arXiv subjects

Zeyu Wang

Publications and source records attributed to Zeyu Wang.

At least 217 records · Page 12Linked to original sources

arXiv2022

Bag of Tricks for FGSM Adversarial Training

Adversarial training (AT) with samples generated by Fast Gradient Sign Method (FGSM), also known as FGSM-AT, is a computationally simple method to train robust networks. However, during its training procedure, an unstable mode of "catastrophic overfitting" has been identified in arXiv:2001.03994 [cs.LG], where the robust accuracy abruptly drops to zero within a single training step. Existing methods use gradient regularizers or random initialization tricks to attenuate this issue, whereas they either take high computational cost or lead to lower robust accuracy. In this work, we provide the first study, which thoroughly examines a collection of tricks from three perspectives: Data Initialization, Network Structure, and Optimization, to overcome the catastrophic overfitting in FGSM-AT. Surprisingly, we find that simple tricks, i.e., a) masking partial pixels (even without randomness), b) setting a large convolution stride and smooth activation functions, or c) regularizing the weights of the first convolutional layer, can effectively tackle the overfitting issue. Extensive results on a range of network architectures validate the effectiveness of each proposed trick, and the combinations of tricks are also investigated. For example, trained with PreActResNet-18 on CIFAR-10, our method attains 49.8% accuracy against PGD-50 attacker and 46.4% accuracy against AutoAttack, demonstrating that pure FGSM-AT is capable of enabling robust learners. The code and models are publicly available at https://github.com/UCSC-VLAA/Bag-of-Tricks-for-FGSM-AT.

cs.CV↗

arXiv2022

Multi-Query Video Retrieval

Retrieving target videos based on text descriptions is a task of great practical value and has received increasing attention over the past few years. Despite recent progress, imperfect annotations in existing video retrieval datasets have posed significant challenges on model evaluation and development. In this paper, we tackle this issue by focusing on the less-studied setting of multi-query video retrieval, where multiple descriptions are provided to the model for searching over the video archive. We first show that multi-query retrieval task effectively mitigates the dataset noise introduced by imperfect annotations and better correlates with human judgement on evaluating retrieval abilities of current models. We then investigate several methods which leverage multiple queries at training time, and demonstrate that the multi-query inspired training can lead to superior performance and better generalization. We hope further investigation in this direction can bring new insights on building systems that perform better in real-world video retrieval applications.

cs.CV↗

arXiv2022

Learning Optimal Treatment Strategies for Sepsis Using Offline Reinforcement Learning in Continuous Space

Sepsis is a leading cause of death in the ICU. It is a disease requiring complex interventions in a short period of time, but its optimal treatment strategy remains uncertain. Evidence suggests that the practices of currently used treatment strategies are problematic and may cause harm to patients. To address this decision problem, we propose a new medical decision model based on historical data to help clinicians recommend the best reference option for real-time treatment. Our model combines offline reinforcement learning and deep reinforcement learning to solve the problem of traditional reinforcement learning in the medical field due to the inability to interact with the environment, while enabling our model to make decisions in a continuous state-action space. We demonstrate that, on average, the treatments recommended by the model are more valuable and reliable than those recommended by clinicians. In a large validation dataset, we find out that the patients whose actual doses from clinicians matched the decisions made by AI has the lowest mortality rates. Our model provides personalized and clinically interpretable treatment decisions for sepsis to improve patient care.

cs.LG↗

arXiv2022

InsMix: Towards Realistic Generative Data Augmentation for Nuclei Instance Segmentation

Nuclei Segmentation from histology images is a fundamental task in digital pathology analysis. However, deep-learning-based nuclei segmentation methods often suffer from limited annotations. This paper proposes a realistic data augmentation method for nuclei segmentation, named InsMix, that follows a Copy-Paste-Smooth principle and performs morphology-constrained generative instance augmentation. Specifically, we propose morphology constraints that enable the augmented images to acquire luxuriant information about nuclei while maintaining their morphology characteristics (e.g., geometry and location). To fully exploit the pixel redundancy of the background and improve the model's robustness, we further propose a background perturbation method, which randomly shuffles the background patches without disordering the original nuclei distribution. To achieve contextual consistency between original and template instances, a smooth-GAN is designed with a foreground similarity encoder (FSE) and a triplet loss. We validated the proposed method on two datasets, i.e., Kumar and CPS datasets. Experimental results demonstrate the effectiveness of each component and the superior performance achieved by our method to the state-of-the-art methods.

eess.IV↗

arXiv2022

Fast multi-source nanophotonic simulations using augmented partial factorization

Full-wave simulations are indispensable for nanophotonics and electromagnetics but are severely constrained on large systems, especially multi-channel ones such as disordered media, aperiodic metasurfaces, and densely packed photonic circuits where each input requires a large-scale simulation. Here we bypass the computationally demanding solution of Maxwell's equations and directly evaluate the full-wave multi-input response, with no approximation. We augment the Maxwell operator with all input source profiles and output projection profiles, followed by a single partial factorization that directly yields the entire multi-input scattering matrix via the Schur complement. This method is simple to implement and applies to any linear partial differential equation. Its advantage grows with size, being 1,000 to 30,000,000 times faster than existing methods for systems with about ten million variables. We use it to realize the first full-wave simulations of entangled-photon backscattering from disorder and all-angle characterizations of high-numerical-aperture metalenses that are thousands of wavelengths wide. This work reveals the significant efficiency gain when we rethink what to compute and enables the exploration of diverse multi-channel systems.

physics.optics↗

arXiv2022

Null Model-Based Data Augmentation for Graph Classification

In network science, the null model is typically used to generate a series of graphs based on randomization as a term of comparison to verify whether a network in question displays some non-trivial features such as community structure. Since such non-trivial features play a significant role in graph classification, the null model could be useful for network data augmentation to enhance classification performance. In this paper, we propose a novel technique that combines the null model with data augmentation for graph classification. Moreover, we propose four standard null model-based augmentation methods and four approximate null model-based augmentation methods to verify and improve the performance of our graph classification technique. Our experiments demonstrate that the proposed augmentation technique has significantly achieved general improvement on the tested datasets. In addition, we find that the standard null model-based augmentation methods always outperform the approximate ones, depending on the design mechanisms of the null models. Our results indicate that the choice of non-trivial features is significant for increasing the performance of augmentation models for different network structures, which also provides a new perspective of data augmentation for studying various graph classification methods.

cs.SI↗

arXiv2022

Whose Language Counts as High Quality? Measuring Language Ideologies in Text Data Selection

Language models increasingly rely on massive web dumps for diverse text data. However, these sources are rife with undesirable content. As such, resources like Wikipedia, books, and newswire often serve as anchors for automatically selecting web text most suitable for language modeling, a process typically referred to as quality filtering. Using a new dataset of U.S. high school newspaper articles -- written by students from across the country -- we investigate whose language is preferred by the quality filter used for GPT-3. We find that newspapers from larger schools, located in wealthier, educated, and urban ZIP codes are more likely to be classified as high quality. We then demonstrate that the filter's measurement of quality is unaligned with other sensible metrics, such as factuality or literary acclaim. We argue that privileging any corpus as high quality entails a language ideology, and more care is needed to construct training corpora for language models, with better transparency and justification for the inclusion or exclusion of various texts.

cs.CL↗

arXiv2022

A Data-driven Adversarial Examples Recognition Framework via Adversarial Feature Genome

Adversarial examples pose many security threats to convolutional neural networks (CNNs). Most defense algorithms prevent these threats by finding differences between the original images and adversarial examples. However, the found differences do not contain features about the classes, so these defense algorithms can only detect adversarial examples without recovering the correct labels. In this regard, we propose the Adversarial Feature Genome (AFG), a novel type of data that contains both the differences and features about classes. This method is inspired by an observed phenomenon, namely the Adversarial Feature Separability (AFS), where the difference between the feature maps of the original images and adversarial examples becomes larger with deeper layers. On top of that, we further develop an adversarial example recognition framework that detects adversarial examples and can recover the correct labels. In the experiments, the detection and classification of adversarial examples by AFGs has an accuracy of more than 90.01\% in various attack scenarios. To the best of our knowledge, our method is the first method that focuses on both attack detecting and recovering. AFG gives a new data-driven perspective to improve the robustness of CNNs. The source code is available at https://github.com/GeoX-Lab/Adv_Fea_Genome.

cs.CV↗

arXiv2022

Optimal monitoring location for risk tracking of geotechnical systems: theory and application to tunneling excavation risks

The maturity of structural health monitoring technology brings ever-increasing opportunities for geotechnical structures and underground infrastructure systems to track the risk of structural failure, such as settlement-induced building damage, based on the monitored data. Reliability updating techniques can offer solutions to estimate the probability of failing to meet a prescribed objective using various types of information that are inclusive of equality and inequality. However, the update in reliability can be highly sensitive to monitoring location. Therefore, there may exist optimal locations in a system for monitoring that yield the maximum value for reliability updating. This paper proposes a computational framework for optimal monitoring location based on an innovative metric called sensitivity of information (SOI) that quantifies the relative change in unconditional and conditional reliability indexes. A state-of-the-practice case of risks posed by tunneling-induced settlement to buildings is explored in-depth to demonstrate and evaluate the computational efficiency of the proposed framework.

stat.AP↗

arXiv2021

Human Activity Recognition Using 3D Orthogonally-projected EfficientNet on Radar Time-Range-Doppler Signature

In radar activity recognition, 2D signal representations such as spectrogram, cepstrum and cadence velocity diagram are often utilized, while range information is often neglected. In this work, we propose to utilize the 3D time-range-Doppler (TRD) representation, and design a 3D Orthogonally-Projected EfficientNet (3D-OPEN) to effectively capture the discriminant information embedded in the 3D TRD cubes for accurate classification. The proposed model aggregates the discriminant information from three orthogonal planes projected from the 3D feature space. It alleviates the difficulty of 3D CNNs in exploiting sparse semantic abstractions directly from the high-dimensional 3D representation. The proposed method is evaluated on the Millimeter-Wave Radar Walking Dataset. It significantly and consistently outperforms the state-of-the-art methods for radar activity recognition.

cs.CV↗

arXiv2021

Chi-square Loss for Softmax: an Echo of Neural Network Structure

Softmax working with cross-entropy is widely used in classification, which evaluates the similarity between two discrete distribution columns (predictions and true labels). Inspired by chi-square test, we designed a new loss function called chi-square loss, which is also works for Softmax. Chi-square loss has a statistical background. We proved that it is unbiased in optimization, and clarified its using conditions (its formula determines that it must work with label smoothing). In addition, we studied the sample distribution of this loss function by visualization and found that the distribution is related to the neural network structure, which is distinct compared to cross-entropy. In the past, the influence of structure was often ignored when visualizing. Chi-square loss can notice changes in neural network structure because it is very strict, and we explained the reason for this strictness. We also studied the influence of label smoothing and discussed the relationship between label smoothing and training accuracy and stability. Since the chi-square loss is very strict, the performance will degrade when dealing samples of very many classes.

cs.LG↗

arXiv2021

maplet: An extensible R toolbox for modular and reproducible omics pipelines

This paper presents maplet, an open-source R package for the creation of highly customizable, fully reproducible statistical pipelines for omics data analysis, with a special focus on metabolomics-based methods. It builds on the SummarizedExperiment data structure to create a centralized pipeline framework for storing data, analysis steps, results, and visualizations. maplet's key design feature is its modularity, which offers several advantages, such as ensuring code quality through the individual maintenance of functions and promoting collaborative development by removing technical barriers to code contribution. With over 90 functions, the package includes a wide range of functionalities, covering many widely used statistical approaches and data visualization techniques.

q-bio.GN↗

arXiv2021

Plot2API: Recommending Graphic API from Plot via Semantic Parsing Guided Neural Network

Plot-based Graphic API recommendation (Plot2API) is an unstudied but meaningful issue, which has several important applications in the context of software engineering and data visualization, such as the plotting guidance of the beginner, graphic API correlation analysis, and code conversion for plotting. Plot2API is a very challenging task, since each plot is often associated with multiple APIs and the appearances of the graphics drawn by the same API can be extremely varied due to the different settings of the parameters. Additionally, the samples of different APIs also suffer from extremely imbalanced. Considering the lack of technologies in Plot2API, we present a novel deep multi-task learning approach named Semantic Parsing Guided Neural Network (SPGNN) which translates the Plot2API issue as a multi-label image classification and an image semantic parsing tasks for the solution. In SPGNN, the recently advanced Convolutional Neural Network (CNN) named EfficientNet is employed as the backbone network for API recommendation. Meanwhile, a semantic parsing module is complemented to exploit the semantic relevant visual information in feature learning and eliminate the appearance-relevant visual information which may confuse the visual-information-based API recommendation. Moreover, the recent data augmentation technique named random erasing is also applied for alleviating the imbalance of API categories. We collect plots with the graphic APIs used to drawn them from Stack Overflow, and release three new Plot2API datasets corresponding to the graphic APIs of R and Python programming languages for evaluating the effectiveness of Plot2API techniques. Extensive experimental results not only demonstrate the superiority of our method over the recent deep learning baselines but also show the practicability of our method in the recommendation of graphic APIs.

cs.SE↗

arXiv2021

SCANN: Synthesis of Compact and Accurate Neural Networks

Deep neural networks (DNNs) have become the driving force behind recent artificial intelligence (AI) research. An important problem with implementing a neural network is the design of its architecture. Typically, such an architecture is obtained manually by exploring its hyperparameter space and kept fixed during training. This approach is time-consuming and inefficient. Another issue is that modern neural networks often contain millions of parameters, whereas many applications and devices require small inference models. However, efforts to migrate DNNs to such devices typically entail a significant loss of classification accuracy. To address these challenges, we propose a two-step neural network synthesis methodology, called DR+SCANN, that combines two complementary approaches to design compact and accurate DNNs. At the core of our framework is the SCANN methodology that uses three basic architecture-changing operations, namely connection growth, neuron growth, and connection pruning, to synthesize feed-forward architectures with arbitrary structure. SCANN encapsulates three synthesis methodologies that apply a repeated grow-and-prune paradigm to three architectural starting points. DR+SCANN combines the SCANN methodology with dataset dimensionality reduction to alleviate the curse of dimensionality. We demonstrate the efficacy of SCANN and DR+SCANN on various image and non-image datasets. We evaluate SCANN on MNIST and ImageNet benchmarks. In addition, we also evaluate the efficacy of using dimensionality reduction alongside SCANN (DR+SCANN) on nine small to medium-size datasets. We also show that our synthesis methodology yields neural networks that are much better at navigating the accuracy vs. energy efficiency space. This would enable neural network-based inference even on Internet-of-Things sensors.

cs.NE↗

arXiv2020

Reflectionless excitation of arbitrary photonic structures: A general theory

We outline a recently developed theory of impedance-matching, or reflectionless excitation of arbitrary finite photonic structures in any dimension. It describes the necessary and sufficient conditions for perfectly reflectionless excitation to be possible, and specifies how many physical parameters must be tuned to achieve this. In the absence of geometric symmetries the tuning of at least one structural parameter will be necessary to achieve reflectionless excitation. The theory employs a recently identified set of complex-frequency solutions of the Maxwell equations as a starting point, which are defined by having zero reflection into a chosen set of input channels, and which are referred to as R-zeros. Tuning is generically necessary in order to move an R-zero to the real-frequency axis, where it becomes a physical steady-state solution, referred to as a Reflectionless Scattering Mode (RSM). Except in single-channel systems, the RSM corresponds to a particular input wavefront, and any other wavefront will generally not be reflectionless. In a structure with parity and time-reversal symmmetry or with parity-time symmetry, generically a subset of R-zeros is real, and reflectionless states exist without structural tuning. Such systems can exhibit symmetry-breaking transitions when two RSMs meet, which corresponds to a recently identified kind of exceptional point at which the shape of the reflection and transmission resonance lineshape is flattened.

physics.optics↗

arXiv2020

Towards Unique and Informative Captioning of Images

Despite considerable progress, state of the art image captioning models produce generic captions, leaving out important image details. Furthermore, these systems may even misrepresent the image in order to produce a simpler caption consisting of common concepts. In this paper, we first analyze both modern captioning systems and evaluation metrics through empirical experiments to quantify these phenomena. We find that modern captioning systems return higher likelihoods for incorrect distractor sentences compared to ground truth captions, and that evaluation metrics like SPICE can be 'topped' using simple captioning systems relying on object detectors. Inspired by these observations, we design a new metric (SPICE-U) by introducing a notion of uniqueness over the concepts generated in a caption. We show that SPICE-U is better correlated with human judgements compared to SPICE, and effectively captures notions of diversity and descriptiveness. Finally, we also demonstrate a general technique to improve any existing captioning model -- by using mutual information as a re-ranking objective during decoding. Empirically, this results in more unique and informative captions, and improves three different state-of-the-art models on SPICE-U as well as average score over existing metrics.

cs.CV↗

arXiv2020

Self-supervised Learning of Detailed 3D Face Reconstruction

In this paper, we present an end-to-end learning framework for detailed 3D face reconstruction from a single image. Our approach uses a 3DMM-based coarse model and a displacement map in UV-space to represent a 3D face. Unlike previous work addressing the problem, our learning framework does not require supervision of surrogate ground-truth 3D models computed with traditional approaches. Instead, we utilize the input image itself as supervision during learning. In the first stage, we combine a photometric loss and a facial perceptual loss between the input face and the rendered face, to regress a 3DMM-based coarse model. In the second stage, both the input image and the regressed texture of the coarse model are unwrapped into UV-space, and then sent through an image-toimage translation network to predict a displacement map in UVspace. The displacement map and the coarse model are used to render a final detailed face, which again can be compared with the original input image to serve as a photometric loss for the second stage. The advantage of learning displacement map in UV-space is that face alignment can be explicitly done during the unwrapping, thus facial details are easier to learn from large amount of data. Extensive experiments demonstrate the superiority of the proposed method over previous work.

cs.CV↗

arXiv2020

Multi-band Weighted $l_p$ Norm Minimization for Image Denoising

Low rank matrix approximation (LRMA) has drawn increasing attention in recent years, due to its wide range of applications in computer vision and machine learning. However, LRMA, achieved by nuclear norm minimization (NNM), tends to over-shrink the rank components with the same threshold and ignore the differences between rank components. To address this problem, we propose a flexible and precise model named multi-band weighted $l_p$ norm minimization (MBWPNM). The proposed MBWPNM not only gives more accurate approximation with a Schatten $p$-norm, but also considers the prior knowledge where different rank components have different importance. We analyze the solution of MBWPNM and prove that MBWPNM is equivalent to a non-convex $l_p$ norm subproblems under certain weight condition, whose global optimum can be solved by a generalized soft-thresholding algorithm. We then adopt the MBWPNM algorithm to color and multispectral image denoising. Extensive experiments on additive white Gaussian noise removal and realistic noise removal demonstrate that the proposed MBWPNM achieves a better performance than several state-of-art algorithms.

cs.CV↗