arXiv subjects

Yu Zhang

Publications and source records attributed to Yu Zhang.

At least 1,171 records · Page 65Linked to original sources

arXiv2022

AnoDFDNet: A Deep Feature Difference Network for Anomaly Detection

This paper proposed a novel anomaly detection (AD) approach of High-speed Train images based on convolutional neural networks and the Vision Transformer. Different from previous AD works, in which anomalies are identified with a single image using classification, segmentation, or object detection methods, the proposed method detects abnormal difference between two images taken at different times of the same region. In other words, we cast anomaly detection problem with a single image into a difference detection problem with two images. The core idea of the proposed method is that the 'anomaly' usually represents an abnormal state instead of a specific object, and this state should be identified by a pair of images. In addition, we introduced a deep feature difference AD network (AnoDFDNet) which sufficiently explored the potential of the Vision Transformer and convolutional neural networks. To verify the effectiveness of the proposed AnoDFDNet, we collected three datasets, a difference dataset (Diff Dataset), a foreign body dataset (FB Dataset), and an oil leakage dataset (OL Dataset). Experimental results on above datasets demonstrate the superiority of proposed method. Source code are available at https://github.com/wangle53/AnoDFDNet.

cs.CV↗

arXiv2022

LibMTL: A Python Library for Multi-Task Learning

This paper presents LibMTL, an open-source Python library built on PyTorch, which provides a unified, comprehensive, reproducible, and extensible implementation framework for Multi-Task Learning (MTL). LibMTL considers different settings and approaches in MTL, and it supports a large number of state-of-the-art MTL methods, including 12 loss weighting strategies, 7 architectures, and 84 combinations of different architectures and loss weighting methods. Moreover, the modular design in LibMTL makes it easy-to-use and well extensible, thus users can easily and fast develop new MTL methods, compare with existing MTL methods fairly, or apply MTL algorithms to real-world applications with the support of LibMTL. The source code and detailed documentations of LibMTL are available at https://github.com/median-research-group/LibMTL and https://libmtl.readthedocs.io, respectively.

cs.LG↗

arXiv2022

OneLabeler: A Flexible System for Building Data Labeling Tools

Labeled datasets are essential for supervised machine learning. Various data labeling tools have been built to collect labels in different usage scenarios. However, developing labeling tools is time-consuming, costly, and expertise-demanding on software development. In this paper, we propose a conceptual framework for data labeling and OneLabeler based on the conceptual framework to support easy building of labeling tools for diverse usage scenarios. The framework consists of common modules and states in labeling tools summarized through coding of existing tools. OneLabeler supports configuration and composition of common software modules through visual programming to build data labeling tools. A module can be a human, machine, or mixed computation procedure in data labeling. We demonstrate the expressiveness and utility of the system through ten example labeling tools built with OneLabeler. A user study with developers provides evidence that OneLabeler supports efficient building of diverse data labeling tools.

cs.HC↗

arXiv2022

Metadata-Induced Contrastive Learning for Zero-Shot Multi-Label Text Classification

Large-scale multi-label text classification (LMTC) aims to associate a document with its relevant labels from a large candidate set. Most existing LMTC approaches rely on massive human-annotated training data, which are often costly to obtain and suffer from a long-tailed label distribution (i.e., many labels occur only a few times in the training set). In this paper, we study LMTC under the zero-shot setting, which does not require any annotated documents with labels and only relies on label surface names and descriptions. To train a classifier that calculates the similarity score between a document and a label, we propose a novel metadata-induced contrastive learning (MICoL) method. Different from previous text-based contrastive learning techniques, MICoL exploits document metadata (e.g., authors, venues, and references of research papers), which are widely available on the Web, to derive similar document-document pairs. Experimental results on two large-scale datasets show that: (1) MICoL significantly outperforms strong zero-shot text classification and contrastive learning baselines; (2) MICoL is on par with the state-of-the-art supervised metadata-aware LMTC method trained on 10K-200K labeled documents; and (3) MICoL tends to predict more infrequent labels than supervised methods, thus alleviates the deteriorated performance on long-tailed labels.

cs.CL↗

arXiv2022

Upper Field Strength Limit of Fast Radio Bursts

Fast radio bursts (FRBs) are cosmological radio transients with unclear generation mechanism. Known characteristics such as their luminosity, duration, spectrum and repetition rate, etc. suggest that FRBs are powerful coherent radio signals at GHz frequencies, but the status of FRBs near source remain unknown. As an extreme astronomical event, FRBs should be accompanied by energy -- comparable or even more powerful x/γ-ray counterparts. Here, particle-in-cell simulations of ultra-strong GHz radio pulse interaction with GeV photons show that at 3*10^12V/cm field-strengths, quantum cascade can generate dense pair plasmas, which greatly dampen the radio pulse. Thus, in the presence of GeV photons in the source region, GHz radio pulses stronger than 3*10^12V/cm cannot escape. This result indicates an upper field-strength limit of FRB at the source.

astro-ph.HE↗

arXiv2022

Topological properties of two-dimensional photonic square lattice without $C_4$ and $M_{x(y)}$ symmetries

Rich topological phenomena, edge states and two types of corner states, are unveiled in a two-dimensional square-lattice dielectric photonic crystal without both $C_4$ and $M_{x(y)}$ symmetries. Specifically, non-trivial type-I corner states, which do not exist in systems with $C_4$ and $M_{x(y)}$ since the degeneracy, are protected by non-zero quadrupole moment, no longer quantized to but less than $0.5$. Excellent properties, e.g. sub-wavelength localization and air-concentrated field distribution, are presented. Type-II corner states, induced by long-range interactions, are easier realized due to asymmetry. This work broadens the topological physics for the symmetries-broken systems and provides potential applications.

physics.optics↗

arXiv2022

Atomic-Scale Visualization of Chiral Charge Density Wave States and Their Reversible Transition

Chirality is essential for various amazing phenomena in life and matter. However,chirality and its switching in electronic superlattices, such as charge density wave(CDW) arrays, remain elusive. In this study, we characterize the chirality transition with atom-resolution imaging in a single-layer NbSe2 CDW pattern by technique of scanning tunneling microscopy. The atomic lattice of the CDW array is found continuous and intact although its chirality is switched. Several intermediate states are tracked by time-resolved imaging, revealing the fast and dynamic chirality transition. Importantly, the switching is reversibly realized with an external electric-field. Our findings unveil the delicate transition process of chiral CDW array in a 2D crystal down to the atomic scale and may be applicable for future nanoscale devices.

cond-mat.mtrl-sci↗

arXiv2022

Estimating Cluster Masses from SDSS Multi-band Images with Transfer Learning

The total masses of galaxy clusters characterize many aspects of astrophysics and the underlying cosmology. It is crucial to obtain reliable and accurate mass estimates for numerous galaxy clusters over a wide range of redshifts and mass scales. We present a transfer-learning approach to estimate cluster masses using the ugriz-band images in the SDSS Data Release 12. The target masses are derived from X-ray or SZ measurements that are only available for a small subset of the clusters. We designed a semi-supervised deep learning model consisting of two convolutional neural networks. In the first network, a feature extractor is trained to classify the SDSS photometric bands. The second network takes the previously trained features as inputs to estimate their total masses. The training and testing processes in this work depend purely on real observational data. Our algorithm reaches a mean absolute error (MAE) of 0.232 dex on average and 0.214 dex for the best fold. The performance is comparable to that given by redMaPPer, 0.192 dex. We have further applied a joint integrated gradient and class activation mapping method to interpret such a two-step neural network. The performance of our algorithm is likely to improve as the size of training dataset increases. This proof-of-concept experiment demonstrates the potential of deep learning in maximizing the scientific return of the current and future large cluster surveys.

astro-ph.IM↗

arXiv2022

Image Steganography based on Style Transfer

Image steganography is the art and science of using images as cover for covert communications. With the development of neural networks, traditional image steganography is more likely to be detected by deep learning-based steganalysis. To improve upon this, we propose image steganography network based on style transfer, and the embedding of secret messages can be disguised as image stylization. We embed secret information while transforming the content image style. In latent space, the secret information is integrated into the latent representation of the cover image to generate the stego images, which are indistinguishable from normal stylized images. It is an end-to-end unsupervised model without pre-training. Extensive experiments on the benchmark dataset demonstrate the reliability, quality and security of stego images generated by our steganographic network.

cs.CV↗

arXiv2022

Improving Confidence Estimation on Out-of-Domain Data for End-to-End Speech Recognition

As end-to-end automatic speech recognition (ASR) models reach promising performance, various downstream tasks rely on good confidence estimators for these systems. Recent research has shown that model-based confidence estimators have a significant advantage over using the output softmax probabilities. If the input data to the speech recogniser is from mismatched acoustic and linguistic conditions, the ASR performance and the corresponding confidence estimators may exhibit severe degradation. Since confidence models are often trained on the same in-domain data as the ASR, generalising to out-of-domain (OOD) scenarios is challenging. By keeping the ASR model untouched, this paper proposes two approaches to improve the model-based confidence estimators on OOD data: using pseudo transcriptions and an additional OOD language model. With an ASR model trained on LibriSpeech, experiments show that the proposed methods can greatly improve the confidence metrics on TED-LIUM and Switchboard datasets while preserving in-domain performance. Furthermore, the improved confidence estimators are better calibrated on OOD data and can provide a much more reliable criterion for data selection.

eess.AS↗

arXiv2022

Ask2Mask: Guided Data Selection for Masked Speech Modeling

Masked speech modeling (MSM) methods such as wav2vec2 or w2v-BERT learn representations over speech frames which are randomly masked within an utterance. While these methods improve performance of Automatic Speech Recognition (ASR) systems, they have one major limitation. They treat all unsupervised speech samples with equal weight, which hinders learning as not all samples have relevant information to learn meaningful representations. In this work, we address this limitation. We propose ask2mask (ATM), a novel approach to focus on specific samples during MSM pre-training. ATM employs an external ASR model or \textit{scorer} to weight unsupervised input samples in two different ways: 1) A fine-grained data selection is performed by masking over the highly confident input frames as chosen by the scorer. This allows the model to learn meaningful representations. 2) ATM is further extended to focus at utterance-level by weighting the final MSM loss with the utterance-level confidence score. We conduct fine-tuning experiments on two well-benchmarked corpora: LibriSpeech (matching the pre-training data) and Commonvoice, TED-LIUM, AMI and CHiME-6 (not matching the pre-training data). The results substantiate the efficacy of ATM on significantly improving the recognition performance under mismatched conditions (up to 11.6\% relative over published results and upto 4.46\% relative over our internal baseline) while still yielding modest improvements under matched conditions.

cs.SD↗

arXiv2022

Path-Aware Graph Attention for HD Maps in Motion Prediction

The success of motion prediction for autonomous driving relies on integration of information from the HD maps. As maps are naturally graph-structured, investigation on graph neural networks (GNNs) for encoding HD maps is burgeoning in recent years. However, unlike many other applications where GNNs have been straightforwardly deployed, HD maps are heterogeneous graphs where vertices (lanes) are connected by edges (lane-lane interaction relationships) of various nature, and most graph-based models are not designed to understand the variety of edge types which provide crucial cues for predicting how the agents would travel the lanes. To overcome this challenge, we propose Path-Aware Graph Attention, a novel attention architecture that infers the attention between two vertices by parsing the sequence of edges forming the paths that connect them. Our analysis illustrates how the proposed attention mechanism can facilitate learning in a didactic problem where existing graph networks like GCN struggle. By improving map encoding, the proposed model surpasses previous state of the art on the Argoverse Motion Forecasting dataset, and won the first place in the 2021 Argoverse Motion Forecasting Competition.

cs.AI↗

arXiv2022

A van der Waals Interface Hosting Two Groups of Magnetic Skyrmions

Multiple magnetic skyrmion phases add an additional degree of freedom for skyrmion based ultrahigh-density spin memory devices. Extending the field to two-dimensional van der Waals magnets is a rewarding challenge, where the realizable degree of freedoms (e.g. thickness, twisting angle and electrical gating) and high skyrmion density result in intriguing new properties and enhanced functionality. We report a van der Waals interface, formed by two 2D ferromagnets Cr2Ge2Te6 and Fe3GeTe2 with a Curie temperature of ~65 K and ~205 K, respectively, hosting two groups of magnetic skyrmions. Two sets of topological Hall effect are observed below 60 K when Cr2Ge2Te6 is magnetically ordered. These two groups of skyrmions are directly imaged using magnetic force microscopy. Interestingly, the magnetic skyrmions persist in the heterostructure in the remanent state with zero applied magnetic field. Our results are promising for the realization of skyrmionic devices based on van der Waals heterostructures hosting multiple skyrmion phases.

cond-mat.mes-hall↗

arXiv2022

CUP: A Conservative Update Policy Algorithm for Safe Reinforcement Learning

Safe reinforcement learning (RL) is still very challenging since it requires the agent to consider both return maximization and safe exploration. In this paper, we propose CUP, a Conservative Update Policy algorithm with a theoretical safety guarantee. We derive the CUP based on the new proposed performance bounds and surrogate functions. Although using bounds as surrogate functions to design safe RL algorithms have appeared in some existing works, we develop them at least three aspects: (i) We provide a rigorous theoretical analysis to extend the surrogate functions to generalized advantage estimator (GAE). GAE significantly reduces variance empirically while maintaining a tolerable level of bias, which is an efficient step for us to design CUP; (ii) The proposed bounds are tighter than existing works, i.e., using the proposed bounds as surrogate functions are better local approximations to the objective and safety constraints. (iii) The proposed CUP provides a non-convex implementation via first-order optimizers, which does not depend on any convex approximation. Finally, extensive experiments show the effectiveness of CUP where the agent satisfies safe constraints. We have opened the source code of CUP at https://github.com/RL-boxes/Safe-RL.

cs.LG↗

arXiv2022

Policy Optimization with Stochastic Mirror Descent

Improving sample efficiency has been a longstanding goal in reinforcement learning. This paper proposes $\mathtt{VRMPO}$ algorithm: a sample efficient policy gradient method with stochastic mirror descent. In $\mathtt{VRMPO}$, a novel variance-reduced policy gradient estimator is presented to improve sample efficiency. We prove that the proposed $\mathtt{VRMPO}$ needs only $\mathcal{O}(ε^{-3})$ sample trajectories to achieve an $ε$-approximate first-order stationary point, which matches the best sample complexity for policy optimization. The extensive experimental results demonstrate that $\mathtt{VRMPO}$ outperforms the state-of-the-art policy gradient methods in various settings.

cs.LG↗

arXiv2022

Topic Discovery via Latent Space Clustering of Pretrained Language Model Representations

Topic models have been the prominent tools for automatic topic discovery from text corpora. Despite their effectiveness, topic models suffer from several limitations including the inability of modeling word ordering information in documents, the difficulty of incorporating external linguistic knowledge, and the lack of both accurate and efficient inference methods for approximating the intractable posterior. Recently, pretrained language models (PLMs) have brought astonishing performance improvements to a wide variety of tasks due to their superior representations of text. Interestingly, there have not been standard approaches to deploy PLMs for topic discovery as better alternatives to topic models. In this paper, we begin by analyzing the challenges of using PLM representations for topic discovery, and then propose a joint latent space learning and clustering framework built upon PLM embeddings. In the latent space, topic-word and document-topic distributions are jointly modeled so that the discovered topics can be interpreted by coherent and distinctive terms and meanwhile serve as meaningful summaries of the documents. Our model effectively leverages the strong representation power and superb linguistic features brought by PLMs for topic discovery, and is conceptually simpler than topic models. On two benchmark datasets in different domains, our model generates significantly more coherent and diverse topics than strong topic models, and offers better topic-wise document representations, based on both automatic and human evaluations.

cs.CL↗

arXiv2022

Simultaneous Detection of Optical Flares of the Magnetically Active M Dwarf Wolf 359

We present detections of stellar flares of Wolf\,359, an M6.5 dwarf in the solar neighborhood (2.41~pc) known to be prone to flares due to surface magnetic activity. The observations were carried out from 2020 April 23 to 29 with a 1-m and a 0.5-m telescope separated by nearly 300~km in Xinjiang, China. In 27~hr of photometric monitoring, a total of 13 optical flares were detected, each with a total energy of $\gtrsim 5 \times 10^{29}$~erg. The measured event rate of about once every two hours is consistent with those reported previously in radio, X-ray and optical wavelengths for this star. One such flare, detected by both telescopes on 26 April, was an energetic event with a released energy of nearly $10^{33}$~erg. The two-telescope lightcurves of this major event sampled at different cadences and exposure timings enabled us to better estimate the intrinsic flare profile, which reached a peak of up to 1.6 times the stellar quiescent brightness, that otherwise would have been underestimated in the observed flare amplitudes of about $0.4$ and $0.8$, respectively, with single telescopes alone. The compromise between fast sampling so as to resolve a flare profile versus a longer integration time for higher photometric signal-to-noise provides a useful guidance in the experimental design of future flare observations.

astro-ph.SR↗

arXiv2022

Predicting Future CSI Feedback For Highly-Mobile Massive MIMO Systems

Massive multiple-input multiple-output (MIMO) system is promising in providing unprecedentedly high data rate. To achieve its full potential, the transceiver needs complete channel state information (CSI) to perform transmit/receive precoding/combining. This requirement, however, is challenging in the practical systems due to the unavoidable processing and feedback delays, which oftentimes degrades the performance to a great extent, especially in the high mobility scenarios. In this paper, we develop a deep learning based channel prediction framework that proactively predicts the downlink channel state information based on the past observed channel sequence. In its core, the model adopts a 3-D convolutional neural network (CNN) based architecture to efficiently learn the temporal, spatial and frequency correlations of downlink channel samples, based on which accurate channel prediction can be performed. Simulation results highlight the potential of the developed learning model in extracting information and predicting future downlink channels directly from the observed past channel sequence, which significantly improves the performance compared to the sample-and-hold approach, and mitigates the impact of the dynamic communication environment.

cs.IT↗