arXiv subjects

Wenhao Hu

Publications and source records attributed to Wenhao Hu.

32 records · Page 2Linked to original sources

arXiv2024

DQA: An Efficient Method for Deep Quantization of Deep Neural Network Activations

Quantization of Deep Neural Network (DNN) activations is a commonly used technique to reduce compute and memory demands during DNN inference, which can be particularly beneficial on resource-constrained devices. To achieve high accuracy, existing methods for quantizing activations rely on complex mathematical computations or perform extensive searches for the best hyper-parameters. However, these expensive operations are impractical on devices with limited computation capabilities, memory capacities, and energy budgets. Furthermore, many existing methods do not focus on sub-6-bit (or deep) quantization. To fill these gaps, in this paper we propose DQA (Deep Quantization of DNN Activations), a new method that focuses on sub-6-bit quantization of activations and leverages simple shifting-based operations and Huffman coding to be efficient and achieve high accuracy. We evaluate DQA with 3, 4, and 5-bit quantization levels and three different DNN models for two different tasks, image classification and image segmentation, on two different datasets. DQA shows significantly better accuracy (up to 29.28%) compared to the direct quantization method and the state-of-the-art NoisyQuant for sub-6-bit quantization.

cs.LG↗

arXiv2024

Blind Inpainting with Object-aware Discrimination for Artificial Marker Removal

Medical images often incorporate doctor-added markers that can hinder AI-based diagnosis. This issue highlights the need of inpainting techniques to restore the corrupted visual contents. However, existing methods require manual mask annotation as input, limiting the application scenarios. In this paper, we propose a novel blind inpainting method that automatically reconstructs visual contents within the corrupted regions without mask input as guidance. Our model includes a blind reconstruction network and an object-aware discriminator for adversarial training. The reconstruction network contains two branches that predict corrupted regions in images and simultaneously restore the missing visual contents. Leveraging the potent recognition capability of a dense object detector, the object-aware discriminator ensures markers undetectable after inpainting. Thus, the restored images closely resemble the clean ones. We evaluate our method on three datasets of various medical imaging modalities, confirming better performance over other state-of-the-art methods.

cs.CV↗

arXiv2024

Designing Efficient LLM Accelerators for Edge Devices

The increase in open-source availability of Large Language Models (LLMs) has enabled users to deploy them on more and more resource-constrained edge devices to reduce reliance on network connections and provide more privacy. However, the high computation and memory demands of LLMs make their execution on resource-constrained edge devices challenging and inefficient. To address this issue, designing new and efficient edge accelerators for LLM inference is crucial. FPGA-based accelerators are ideal for LLM acceleration due to their reconfigurability, as they enable model-specific optimizations and higher performance per watt. However, creating and integrating FPGA-based accelerators for LLMs (particularly on edge devices) has proven challenging, mainly due to the limited hardware design flows for LLMs in existing FPGA platforms. To tackle this issue, in this paper we first propose a new design platform, named SECDA-LLM, that utilizes the SECDA methodology to streamline the process of designing, integrating, and deploying efficient FPGA-based LLM accelerators for the llama.cpp inference framework. We then demonstrate, through a case study, the potential benefits of SECDA-LLM by creating a new MatMul accelerator that supports block floating point quantized operations for LLMs. Our initial accelerator design, deployed on the PYNQ-Z1 board, reduces latency 1.7 seconds per token or ~2 seconds per word) by 11x over the dual-core Arm NEON-based CPU execution for the TinyLlama model.

cs.AR↗

arXiv2024

CityCraft: A Real Crafter for 3D City Generation

City scene generation has gained significant attention in autonomous driving, smart city development, and traffic simulation. It helps enhance infrastructure planning and monitoring solutions. Existing methods have employed a two-stage process involving city layout generation, typically using Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), or Transformers, followed by neural rendering. These techniques often exhibit limited diversity and noticeable artifacts in the rendered city scenes. The rendered scenes lack variety, resembling the training images, resulting in monotonous styles. Additionally, these methods lack planning capabilities, leading to less realistic generated scenes. In this paper, we introduce CityCraft, an innovative framework designed to enhance both the diversity and quality of urban scene generation. Our approach integrates three key stages: initially, a diffusion transformer (DiT) model is deployed to generate diverse and controllable 2D city layouts. Subsequently, a Large Language Model(LLM) is utilized to strategically make land-use plans within these layouts based on user prompts and language guidelines. Based on the generated layout and city plan, we utilize the asset retrieval module and Blender for precise asset placement and scene construction. Furthermore, we contribute two new datasets to the field: 1)CityCraft-OSM dataset including 2D semantic layouts of urban areas, corresponding satellite images, and detailed annotations. 2) CityCraft-Buildings dataset, featuring thousands of diverse, high-quality 3D building assets. CityCraft achieves state-of-the-art performance in generating realistic 3D cities.

cs.CV↗

arXiv2023

A Survey of Deep Learning in Sports Applications: Perception, Comprehension, and Decision

Deep learning has the potential to revolutionize sports performance, with applications ranging from perception and comprehension to decision. This paper presents a comprehensive survey of deep learning in sports performance, focusing on three main aspects: algorithms, datasets and virtual environments, and challenges. Firstly, we discuss the hierarchical structure of deep learning algorithms in sports performance which includes perception, comprehension and decision while comparing their strengths and weaknesses. Secondly, we list widely used existing datasets in sports and highlight their characteristics and limitations. Finally, we summarize current challenges and point out future trends of deep learning in sports. Our survey provides valuable reference material for researchers interested in deep learning in sports applications.

cs.CV↗

arXiv2023

Improving 2D face recognition via fine-level facial depth generation and RGB-D complementary feature learning

Face recognition in complex scenes suffers severe challenges coming from perturbations such as pose deformation, ill illumination, partial occlusion. Some methods utilize depth estimation to obtain depth corresponding to RGB to improve the accuracy of face recognition. However, the depth generated by them suffer from image blur, which introduces noise in subsequent RGB-D face recognition tasks. In addition, existing RGB-D face recognition methods are unable to fully extract complementary features. In this paper, we propose a fine-grained facial depth generation network and an improved multimodal complementary feature learning network. Extensive experiments on the Lock3DFace dataset and the IIIT-D dataset show that the proposed FFDGNet and I MCFLNet can improve the accuracy of RGB-D face recognition while achieving the state-of-the-art performance.

cs.CV↗

arXiv2023

Deep Learning Methods for Small Molecule Drug Discovery: A Survey

With the development of computer-assisted techniques, research communities including biochemistry and deep learning have been devoted into the drug discovery field for over a decade. Various applications of deep learning have drawn great attention in drug discovery, such as molecule generation, molecular property prediction, retrosynthesis prediction, and reaction prediction. While most existing surveys only focus on one of the applications, limiting the view of researchers in the community. In this paper, we present a comprehensive review on the aforementioned four aspects, and discuss the relationships among different applications. The latest literature and classical benchmarks are presented for better understanding the development of variety of approaches. We commence by summarizing the molecule representation format in these works, followed by an introduction of recent proposed approaches for each of the four tasks. Furthermore, we review a variety of commonly used datasets and evaluation metrics and compare the performance of deep learning-based models. Finally, we conclude by identifying remaining challenges and discussing the future trend for deep learning methods in drug discovery.

cs.LG↗

arXiv2021

Feasible Architecture for Quantum Fully Convolutional Networks

Fully convolutional networks are robust in performing semantic segmentation, with many applications from signal processing to computer vision. From the fundamental principles of variational quantum algorithms, we propose a feasible pure quantum architecture that can be operated on noisy intermediate-scale quantum devices. In this work, a parameterized quantum circuit consisting of three layers, convolutional, pooling, and upsampling, is characterized by generative one-qubit and two-qubit gates and driven by a classical optimizer. This architecture supplies a solution for realizing the dynamical programming on a one-way quantum computer and maximally taking advantage of quantum computing throughout the calculation. Moreover, our algorithm works on many physical platforms, and particularly the upsampling layer can use either conventional qubits or multiple-level systems. Through numerical simulations, our study represents the successful training of a pure quantum fully convolutional network and discusses advantages by comparing it with the hybrid solution.

quant-ph↗

arXiv2020

The Trace Criterion for Kernel Bandwidth Selection for Support Vector Data Description

Support vector data description (SVDD) is a popular anomaly detection technique. The SVDD classifier partitions the whole data space into an inlier region, which consists of the region near the training data, and an outlier region, which consists of points away from the training data. The computation of the SVDD classifier requires a kernel function, for which the Gaussian kernel is a common choice. The Gaussian kernel has a bandwidth parameter, and it is important to set the value of this parameter correctly for good results. A small bandwidth leads to overfitting such that the resulting SVDD classifier overestimates the number of anomalies, whereas a large bandwidth leads to underfitting and an inability to detect many anomalies. In this paper, we present a new unsupervised method for selecting the Gaussian kernel bandwidth. Our method exploits a low-rank representation of the kernel matrix to suggest a kernel bandwidth value. Our new technique is competitive with the current state of the art for low-dimensional data and performs extremely well for many classes of high-dimensional data. Because the mathematical formulation of SVDD is identical with the mathematical formulation of one-class support vector machines (OCSVM) when the Gaussian kernel is used, our method is equally applicable to Gaussian kernel bandwidth tuning for OCSVM.

stat.ML↗

arXiv2018

A fast algorithm for computing distance correlation

Classical dependence measures such as Pearson correlation, Spearman's $ρ$, and Kendall's $τ$ can detect only monotonic or linear dependence. To overcome these limitations, Szekely et al.(2007) proposed distance covariance as a weighted $L_2$ distance between the joint characteristic function and the product of marginal distributions. The distance covariance is $0$ if and only if two random vectors ${X}$ and ${Y}$ are independent. This measure has the power to detect the presence of a dependence structure when the sample size is large enough. They further showed that the sample distance covariance can be calculated simply from modified Euclidean distances, which typically requires $\mathcal{O}(n^2)$ cost. The quadratic computing time greatly limits the application of distance covariance to large data. In this paper, we present a simple exact $\mathcal{O}(n\log(n))$ algorithm to calculate the sample distance covariance between two univariate random variables. The proposed method essentially consists of two sorting steps, so it is easy to implement. Empirical results show that the proposed algorithm is significantly faster than state-of-the-art methods. The algorithm's speed will enable researchers to explore complicated dependence structures in large datasets.

stat.CO↗

arXiv2018

Fast Incremental SVDD Learning Algorithm with the Gaussian Kernel

Support vector data description (SVDD) is a machine learning technique that is used for single-class classification and outlier detection. The idea of SVDD is to find a set of support vectors that defines a boundary around data. When dealing with online or large data, existing batch SVDD methods have to be rerun in each iteration. We propose an incremental learning algorithm for SVDD that uses the Gaussian kernel. This algorithm builds on the observation that all support vectors on the boundary have the same distance to the center of sphere in a higher-dimensional feature space as mapped by the Gaussian kernel function. Each iteration involves only the existing support vectors and the new data point. Moreover, the algorithm is based solely on matrix manipulations; the support vectors and their corresponding Lagrange multiplier $α_i$'s are automatically selected and determined in each iteration. It can be seen that the complexity of our algorithm in each iteration is only $O(k^2)$, where $k$ is the number of support vectors. Experimental results on some real data sets indicate that FISVDD demonstrates significant gains in efficiency with almost no loss in either outlier detection accuracy or objective function value.

stat.ML↗

arXiv2018

Variable selection using pseudo-variables

Penalized regression has become a standard tool for model building across a wide range of application domains. Common practice is to tune the amount of penalization to tradeoff bias and variance or to optimize some other measure of performance of the estimated model. An advantage of such automated model-building procedures is that their operating characteristics are well-defined, i.e., completely data-driven, and thereby they can be systematically studied. However, in many applications it is desirable to incorporate domain knowledge into the model building process; one way to do this is to characterize each model along the solution path of a penalized regression estimator in terms of an operating characteristic that is meaningful within a domain context and then to allow domain experts to choose from among these models using these operating characteristics as well as other factors not available to the estimation algorithm. We derive an estimator of the false selection rate for each model along the solution path using a novel variable addition method. The proposed estimator applies to both fixed and random designs and allows for $p \gg n$. The proposed estimator can be used to estimate a model with a pre-specified false selection rate or can be overlaid on the solution path to facilitate interactive model exploration. We characterize the asymptotic behavior of the proposed estimator in the case of a linear model under a fixed design; however, simulation experiments show that the proposed estimator provides consistently more accurate estimates of the false selection rate than competing methods across a wide range of models.

stat.ME↗

arXiv2016

Strain-based Spin Manipulation on Substitutional Nickel in Silicon Carbide

By using the full potential linear augmented plane wave (FP-LAPW) method and full potential local orbital minimum basis (FP-LOMB) method within generalized gradient approximation (GGA), we studied the electronic structures and magnetic properties of nickel and chromium single dopants in polytypes of silicon carbide (SiC). The magnetic phases of defects are found to be strongly dependent on the external stress on the supercell. In 3C-SiC, the Ni single dopant exhibits an anti-ferromagnetic (AFM) to ferromagnetic (FM) transition at a moderate compressive and tensile hydrostatic strain in Si-sub and C-sub cases. In contrast, the Ni single dopant in 4H-SiC is stably in the nonmagnetic phase under external stress. The Cr single dopant is also insensitive to the applied stress but stably in the magnetic phase. This strain controlled magnetic transition makes the Ni single dopant a novel scheme of qubit.

cond-mat.mtrl-sci↗

arXiv2014

Synthesis of Luminescent Eu defects in diamond

Lanthanides are vital components in lighting, imaging technologies and future quantum memory applications due to their narrow optical transitions and long spin coherence times. Recently, diamond has become a preeminent platform for realization of many experiments in quantum information science. In this work, we demonstrate a promising approach to incorporate Eu ions into single crystal diamond and nanodiamonds, providing a means to harness the exceptional characteristics of both lanthanides and diamond in a single material. Polyelectrolytes are used to electrostatically assemble Eu(III) chelate molecules on diamond and subsequently chemical vapor deposition is employed for the growth of a high quality diamond crystal. Photoluminescence, cathodoluminescence and time resolved fluorescence measurements show that the Eu atoms retain the characteristic optical signature of Eu(III) upon incorporation into the diamond lattice. Computational modelling supports the experimental findings, corroborating that Eu3+ in diamond is a stable configuration within the diamond bandgap. The versatility of the synthetic technique is further illustrated through the creation of the well-studied Cr defect center. Together these defect centers demonstrate the outstanding chemical control over the incorporation of impurities into diamond enabled by the electrostatic assembly together with chemical vapour deposition growth.

cond-mat.mtrl-sci↗