Searcharxiv⌕ Search

arXiv subjects

Qi Chen

Publications and source records attributed to Qi Chen.

At least 271 records · Page 15Linked to original sources
arXiv2022

Proposal-free Lidar Panoptic Segmentation with Pillar-level Affinity

We propose a simple yet effective proposal-free architecture for lidar panoptic segmentation. We jointly optimize both semantic segmentation and class-agnostic instance classification in a single network using a pillar-based bird's-eye view representation. The instance classification head learns pairwise affinity between pillars to determine whether the pillars belong to the same instance or not. We further propose a local clustering algorithm to propagate instance ids by merging semantic segmentation and affinity predictions. Our experiments on nuScenes dataset show that our approach outperforms previous proposal-free methods and is comparable to proposal-based methods which requires extra annotation from object detection.

cs.CV↗
arXiv2022

PolarStream: Streaming Lidar Object Detection and Segmentation with Polar Pillars

Recent works recognized lidars as an inherently streaming data source and showed that the end-to-end latency of lidar perception models can be reduced significantly by operating on wedge-shaped point cloud sectors rather then the full point cloud. However, due to use of cartesian coordinate systems these methods represent the sectors as rectangular regions, wasting memory and compute. In this work we propose using a polar coordinate system and make two key improvements on this design. First, we increase the spatial context by using multi-scale padding from neighboring sectors: preceding sector from the current scan and/or the following sector from the past scan. Second, we improve the core polar convolutional architecture by introducing feature undistortion and range stratified convolutions. Experimental results on the nuScenes dataset show significant improvements over other streaming based methods. We also achieve comparable results to existing non-streaming methods but with lower latencies. The code and pretrained models are available at \url{https://github.com/motional/polarstream}.

cs.CV↗
arXiv2022

Self-supervised Image-specific Prototype Exploration for Weakly Supervised Semantic Segmentation

Weakly Supervised Semantic Segmentation (WSSS) based on image-level labels has attracted much attention due to low annotation costs. Existing methods often rely on Class Activation Mapping (CAM) that measures the correlation between image pixels and classifier weight. However, the classifier focuses only on the discriminative regions while ignoring other useful information in each image, resulting in incomplete localization maps. To address this issue, we propose a Self-supervised Image-specific Prototype Exploration (SIPE) that consists of an Image-specific Prototype Exploration (IPE) and a General-Specific Consistency (GSC) loss. Specifically, IPE tailors prototypes for every image to capture complete regions, formed our Image-Specific CAM (IS-CAM), which is realized by two sequential steps. In addition, GSC is proposed to construct the consistency of general CAM and our specific IS-CAM, which further optimizes the feature representation and empowers a self-correction ability of prototype exploration. Extensive experiments are conducted on PASCAL VOC 2012 and MS COCO 2014 segmentation benchmark and results show our SIPE achieves new state-of-the-art performance using only image-level labels. The code is available at https://github.com/chenqi1126/SIPE.

cs.CV↗
arXiv2021

Generalization Bounds For Meta-Learning: An Information-Theoretic Analysis

We derive a novel information-theoretic analysis of the generalization property of meta-learning algorithms. Concretely, our analysis proposes a generic understanding of both the conventional learning-to-learn framework and the modern model-agnostic meta-learning (MAML) algorithms. Moreover, we provide a data-dependent generalization bound for a stochastic variant of MAML, which is non-vacuous for deep few-shot learning. As compared to previous bounds that depend on the square norm of gradients, empirical validations on both simulated data and a well-known few-shot benchmark show that our bound is orders of magnitude tighter in most situations.

cs.LG↗
arXiv2021

V2C: Visual Voice Cloning

Existing Voice Cloning (VC) tasks aim to convert a paragraph text to a speech with desired voice specified by a reference audio. This has significantly boosted the development of artificial speech applications. However, there also exist many scenarios that cannot be well reflected by these VC tasks, such as movie dubbing, which requires the speech to be with emotions consistent with the movie plots. To fill this gap, in this work we propose a new task named Visual Voice Cloning (V2C), which seeks to convert a paragraph of text to a speech with both desired voice specified by a reference audio and desired emotion specified by a reference video. To facilitate research in this field, we construct a dataset, V2C-Animation, and propose a strong baseline based on existing state-of-the-art (SoTA) VC techniques. Our dataset contains 10,217 animated movie clips covering a large variety of genres (e.g., Comedy, Fantasy) and emotions (e.g., happy, sad). We further design a set of evaluation metrics, named MCD-DTW-SL, which help evaluate the similarity between ground-truth speeches and the synthesised ones. Extensive experimental results show that even SoTA VC methods cannot generate satisfying speeches for our V2C task. We hope the proposed new task together with the constructed dataset and evaluation metric will facilitate the research in the field of voice cloning and the broader vision-and-language community.

cs.CV↗
arXiv2021

SPANN: Highly-efficient Billion-scale Approximate Nearest Neighbor Search

The in-memory algorithms for approximate nearest neighbor search (ANNS) have achieved great success for fast high-recall search, but are extremely expensive when handling very large scale database. Thus, there is an increasing request for the hybrid ANNS solutions with small memory and inexpensive solid-state drive (SSD). In this paper, we present a simple but efficient memory-disk hybrid indexing and search system, named SPANN, that follows the inverted index methodology. It stores the centroid points of the posting lists in the memory and the large posting lists in the disk. We guarantee both disk-access efficiency (low latency) and high recall by effectively reducing the disk-access number and retrieving high-quality posting lists. In the index-building stage, we adopt a hierarchical balanced clustering algorithm to balance the length of posting lists and augment the posting list by adding the points in the closure of the corresponding clusters. In the search stage, we use a query-aware scheme to dynamically prune the access of unnecessary posting lists. Experiment results demonstrate that SPANN is 2$\times$ faster than the state-of-the-art ANNS solution DiskANN to reach the same recall quality $90\%$ with same memory cost in three billion-scale datasets. It can reach $90\%$ recall@1 and recall@10 in just around one millisecond with only 32GB memory cost. Code is available at: {\footnotesize\color{blue}{\url{https://github.com/microsoft/SPTAG}}}.

cs.DB↗
arXiv2021

Heterogeneously integrated, superconducting silicon-photonic platform for measurement-device-independent quantum key distribution

Integrated photonics provides a route both to miniaturize quantum key distribution (QKD) devices and to enhance their performance. A key element for achieving discrete-variable QKD is a single-photon detector. It is highly desirable to integrate detectors onto a photonic chip to enable the realization of practical and scalable quantum networks. We realize an integrated heterogeneous superconducting-silicon-photonic chip. Harnessing the unique high-speed feature of our optical waveguide-integrated superconducting detector, we perform the first optimal Bell-state measurement (BSM) of time-bin encoded qubits generated from two independent lasers. The optimal BSM enables an increased key rate of measurement-device-independent QKD, which is immune to all attacks against the detection system, and hence provides the basis for a QKD network with untrusted relays. Together with the time-multiplexed technique, we have enhanced the sifted key rate by almost one order of magnitude. With a 125 MHz clock rate, we obtain a secure key rate of 6.166 kbps over 24.0 dB loss, which is comparable to the state-of-the-art MDI-QKD experimental results with GHz clock rate. Combined with integrated QKD transmitters, a scalable, chip-based and cost-effective QKD network should become realizable in the near future.

quant-ph↗
arXiv2021

GLOW : Global Weighted Self-Attention Network for Web Search

Deep matching models aim to facilitate search engines retrieving more relevant documents by mapping queries and documents into semantic vectors in the first-stage retrieval. When leveraging BERT as the deep matching model, the attention score across two words are solely built upon local contextualized word embeddings. It lacks prior global knowledge to distinguish the importance of different words, which has been proved to play a critical role in information retrieval tasks. In addition to this, BERT only performs attention across sub-words tokens which weakens whole word attention representation. We propose a novel Global Weighted Self-Attention (GLOW) network for web document search. GLOW fuses global corpus statistics into the deep matching model. By adding prior weights into attention generation from global information, like BM25, GLOW successfully learns weighted attention scores jointly with query matrix $Q$ and key matrix $K$. We also present an efficient whole word weight sharing solution to bring prior whole word knowledge into sub-words level attention. It aids Transformer to learn whole word level attention. To make our models applicable to complicated web search scenarios, we introduce combined fields representation to accommodate documents with multiple fields even with variable number of instances. We demonstrate GLOW is more efficient to capture the topical and semantic representation both in queries and documents. Intrinsic evaluation and experiments conducted on public data sets reveal GLOW to be a general framework for document retrieve task. It significantly outperforms BERT and other competitive baselines by a large margin while retaining the same model complexity with BERT.

cs.IR↗
arXiv2021

Contrastive Neural Architecture Search with Neural Architecture Comparators

One of the key steps in Neural Architecture Search (NAS) is to estimate the performance of candidate architectures. Existing methods either directly use the validation performance or learn a predictor to estimate the performance. However, these methods can be either computationally expensive or very inaccurate, which may severely affect the search efficiency and performance. Moreover, as it is very difficult to annotate architectures with accurate performance on specific tasks, learning a promising performance predictor is often non-trivial due to the lack of labeled data. In this paper, we argue that it may not be necessary to estimate the absolute performance for NAS. On the contrary, we may need only to understand whether an architecture is better than a baseline one. However, how to exploit this comparison information as the reward and how to well use the limited labeled data remains two great challenges. In this paper, we propose a novel Contrastive Neural Architecture Search (CTNAS) method which performs architecture search by taking the comparison results between architectures as the reward. Specifically, we design and learn a Neural Architecture Comparator (NAC) to compute the probability of candidate architectures being better than a baseline one. Moreover, we present a baseline updating scheme to improve the baseline iteratively in a curriculum learning manner. More critically, we theoretically show that learning NAC is equivalent to optimizing the ranking over architectures. Extensive experiments in three search spaces demonstrate the superiority of our CTNAS over existing methods.

cs.CV↗
arXiv2021

Pareto-Frontier-aware Neural Architecture Generation for Diverse Budgets

Designing feasible and effective architectures under diverse computation budgets incurred by different applications/devices is essential for deploying deep models in practice. Existing methods often perform an independent architecture search for each target budget, which is very inefficient yet unnecessary. Moreover, the repeated independent search manner would inevitably ignore the common knowledge among different search processes and hamper the search performance. To address these issues, we seek to train a general architecture generator that automatically produces effective architectures for an arbitrary budget merely via model inference. To this end, we propose a Pareto-Frontier-aware Neural Architecture Generator (NAG) which takes an arbitrary budget as input and produces the Pareto optimal architecture for the target budget. We train NAG by learning the Pareto frontier (i.e., the set of Pareto optimal architectures) over model performance and computational cost (e.g., latency). Extensive experiments on three platforms (i.e., mobile, CPU, and GPU) show the superiority of the proposed method over existing NAS methods.

cs.LG↗
arXiv2021

Towards Accurate and Compact Architectures via Neural Architecture Transformer

Designing effective architectures is one of the key factors behind the success of deep neural networks. Existing deep architectures are either manually designed or automatically searched by some Neural Architecture Search (NAS) methods. However, even a well-designed/searched architecture may still contain many nonsignificant or redundant modules/operations. Thus, it is necessary to optimize the operations inside an architecture to improve the performance without introducing extra computational cost. To this end, we have proposed a Neural Architecture Transformer (NAT) method which casts the optimization problem into a Markov Decision Process (MDP) and seeks to replace the redundant operations with more efficient operations, such as skip or null connection. Note that NAT only considers a small number of possible transitions and thus comes with a limited search/transition space. As a result, such a small search space may hamper the performance of architecture optimization. To address this issue, we propose a Neural Architecture Transformer++ (NAT++) method which further enlarges the set of candidate transitions to improve the performance of architecture optimization. Specifically, we present a two-level transition rule to obtain valid transitions, i.e., allowing operations to have more efficient types (e.g., convolution->separable convolution) or smaller kernel sizes (e.g., 5x5->3x3). Note that different operations may have different valid transitions. We further propose a Binary-Masked Softmax (BMSoftmax) layer to omit the possible invalid transitions. Extensive experiments on several benchmark datasets show that the transformed architecture significantly outperforms both its original counterpart and the architectures optimized by existing methods.

cs.CV↗
arXiv2021

StrokeGAN: Reducing Mode Collapse in Chinese Font Generation via Stroke Encoding

The generation of stylish Chinese fonts is an important problem involved in many applications. Most of existing generation methods are based on the deep generative models, particularly, the generative adversarial networks (GAN) based models. However, these deep generative models may suffer from the mode collapse issue, which significantly degrades the diversity and quality of generated results. In this paper, we introduce a one-bit stroke encoding to capture the key mode information of Chinese characters and then incorporate it into CycleGAN, a popular deep generative model for Chinese font generation. As a result we propose an efficient method called StrokeGAN, mainly motivated by the observation that the stroke encoding contains amount of mode information of Chinese characters. In order to reconstruct the one-bit stroke encoding of the associated generated characters, we introduce a stroke-encoding reconstruction loss imposed on the discriminator. Equipped with such one-bit stroke encoding and stroke-encoding reconstruction loss, the mode collapse issue of CycleGAN can be significantly alleviated, with an improved preservation of strokes and diversity of generated characters. The effectiveness of StrokeGAN is demonstrated by a series of generation tasks over nine datasets with different fonts. The numerical results demonstrate that StrokeGAN generally outperforms the state-of-the-art methods in terms of content and recognition accuracies, as well as certain stroke error, and also generates more realistic characters.

cs.CV↗
arXiv2021

On the $L^\infty$ stability of Prandtl expansions in Gevrey class

In this paper, we prove the $L^\infty\cap L^2$ stability of Prandtl expansions of shear flow type as $\big(U(y/\sqrtν),0\big)$ for the initial perturbation in the Gevrey class, where $U(y)$ is a monotone and concave function and $ν$ is the viscosity coefficient. To this end, we develop the direct resolvent estimate method for the linearized Orr-Sommerfeld operator instead of the Rayleigh-Airy iteration method introduced by Grenier, Guo and Nguyen.

math.AP↗
arXiv2020

Mid-infrared single photon detector with superconductor Mo$_{80}$Si$_{20}$ nanowire

A mid-infrared single photon detector (MIR-SNSPD) was reported based on 30 nm-wide superconductor molybdenum silicide nanowires in this work. Saturated quantum efficiencies (QEs) were achieved at the wavelength ranging from 1.55 to 5.07 micrometer in experiments. At the same time, the intrinsic dark count rate (DCR) was below 100 cps. Thus, this device produced a noise equivalent power (NEP) of 4.5 * 10-19 W/sqrt(Hz). The results provide the foundation of developing 10 micrometer-SNSPD for the applications of infrared astronomy observation.

physics.app-ph↗
arXiv2020

Multi-frequency Point Supported LLRF Front-end for CiADS Wide-bandwidth Application

The China initiative Accelerator Driven System, CiADS, physics design adopts $162.5 \,\mathrm{MHz}$, $325 \,\mathrm{MHz}$, and $650 \,\mathrm{MHz}$ cavities, which are driven by the corresponding radio frequency (RF) power system, requiring frequency translation front-end for the RF station. For that application, a general-purpose design front-end prototype has been developed to evaluate the multi-frequency point supported design feasibility. The difficult parts to achieve the requirements of the general-purpose design are reasonable device selection and balanced design. With a carefully selected low-noise wide-band RF mixer and amplifier to balance the performance of multi-frequency supported down-conversion, specially designed local oscillator (LO) distribution net to increase isolation between adjacent channels, and external band-pass filter to realize expected up-conversion frequencies, high maintenance and modular front-end general-purpose design has been implemented. Results of standard parameters show an $R^2$ value of at least $99.991\%$ in the range of $-60 \sim 10\,\mathrm{dBm}$ for linearity, up to $18\,\mathrm{dBm}$ for P1dB, and up to $89\,\mathrm{dBc}$ for crosstalk between adjacent channels. The phase noise spectrum is lower than $80\,\mathrm{dBc}$ in the range of $0 \sim 1\,\mathrm{MHz}$, and cumulative phase noise is $0.006^\circ$; amplitude and phase stability are $0.022\%$ and $0.034^\circ$, respectively.

physics.acc-ph↗
arXiv2020

Object as Hotspots: An Anchor-Free 3D Object Detection Approach via Firing of Hotspots

Accurate 3D object detection in LiDAR based point clouds suffers from the challenges of data sparsity and irregularities. Existing methods strive to organize the points regularly, e.g. voxelize, pass them through a designed 2D/3D neural network, and then define object-level anchors that predict offsets of 3D bounding boxes using collective evidences from all the points on the objects of interest. Contrary to the state-of-the-art anchor-based methods, based on the very nature of data sparsity, we observe that even points on an individual object part are informative about semantic information of the object. We thus argue in this paper for an approach opposite to existing methods using object-level anchors. Inspired by compositional models, which represent an object as parts and their spatial relations, we propose to represent an object as composition of its interior non-empty voxels, termed hotspots, and the spatial relations of hotspots. This gives rise to the representation of Object as Hotspots (OHS). Based on OHS, we further propose an anchor-free detection head with a novel ground truth assignment strategy that deals with inter-object point-sparsity imbalance to prevent the network from biasing towards objects with more points. Experimental results show that our proposed method works remarkably well on objects with a small number of points. Notably, our approach ranked 1st on KITTI 3D Detection Benchmark for cyclist and pedestrian detection, and achieved state-of-the-art performance on NuScenes 3D Detection Benchmark.

cs.CV↗
arXiv2020

CoFF: Cooperative Spatial Feature Fusion for 3D Object Detection on Autonomous Vehicles

To reduce the amount of transmitted data, feature map based fusion is recently proposed as a practical solution to cooperative 3D object detection by autonomous vehicles. The precision of object detection, however, may require significant improvement, especially for objects that are far away or occluded. To address this critical issue for the safety of autonomous vehicles and human beings, we propose a cooperative spatial feature fusion (CoFF) method for autonomous vehicles to effectively fuse feature maps for achieving a higher 3D object detection performance. Specially, CoFF differentiates weights among feature maps for a more guided fusion, based on how much new semantic information is provided by the received feature maps. It also enhances the inconspicuous features corresponding to far/occluded objects to improve their detection precision. Experimental results show that CoFF achieves a significant improvement in terms of both detection precision and effective detection range for autonomous vehicles, compared to previous feature fusion solutions.

cs.CV↗