and in two programming languages (Python and Java). Our proposed approach, called ADG-Seq2Seq, yields significant improvements over existing state-of-the-art methods and maintains its performance as the length of the target code increases. Extensive ablation tests show that the proposed ADG embedding is effective and outperforms the baselines.

cs.SE↗

arXiv2020

Learning Control Barrier Functions from Expert Demonstrations

Inspired by the success of imitation and inverse reinforcement learning in replicating expert behavior through optimal control, we propose a learning based approach to safe controller synthesis based on control barrier functions (CBFs). We consider the setting of a known nonlinear control affine dynamical system and assume that we have access to safe trajectories generated by an expert - a practical example of such a setting would be a kinematic model of a self-driving vehicle with safe trajectories (e.g., trajectories that avoid collisions with obstacles in the environment) generated by a human driver. We then propose and analyze an optimization-based approach to learning a CBF that enjoys provable safety guarantees under suitable Lipschitz smoothness assumptions on the underlying dynamical system. A strength of our approach is that it is agnostic to the parameterization used to represent the CBF, assuming only that the Lipschitz constant of such functions can be efficiently bounded. Furthermore, if the CBF parameterization is convex, then under mild assumptions, so is our learning process. We end with extensive numerical evaluations of our results on both planar and realistic examples, using both random feature and deep neural network parameterizations of the CBF. To the best of our knowledge, these are the first results that learn provably safe control barrier functions from data.

eess.SY↗

arXiv2020

Learning Hybrid Control Barrier Functions from Data

Motivated by the lack of systematic tools to obtain safe control laws for hybrid systems, we propose an optimization-based framework for learning certifiably safe control laws from data. In particular, we assume a setting in which the system dynamics are known and in which data exhibiting safe system behavior is available. We propose hybrid control barrier functions for hybrid systems as a means to synthesize safe control inputs. Based on this notion, we present an optimization-based framework to learn such hybrid control barrier functions from data. Importantly, we identify sufficient conditions on the data such that feasibility of the optimization problem ensures correctness of the learned hybrid control barrier functions, and hence the safety of the system. We illustrate our findings in two simulations studies, including a compass gait walker.

eess.SY↗

arXiv2020

Quasinormal Coupled Mode Theory

Coupled mode theory (CMT) is a powerful framework for decomposing interactions between electromagnetic waves and scattering bodies into resonances and their couplings with power-carrying channels. It has widespread use in few-resonance, weakly coupled resonator systems across nanophotonics, but cannot be applied to the complex scatterers of emerging importance. We use quasinormal modes to develop an exact, ab initio generalized coupled mode theory from Maxwell's equations. This quasinormal coupled mode theory, which we denote "QCMT", enables a direct, mode-based construction of scattering matrices without resorting to external solvers or data. We consider canonical scattering bodies, for which we show that a CMT model will necessarily be highly inaccurate, whereas QCMT exhibits near-perfect accuracy.

physics.optics↗