arXiv subjects

Wei Han

Publications and source records attributed to Wei Han.

At least 163 records · Page 9Linked to original sources

arXiv2020

Finding the Evidence: Localization-aware Answer Prediction for Text Visual Question Answering

Image text carries essential information to understand the scene and perform reasoning. Text-based visual question answering (text VQA) task focuses on visual questions that require reading text in images. Existing text VQA systems generate an answer by selecting from optical character recognition (OCR) texts or a fixed vocabulary. Positional information of text is underused and there is a lack of evidence for the generated answer. As such, this paper proposes a localization-aware answer prediction network (LaAP-Net) to address this challenge. Our LaAP-Net not only generates the answer to the question but also predicts a bounding box as evidence of the generated answer. Moreover, a context-enriched OCR representation (COR) for multimodal fusion is proposed to facilitate the localization task. Our proposed LaAP-Net outperforms existing approaches on three benchmark datasets for the text VQA task by a noticeable margin.

cs.CV↗

arXiv2020

Rate-Splitting Multiple Access for Downlink Multi-Antenna Communications: Physical Layer Design and Link-level Simulations

Rate-Splitting Multiple Access (RSMA) is an emerging flexible, robust and powerful multiple access scheme for downlink multi-antenna wireless networks. RSMA relies on multi-antenna Rate-Splitting (RS) strategies at the transmitter and Successive Interference Cancellation (SIC) at the receivers, and has the unique ability to partially decode interference and partially treat interference as noise so as to softly bridge the two extremes of fully decoding interference (as in Non-Orthogonal Multiple Access, NOMA) and treating interference as noise (as in Space Division Multiple Access, SDMA or Multi-User Multiple-Input Multiple-Output, MU-MIMO). RSMA has been shown to provide significant room for spectral efficiency, energy efficiency, Quality-of-Service enhancements, robustness to Channel State Information (CSI) imperfections, as well as feedback overhead and complexity reduction, in a wide range of network loads (underloaded and overloaded regimes) and user deployments (with a diversity of channel directions, channel strengths and qualities). RSMA is also deeply rooted and motivated by recent advances in understanding the fundamental limits of multi-antenna networks with imperfect CSI at the Transmitter (CSIT). In this work, we leverage recent results on the optimization of RSMA and design for the first time its physical layer, accounting for modulation, coding (using polar codes), message split, adaptive modulation and coding, and SIC receiver. Link-level evaluations confirm the significant throughput benefits of RSMA over various baselines as SDMA and NOMA.

cs.IT↗

arXiv2020

Rate-Splitting Multiple Access: A New Frontier for the PHY Layer of 6G

In order to efficiently cope with the high throughput, reliability, heterogeneity of Quality-of-Service (QoS), and massive connectivity requirements of future 6G multi-antenna wireless networks, multiple access and multiuser communication system design need to depart from conventional interference management strategies, namely fully treat interference as noise (as commonly used in 4G/5G, MU-MIMO, CoMP, Massive MIMO, millimetre wave MIMO) and fully decode interference (as in Non-Orthogonal Multiple Access, NOMA). This paper is dedicated to the theory and applications of a more general and powerful transmission framework based on Rate-Splitting Multiple Access (RSMA) that splits messages into common and private parts and enables to partially decode interference and treat remaining part of the interference as noise. This enables RSMA to softly bridge and therefore reconcile the two extreme strategies of fully decode interference and treat interference as noise and provide room for spectral efficiency, energy efficiency and QoS enhancements, robustness to imperfect Channel State Information at the Transmitter (CSIT), and complexity reduction. We give an overview of RSMA and its potential to address the requirements of 6G. This paper provides an overview of RSMA and its potential to address the requirements of 6G.

cs.IT↗

arXiv2020

ContextNet: Improving Convolutional Neural Networks for Automatic Speech Recognition with Global Context

Convolutional neural networks (CNN) have shown promising results for end-to-end speech recognition, albeit still behind other state-of-the-art methods in performance. In this paper, we study how to bridge this gap and go beyond with a novel CNN-RNN-transducer architecture, which we call ContextNet. ContextNet features a fully convolutional encoder that incorporates global context information into convolution layers by adding squeeze-and-excitation modules. In addition, we propose a simple scaling method that scales the widths of ContextNet that achieves good trade-off between computation and accuracy. We demonstrate that on the widely used LibriSpeech benchmark, ContextNet achieves a word error rate (WER) of 2.1%/4.6% without external language model (LM), 1.9%/4.1% with LM and 2.9%/7.0% with only 10M parameters on the clean/noisy LibriSpeech test sets. This compares to the previous best published system of 2.0%/4.6% with LM and 3.9%/11.3% with 20M parameters. The superiority of the proposed ContextNet model is also verified on a much larger internal dataset.

eess.AS↗

arXiv2020

Conformer: Convolution-augmented Transformer for Speech Recognition

Recently Transformer and Convolution neural network (CNN) based models have shown promising results in Automatic Speech Recognition (ASR), outperforming Recurrent neural networks (RNNs). Transformer models are good at capturing content-based global interactions, while CNNs exploit local features effectively. In this work, we achieve the best of both worlds by studying how to combine convolution neural networks and transformers to model both local and global dependencies of an audio sequence in a parameter-efficient way. To this regard, we propose the convolution-augmented transformer for speech recognition, named Conformer. Conformer significantly outperforms the previous Transformer and CNN based models achieving state-of-the-art accuracies. On the widely used LibriSpeech benchmark, our model achieves WER of 2.1%/4.3% without using a language model and 1.9%/3.9% with an external language model on test/testother. We also observe competitive performance of 2.7%/6.3% with a small model of only 10M parameters.

eess.AS↗

arXiv2020

Thermal generation, manipulation and detection of skyrmions

Recent years have witnessed significant progresses in realizing skyrmions in chiral magnets1-4 and asymmetric magnetic multilayers5-13, as well as their electrical manipulation2,7,8,10. Equally important, thermal generation, manipulation and detection of skyrmions can be exploited for prototypical new architecture with integrated computation14 and energy harvesting15. It has yet to verify if skyrmions can be purely generated by heating16,17, and if their resultant direction of motion driven by temperature gradients follows the diffusion or, oppositely, the magnonic spin torque17-21. Here, we address these important issues in microstructured devices made of multilayers: (Ta_CoFeB_MgO)15, (Pt_CoFeB_MgO_Ta)15 and (Pt_Co_Ta)15 integrated with on-chip heaters, by using a full-field soft X-ray microscopy. The thermal generation of densely packed skyrmions is attributed to the low energy barrier at the device edge, together with the thermally induced morphological transition from stripe domains to skyrmions. The unidirectional diffusion of skyrmions from the hot region towards the cold region is experimentally observed. It can be theoretically explained by the combined contribution from repulsive forces between skyrmions, and thermal spin-orbit torques in competing with magnonic spin torques17,18,20,21 and entropic forces22. These thermally generated skyrmions can be further electrically detected by measuring the accompanied anomalous Nernst voltages23. The on-chip thermoelectric generation, manipulation and detection of skyrmions could open another exciting avenue for enabling skyrmionics, and promote interdisciplinary studies among spin caloritronics15, magnonics24 and skyrmionics3,4,12.

cond-mat.mes-hall↗

arXiv2020

Scalability in Perception for Autonomous Driving: Waymo Open Dataset

The research community has increasing interest in autonomous driving research, despite the resource intensity of obtaining representative real world data. Existing self-driving datasets are limited in the scale and variation of the environments they capture, even though generalization within and between operating regions is crucial to the overall viability of the technology. In an effort to help align the research community's contributions with real-world self-driving problems, we introduce a new large scale, high quality, diverse dataset. Our new dataset consists of 1150 scenes that each span 20 seconds, consisting of well synchronized and calibrated high quality LiDAR and camera data captured across a range of urban and suburban geographies. It is 15x more diverse than the largest camera+LiDAR dataset available based on our proposed diversity metric. We exhaustively annotated this data with 2D (camera image) and 3D (LiDAR) bounding boxes, with consistent identifiers across frames. Finally, we provide strong baselines for 2D as well as 3D detection and tracking tasks. We further study the effects of dataset size and generalization across geographies on 3D detection methods. Find data, code and more up-to-date information at http://www.waymo.com/open.

cs.CV↗

arXiv2020

Streaming Object Detection for 3-D Point Clouds

Autonomous vehicles operate in a dynamic environment, where the speed with which a vehicle can perceive and react impacts the safety and efficacy of the system. LiDAR provides a prominent sensory modality that informs many existing perceptual systems including object detection, segmentation, motion estimation, and action recognition. The latency for perceptual systems based on point cloud data can be dominated by the amount of time for a complete rotational scan (e.g. 100 ms). This built-in data capture latency is artificial, and based on treating the point cloud as a camera image in order to leverage camera-inspired architectures. However, unlike camera sensors, most LiDAR point cloud data is natively a streaming data source in which laser reflections are sequentially recorded based on the precession of the laser beam. In this work, we explore how to build an object detector that removes this artificial latency constraint, and instead operates on native streaming data in order to significantly reduce latency. This approach has the added benefit of reducing the peak computational burden on inference hardware by spreading the computation over the acquisition time for a scan. We demonstrate a family of streaming detection systems based on sequential modeling through a series of modifications to the traditional detection meta-architecture. We highlight how this model may achieve competitive if not superior predictive performance with state-of-the-art, traditional non-streaming detection systems while achieving significant latency gains (e.g. 1/15'th - 1/3'rd of peak latency). Our results show that operating on LiDAR data in its native streaming formulation offers several advantages for self driving object detection -- advantages that we hope will be useful for any LiDAR perception system where minimizing latency is critical for safe and efficient operation.

cs.CV↗

arXiv2020

Discovery of two-dimensional anisotropic superconductivity at KTaO$_3$ (111) interfaces

The unique electronic structure found at interfaces between materials can allow unconventional quantum states to emerge. Here we observe superconductivity in electron gases formed at interfaces between (111) oriented KTaO$_3$ and insulating overlayers of either EuO or LaAlO$_3$. The superconducting transition temperature, approaching 2.2 K, is about one order of magnitude higher than that of the LaAlO$_3$/SrTiO$_3$ system. Strikingly, similar electron gases at (001) KTaO$_3$ interfaces remain normal down to 25 mK. The critical field and current-voltage measurements indicate that the superconductivity is two dimensional. Higher mobility EuO/KTaO$_3$ (111) samples show a large in-plane anisotropy in transport properties at low temperatures prior to onset of superconductivity, suggesting the emergence of a stripe like phase where the superconductivity is nearly homogeneous in one direction, but strongly modulated in the other.

cond-mat.supr-con↗

arXiv2020

FFusionCGAN: An end-to-end fusion method for few-focus images using conditional GAN in cytopathological digital slides

Multi-focus image fusion technologies compress different focus depth images into an image in which most objects are in focus. However, although existing image fusion techniques, including traditional algorithms and deep learning-based algorithms, can generate high-quality fused images, they need multiple images with different focus depths in the same field of view. This criterion may not be met in some cases where time efficiency is required or the hardware is insufficient. The problem is especially prominent in large-size whole slide images. This paper focused on the multi-focus image fusion of cytopathological digital slide images, and proposed a novel method for generating fused images from single-focus or few-focus images based on conditional generative adversarial network (GAN). Through the adversarial learning of the generator and discriminator, the method is capable of generating fused images with clear textures and large depth of field. Combined with the characteristics of cytopathological images, this paper designs a new generator architecture combining U-Net and DenseBlock, which can effectively improve the network's receptive field and comprehensively encode image features. Meanwhile, this paper develops a semantic segmentation network that identifies the blurred regions in cytopathological images. By integrating the network into the generative model, the quality of the generated fused images is effectively improved. Our method can generate fused images from only single-focus or few-focus images, thereby avoiding the problem of collecting multiple images of different focus depths with increased time and hardware costs. Furthermore, our model is designed to learn the direct mapping of input source images to fused images without the need to manually design complex activity level measurements and fusion rules as in traditional methods.

cs.CV↗

arXiv2019

Highly Conductive Co3Se4 Embedded in N-doped 3D Interconnected Carbonaceous Networks for Enhanced Lithium and Sodium Storage

Traditional cobalt selenides as active materials in lithium-ion batteries (LIBs) and sodium-ion batteries (SIBs) would suffer from drastic volume expansions and large stacking effects, leading to a low cycling stability. In this work, we utilized a facile template method for preparing Co3Se4@N-CN (CSNC) that encapsulated Co3Se4 nanoparticles into 3D interconnected nitrogen-doped carbon network (N-CN). Satisfactorily, it possesses excellent cycling stability with enhanced lithium and sodium energy storage capacity. As an anode material in LIBs, CSNC exhibited a prominent reversible discharge performance of 1313.5 mAh g-1 after 100 cycles at 0.1 A g-1 and 835.6 mAh g-1 after 500 cycles at 1.0 A g-1. Interestingly, according to the analysis from cyclic voltammetry, the in-situ generated Se might provide extra capacity that leaded to a rising trend of capacity. When utilized as an anode in SIBs, CSNC delivered an outstanding capacity of 448.7 mAh g-1 after 100 cycles at 0.1 A g-1 and could retain 328.9 mAh g-1 (77.2% of that of 0.1 A g-1) even at a high current density of 5.0 A g-1. The results demonstrate that CSNC is a superior anode material in LIBs and SIBs with great promise. More importantly, this strategy opens up an effective avenue for the design of transition metal selenide/carbonaceous composites for advanced battery storage systems.

physics.app-ph↗

arXiv2019

StarNet: Targeted Computation for Object Detection in Point Clouds

Detecting objects from LiDAR point clouds is an important component of self-driving car technology as LiDAR provides high resolution spatial information. Previous work on point-cloud 3D object detection has re-purposed convolutional approaches from traditional camera imagery. In this work, we present an object detection system called StarNet designed specifically to take advantage of the sparse and 3D nature of point cloud data. StarNet is entirely point-based, uses no global information, has data dependent anchors, and uses sampling instead of learned region proposals. We demonstrate how this design leads to competitive or superior performance on the large Waymo Open Dataset and the KITTI detection dataset, as compared to convolutional baselines. In particular, we show how our detector can outperform a competitive baseline on Pedestrian detection on the Waymo Open Dataset by more than 7 absolute mAP while being more computationally efficient. We show how our redesign---namely using only local information and using sampling instead of learned proposals---leads to a significantly more flexible and adaptable system: we demonstrate how we can vary the computational cost of a single trained StarNet without retraining, and how we can target proposals towards areas of interest with priors and heuristics. Finally, we show how our design allows for incorporating temporal context by using detections from previous frames to target computation of the detector, which leads to further improvements in performance without additional computational cost.

cs.CV↗

arXiv2019

A comparison of end-to-end models for long-form speech recognition

End-to-end automatic speech recognition (ASR) models, including both attention-based models and the recurrent neural network transducer (RNN-T), have shown superior performance compared to conventional systems. However, previous studies have focused primarily on short utterances that typically last for just a few seconds or, at most, a few tens of seconds. Whether such architectures are practical on long utterances that last from minutes to hours remains an open question. In this paper, we both investigate and improve the performance of end-to-end models on long-form transcription. We first present an empirical comparison of different end-to-end models on a real world long-form task and demonstrate that the RNN-T model is much more robust than attention-based systems in this regime. We next explore two improvements to attention-based systems that significantly improve its performance: restricting the attention to be monotonic, and applying a novel decoding algorithm that breaks long utterances into shorter overlapping segments. Combining these two improvements, we show that attention-based end-to-end models can be very competitive to RNN-T on long-form speech recognition.

eess.AS↗

arXiv2019

Statistical Inference in Mean-Field Variational Bayes

We conduct non-asymptotic analysis on the mean-field variational inference for approximating posterior distributions in complex Bayesian models that may involve latent variables. We show that the mean-field approximation to the posterior can be well-approximated relative to the Kullback-Leibler divergence discrepancy measure by a normal distribution whose center is the maximum likelihood estimator (MLE). In particular, our results imply that the center of the mean-field approximation matches the MLE up to higher-order terms and there is essentially no loss of efficiency in using it as a point estimator for the parameter in any regular parametric model with latent variables. We also propose a new class of variational weighted likelihood bootstrap (VWLB) methods for quantifying the uncertainty in the mean-field variational inference. The proposed VWLB can be viewed as a new sampling scheme that produces independent samples for approximating the posterior. Comparing with traditional sampling algorithms such Markov Chain Monte Carlo, VWLB can be implemented in parallel and is free of tuning.

math.ST↗

arXiv2019

Anomalous Hall effect mechanisms in quasi-2D van der Waals ferromagnet Fe0.29TaS2

The recent emergence of two-dimensional (2D) van der Waals ferromagnets has provided a new platform for exploring magnetism in the flatland and for designing 2D ferromagnet-based spintronics devices. Despite intensive studies, the anomalous Hall effect (AHE) mechanisms in 2D van der Waals ferromagnets have not been investigated yet. In this paper, we report the AHE mechanisms in quasi-2D van der Waals ferromagnet Fe0.29TaS2 via systematically measuring Fe0.29TaS2 devices with thickness from 14 nm to bulk single crystal. The AHE mechanisms are investigated via the scaling relationship between the anomalous Hall and channel conductivities. As the Fe0.29TaS2 thickness decreases, the major AHE mechanism changes from extrinsic scattering to intrinsic contribution. The crossover of the AHE mechanisms is found to be highly associated with the channel conductivities as the Fe0.29TaS2 thickness varies.

cond-mat.mtrl-sci↗

arXiv2019

Interface ferromagnetism and anomalous Hall effect of CdO/ferromagnetic insulator heterostructures

The experimental observation of quantum anomalous Hall effect (QAHE) in magnetic topological insulators has stimulated enormous interest in condensed-matter physics and materials science. For the purpose of realizing high-temperature QAHE, several material candidates have been proposed, among which the interface states in the CdO/ferromagnetic insulator heterostructures are particularly interesting and favorable for technological applications. Here, we report the experimental observation of the interfacial ferromagnetism and anomalous Hall effect in the Fe3O4/CdO/Fe3O4 heterostructures grown via oxide molecular-beam epitaxy. Systematical variation of the CdO thickness reveals the interface ferromagnetism as the major cause for the observed planar magnetoresistance and anomalous Hall effect. Our results might pave the way to engineer oxide interface states for the exploration of QAHE towards exotic quantum-physical phenomena and potential applications.

cond-mat.mtrl-sci↗

arXiv2019

Probing Magnetism in Insulating Cr2Ge2Te6 by Induced Anomalous Hall Effect in Pt

Two-dimensional ferromagnet Cr2Ge2Te6 (CGT) is so resistive below its Curie temperature that probing its magnetism by electrical transport becomes extremely difficult. By forming heterostructures with Pt, however, we observe clear anomalous Hall effect (AHE) in 5 nm thick Pt deposited on thin (< 50 nm) exfoliated flakes of CGT. The AHE hysteresis loops persist to ~ 60 K, which matches well to the Curie temperature of CGT obtained from the bulk magnetization measurements. The slanted AHE loops with a narrow opening indicate magnetic domain formation, which is confirmed by low-temperature magnetic force microscopy (MFM) imaging. These results clearly demonstrate that CGT imprints its magnetization in the AHE signal of the Pt layer. Density functional theory calculations of CGT/Pt heterostructures suggest that the induced ferromagnetism in Pt may be primarily responsible for the observed AHE. Our results establish a powerful way of investigating magnetism in 2D insulating ferromagnets which can potentially work for monolayer devices.

cond-mat.mes-hall↗

arXiv2019

Magnon transport in quasi-two-dimensional van der Waals antiferromagnets

The recent emergence of 2D van der Waals magnets down to atomic layer thickness provides an exciting platform for exploring quantum magnetism and spintronics applications. The van der Waals nature stabilizes the long-range ferromagnetic order as a result of magnetic anisotropy. Furthermore, giant tunneling magnetoresistance and electrical control of magnetism have been reported. However, the potential of 2D van der Waals magnets for magnonics, magnon-based spintronics, has not been explored yet. Here, we report the experimental observation of long-distance magnon transport in quasi-twodimensional van der Waals antiferromagnet MnPS3, which demonstrates the 2D magnets as promising material candidates for magnonics. As the 2D MnPS3 thickness decreases, a shorter magnon diffusion length is observed, which could be attributed to the surface-impurity-induced magnon scattering. Our results could pave the way for exploring quantum magnonics phenomena and designing future magnonics devices based on 2D van der Waals magnets.

cond-mat.mtrl-sci↗