arXiv subjects

Yi Yang

Publications and source records attributed to Yi Yang.

At least 937 records · Page 52Linked to original sources

arXiv2020

Query-efficient Meta Attack to Deep Neural Networks

Black-box attack methods aim to infer suitable attack patterns to targeted DNN models by only using output feedback of the models and the corresponding input queries. However, due to lack of prior and inefficiency in leveraging the query and feedback information, existing methods are mostly query-intensive for obtaining effective attack patterns. In this work, we propose a meta attack approach that is capable of attacking a targeted model with much fewer queries. Its high queryefficiency stems from effective utilization of meta learning approaches in learning generalizable prior abstraction from the previously observed attack patterns and exploiting such prior to help infer attack patterns from only a few queries and outputs. Extensive experiments on MNIST, CIFAR10 and tiny-Imagenet demonstrate that our meta-attack method can remarkably reduce the number of model queries without sacrificing the attack performance. Besides, the obtained meta attacker is not restricted to a particular model but can be used easily with a fast adaptive ability to attack a variety of models.The code of our work is available at https://github.com/dydjw9/MetaAttack_ICLR2020/.

cs.CV↗

arXiv2020

Direct measurement of temporal correlations above the spin-glass transition by coherent resonant magnetic x-ray spectroscopy

In the 1970s a new paradigm was introduced that interacting quenched systems, such as a spin-glass, have a phase transition in which long time memory of spatial patterns is realized without spatial correlations. The principal methods to study the spin-glass transition, besides some elaborate and elegant theoretical constructions, have been numerical computer simulations and neutron spin echo measurements . We show here that the dynamical correlations of the spin-glass transition are embedded in measurements of the four-spin correlations at very long times. This information is directly available in the temporal correlations of the intensity, which encode the spin-orientation memory, obtained by the technique of resonant magnetic x-ray photon correlation spectroscopy (RM- XPCS). We have implemented this method to observe and accurately characterize the critical slowing down of the spin orientation fluctuations in the classic metallic spin glass alloy Cu(Mn) over time scales of 1 to 1000 secs. Our method opens the way for studying phase transitions in systems such as spin ices, and quantum spin liquids, as well as the structural glass transition.

cond-mat.str-el↗

arXiv2020

Dynamic Inference: A New Approach Toward Efficient Video Action Recognition

Though action recognition in videos has achieved great success recently, it remains a challenging task due to the massive computational cost. Designing lightweight networks is a possible solution, but it may degrade the recognition performance. In this paper, we innovatively propose a general dynamic inference idea to improve inference efficiency by leveraging the variation in the distinguishability of different videos. The dynamic inference approach can be achieved from aspects of the network depth and the number of input video frames, or even in a joint input-wise and network depth-wise manner. In a nutshell, we treat input frames and network depth of the computational graph as a 2-dimensional grid, and several checkpoints are placed on this grid in advance with a prediction module. The inference is carried out progressively on the grid by following some predefined route, whenever the inference process comes across a checkpoint, an early prediction can be made depending on whether the early stop criteria meets. For the proof-of-concept purpose, we instantiate three dynamic inference frameworks using two well-known backbone CNNs. In these instances, we overcome the drawback of limited temporal coverage resulted from an early prediction by a novel frame permutation scheme, and alleviate the conflict between progressive computation and video temporal relation modeling by introducing an online temporal shift module. Extensive experiments are conducted to thoroughly analyze the effectiveness of our ideas and to inspire future research efforts. Results on various datasets also evident the superiority of our approach.

cs.CV↗

arXiv2020

Symbiotic Attention with Privileged Information for Egocentric Action Recognition

Egocentric video recognition is a natural testbed for diverse interaction reasoning. Due to the large action vocabulary in egocentric video datasets, recent studies usually utilize a two-branch structure for action recognition, ie, one branch for verb classification and the other branch for noun classification. However, correlation studies between the verb and the noun branches have been largely ignored. Besides, the two branches fail to exploit local features due to the absence of a position-aware attention mechanism. In this paper, we propose a novel Symbiotic Attention framework leveraging Privileged information (SAP) for egocentric video recognition. Finer position-aware object detection features can facilitate the understanding of actor's interaction with the object. We introduce these features in action recognition and regard them as privileged information. Our framework enables mutual communication among the verb branch, the noun branch, and the privileged information. This communication process not only injects local details into global features but also exploits implicit guidance about the spatio-temporal position of an on-going action. We introduce novel symbiotic attention (SA) to enable effective communication. It first normalizes the detection guided features on one branch to underline the action-relevant information from the other branch. SA adaptively enhances the interactions among the three sources. To further catalyze this communication, spatial relations are uncovered for the selection of most action-relevant information. It identifies the most valuable and discriminative feature for classification. We validate the effectiveness of our SAP quantitatively and qualitatively. Notably, it achieves the state-of-the-art on two large-scale egocentric video datasets.

cs.CV↗

arXiv2020

Progressive Local Filter Pruning for Image Retrieval Acceleration

This paper focuses on network pruning for image retrieval acceleration. Prevailing image retrieval works target at the discriminative feature learning, while little attention is paid to how to accelerate the model inference, which should be taken into consideration in real-world practice. The challenge of pruning image retrieval models is that the middle-level feature should be preserved as much as possible. Such different requirements of the retrieval and classification model make the traditional pruning methods not that suitable for our task. To solve the problem, we propose a new Progressive Local Filter Pruning (PLFP) method for image retrieval acceleration. Specifically, layer by layer, we analyze the local geometric properties of each filter and select the one that can be replaced by the neighbors. Then we progressively prune the filter by gradually changing the filter weights. In this way, the representation ability of the model is preserved. To verify this, we evaluate our method on two widely-used image retrieval datasets,i.e., Oxford5k and Paris6K, and one person re-identification dataset,i.e., Market-1501. The proposed method arrives with superior performance to the conventional pruning methods, suggesting the effectiveness of the proposed method for image retrieval.

cs.CV↗

arXiv2020

NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search

Neural architecture search (NAS) has achieved breakthrough success in a great number of applications in the past few years. It could be time to take a step back and analyze the good and bad aspects in the field of NAS. A variety of algorithms search architectures under different search space. These searched architectures are trained using different setups, e.g., hyper-parameters, data augmentation, regularization. This raises a comparability problem when comparing the performance of various NAS algorithms. NAS-Bench-101 has shown success to alleviate this problem. In this work, we propose an extension to NAS-Bench-101: NAS-Bench-201 with a different search space, results on multiple datasets, and more diagnostic information. NAS-Bench-201 has a fixed search space and provides a unified benchmark for almost any up-to-date NAS algorithms. The design of our search space is inspired from the one used in the most popular cell-based searching algorithms, where a cell is represented as a DAG. Each edge here is associated with an operation selected from a predefined operation set. For it to be applicable for all NAS algorithms, the search space defined in NAS-Bench-201 includes all possible architectures generated by 4 nodes and 5 associated operation options, which results in 15,625 candidates in total. The training log and the performance for each architecture candidate are provided for three datasets. This allows researchers to avoid unnecessary repetitive training for selected candidate and focus solely on the search algorithm itself. The training time saved for every candidate also largely improves the efficiency of many methods. We provide additional diagnostic information such as fine-grained loss and accuracy, which can give inspirations to new designs of NAS algorithms. In further support, we have analyzed it from many aspects and benchmarked 10 recent NAS algorithms.

cs.CV↗

arXiv2020

Early Ultra-Violet observations of type IIn supernovae constrain the asphericity of their circumstellar material

We present a survey of the early evolution of 12 Type IIn supernovae (SNe IIn) in the Ultra-Violet (UV) and visible light. We use this survey to constrain the geometry of the circumstellar material (CSM) surrounding SN IIn explosions, which may shed light on their progenitor diversity. In order to distinguish between aspherical and spherical circumstellar material (CSM), we estimate the blackbody radius temporal evolution of the SNe IIn of our sample, following the method introduced by Soumagnac et al. We find that higher luminosity objects tend to show evidence for aspherical CSM. Depending on whether this correlation is due to physical reasons or to some selection bias, we derive a lower limit between 35% and 66% on the fraction of SNe IIn showing evidence for aspherical CSM. This result suggests that asphericity of the CSM surrounding SNe IIn is common - consistent with data from resolved images of stars undergoing considerable mass loss. It should be taken into account for more realistic modelling of these events.

astro-ph.HE↗

arXiv2020

Similarity-preserving Image-image Domain Adaptation for Person Re-identification

This article studies the domain adaptation problem in person re-identification (re-ID) under a "learning via translation" framework, consisting of two components, 1) translating the labeled images from the source to the target domain in an unsupervised manner, 2) learning a re-ID model using the translated images. The objective is to preserve the underlying human identity information after image translation, so that translated images with labels are effective for feature learning on the target domain. To this end, we propose a similarity preserving generative adversarial network (SPGAN) and its end-to-end trainable version, eSPGAN. Both aiming at similarity preserving, SPGAN enforces this property by heuristic constraints, while eSPGAN does so by optimally facilitating the re-ID model learning. More specifically, SPGAN separately undertakes the two components in the "learning via translation" framework. It first preserves two types of unsupervised similarity, namely, self-similarity of an image before and after translation, and domain-dissimilarity of a translated source image and a target image. It then learns a re-ID model using existing networks. In comparison, eSPGAN seamlessly integrates image translation and re-ID model learning. During the end-to-end training of eSPGAN, re-ID learning guides image translation to preserve the underlying identity information of an image. Meanwhile, image translation improves re-ID learning by providing identity-preserving training samples of the target domain style. In the experiment, we show that identities of the fake images generated by SPGAN and eSPGAN are well preserved. Based on this, we report the new state-of-the-art domain adaptation results on two large-scale person re-ID datasets.

cs.CV↗

arXiv2019

Very Long Natural Scenery Image Prediction by Outpainting

Comparing to image inpainting, image outpainting receives less attention due to two challenges in it. The first challenge is how to keep the spatial and content consistency between generated images and original input. The second challenge is how to maintain high quality in generated results, especially for multi-step generations in which generated regions are spatially far away from the initial input. To solve the two problems, we devise some innovative modules, named Skip Horizontal Connection and Recurrent Content Transfer, and integrate them into our designed encoder-decoder structure. By this design, our network can generate highly realistic outpainting prediction effectively and efficiently. Other than that, our method can generate new images with very long sizes while keeping the same style and semantic content as the given input. To test the effectiveness of the proposed architecture, we collect a new scenery dataset with diverse, complicated natural scenes. The experimental results on this dataset have demonstrated the efficacy of our proposed network. The code and dataset are available from https://github.com/z-x-yang/NS-Outpainting.

cs.CV↗

arXiv2019

Utterance-level Permutation Invariant Training with Latency-controlled BLSTM for Single-channel Multi-talker Speech Separation

Utterance-level permutation invariant training (uPIT) has achieved promising progress on single-channel multi-talker speech separation task. Long short-term memory (LSTM) and bidirectional LSTM (BLSTM) are widely used as the separation networks of uPIT, i.e. uPIT-LSTM and uPIT-BLSTM. uPIT-LSTM has lower latency but worse performance, while uPIT-BLSTM has better performance but higher latency. In this paper, we propose using latency-controlled BLSTM (LC-BLSTM) during inference to fulfill low-latency and good-performance speech separation. To find a better training strategy for BLSTM-based separation network, chunk-level PIT (cPIT) and uPIT are compared. The experimental results show that uPIT outperforms cPIT when LC-BLSTM is used during inference. It is also found that the inter-chunk speaker tracing (ST) can further improve the separation performance of uPIT-LC-BLSTM. Evaluated on the WSJ0 two-talker mixed-speech separation task, the absolute gap of signal-to-distortion ratio (SDR) between uPIT-BLSTM and uPIT-LC-BLSTM is reduced to within 0.7 dB.

cs.SD↗

arXiv2019

High-Resolution Near-Infrared Polarimetry and Sub-Millimeter Imaging of FS Tau A: Possible Streamers in Misaligned Circumbinary Disk System

We analyzed the young (2.8-Myr-old) binary system FS Tau A using near-infrared (H-band) high-contrast polarimetry data from Subaru/HiCIAO and sub-millimeter CO (J=2-1) line emission data from ALMA. Both the near-infrared and sub-millimeter observations reveal several clear structures extending to $\sim$240 AU from the stars. Based on these observations at different wavelengths, we report the following discoveries. One arm-like structure detected in the near-infrared band initially extends from the south of the binary with a subsequent turn to the northeast, corresponding to two bar-like structures detected in ALMA observations with an LSRK velocity of 1.19-5.64 km/s. Another feature detected in the near-infrared band extends initially from the north of the binary, relating to an arm-like structure detected in ALMA observations with an LSRK velocity of 8.17-16.43 km/s. From their shapes and velocities, we suggest that these structures can mostly be explained by two streamers that connect the outer circumbinary disk and the central binary components. These discoveries will be helpful for understanding the evolution of streamers and circumstellar disks in young binary systems.

astro-ph.SR↗

arXiv2019

SUBARU Near-Infrared Imaging Polarimetry of Misaligned Disks Around The SR24 Hierarchical Triple System

The SR24 multi-star system hosts both circumprimary and circumsecondary disks, which are strongly misaligned from each other. The circumsecondary disk is circumbinary in nature. Interestingly, both disks are interacting, and they possibly rotate in opposite directions. To investigate the nature of this unique twin disk system, we present 0.''1 resolution near-infrared polarized intensity images of the circumstellar structures around SR24, obtained with HiCIAO mounted on the Subaru 8.2 m telescope. Both the circumprimary disk and the circumsecondary disk are resolved and have elongated features. While the position angle of the major axis and radius of the NIR polarization disk around SR24S are 55$^{\circ}$ and 137 au, respectively, those around SR24N are 110$^{\circ}$ and 34 au, respectively. With regard to overall morphology, the circumprimary disk around SR24S shows strong asymmetry, whereas the circumsecondary disk around SR24N shows relatively strong symmetry. Our NIR observations confirm the previous claim that the circumprimary and circumsecondary disks are misaligned from each other. Both the circumprimary and circumsecondary disks show similar structures in $^{12}$CO observations in terms of its size and elongation direction. This consistency is because both NIR and $^{12}$CO are tracing surface layers of the flared disks. As the radius of the polarization disk around SR24N is roughly consistent with the size of the outer Roche lobe, it is natural to interpret the polarization disk around SR24N as a circumbinary disk surrounding the SR24Nb-Nc system.

astro-ph.SR↗

arXiv2019

PointRNN: Point Recurrent Neural Network for Moving Point Cloud Processing

In this paper, we introduce a Point Recurrent Neural Network (PointRNN) for moving point cloud processing. At each time step, PointRNN takes point coordinates $\boldsymbol{P} \in \mathbb{R}^{n \times 3}$ and point features $\boldsymbol{X} \in \mathbb{R}^{n \times d}$ as input ($n$ and $d$ denote the number of points and the number of feature channels, respectively). The state of PointRNN is composed of point coordinates $\boldsymbol{P}$ and point states $\boldsymbol{S} \in \mathbb{R}^{n \times d'}$ ($d'$ denotes the number of state channels). Similarly, the output of PointRNN is composed of $\boldsymbol{P}$ and new point features $\boldsymbol{Y} \in \mathbb{R}^{n \times d''}$ ($d''$ denotes the number of new feature channels). Since point clouds are orderless, point features and states from two time steps can not be directly operated. Therefore, a point-based spatiotemporally-local correlation is adopted to aggregate point features and states according to point coordinates. We further propose two variants of PointRNN, i.e., Point Gated Recurrent Unit (PointGRU) and Point Long Short-Term Memory (PointLSTM). We apply PointRNN, PointGRU and PointLSTM to moving point cloud prediction, which aims to predict the future trajectories of points in a set given their history movements. Experimental results show that PointRNN, PointGRU and PointLSTM are able to produce correct predictions on both synthetic and real-world datasets, demonstrating their ability to model point cloud sequences. The code has been released at \url{https://github.com/hehefan/PointRNN}.

cs.CV↗

arXiv2019

Tweedie Gradient Boosting for Extremely Unbalanced Zero-inflated Data

Tweedie's compound Poisson model is a popular method to model insurance claims with probability mass at zero and nonnegative, highly right-skewed distribution. In particular, it is not uncommon to have extremely unbalanced data with excessively large proportion of zero claims, and even traditional Tweedie model may not be satisfactory for fitting the data. In this paper, we propose a boosting-assisted zero-inflated Tweedie model, called EMTboost, that allows zero probability mass to exceed a traditional model. We makes a nonparametric assumption on its Tweedie model component, that unlike a linear model, is able to capture nonlinearities, discontinuities, and complex higher order interactions among predictors. A specialized Expectation-Maximization algorithm is developed that integrates a blockwise coordinate descent strategy and a gradient tree-boosting algorithm to estimate key model parameters. We use extensive simulation and data analysis on synthetic zero-inflated auto-insurance claim data to illustrate our method's prediction performance.

stat.CO↗

arXiv2019

Asymptotic Soft Filter Pruning for Deep Convolutional Neural Networks

Deeper and wider Convolutional Neural Networks (CNNs) achieve superior performance but bring expensive computation cost. Accelerating such over-parameterized neural network has received increased attention. A typical pruning algorithm is a three-stage pipeline, i.e., training, pruning, and retraining. Prevailing approaches fix the pruned filters to zero during retraining, and thus significantly reduce the optimization space. Besides, they directly prune a large number of filters at first, which would cause unrecoverable information loss. To solve these problems, we propose an Asymptotic Soft Filter Pruning (ASFP) method to accelerate the inference procedure of the deep neural networks. First, we update the pruned filters during the retraining stage. As a result, the optimization space of the pruned model would not be reduced but be the same as that of the original model. In this way, the model has enough capacity to learn from the training data. Second, we prune the network asymptotically. We prune few filters at first and asymptotically prune more filters during the training procedure. With asymptotic pruning, the information of the training set would be gradually concentrated in the remaining filters, so the subsequent training and pruning process would be stable. Experiments show the effectiveness of our ASFP on image classification benchmarks. Notably, on ILSVRC-2012, our ASFP reduces more than 40% FLOPs on ResNet-50 with only 0.14% top-5 accuracy degradation, which is higher than the soft filter pruning (SFP) by 8%.

cs.CV↗

arXiv2019

Self-Correction for Human Parsing

Labeling pixel-level masks for fine-grained semantic segmentation tasks, e.g. human parsing, remains a challenging task. The ambiguous boundary between different semantic parts and those categories with similar appearance usually are confusing, leading to unexpected noises in ground truth masks. To tackle the problem of learning with label noises, this work introduces a purification strategy, called Self-Correction for Human Parsing (SCHP), to progressively promote the reliability of the supervised labels as well as the learned models. In particular, starting from a model trained with inaccurate annotations as initialization, we design a cyclically learning scheduler to infer more reliable pseudo-masks by iteratively aggregating the current learned model with the former optimal one in an online manner. Besides, those correspondingly corrected labels can in turn to further boost the model performance. In this way, the models and the labels will reciprocally become more robust and accurate during the self-correction learning cycles. Benefiting from the superiority of SCHP, we achieve the best performance on two popular single-person human parsing benchmarks, including LIP and Pascal-Person-Part datasets. Our overall system ranks 1st in CVPR2019 LIP Challenge. Code is available at https://github.com/PeikeLi/Self-Correction-Human-Parsing.

cs.CV↗

arXiv2019

Network Pruning via Transformable Architecture Search

Network pruning reduces the computation costs of an over-parameterized network without performance damage. Prevailing pruning algorithms pre-define the width and depth of the pruned networks, and then transfer parameters from the unpruned network to pruned networks. To break the structure limitation of the pruned networks, we propose to apply neural architecture search to search directly for a network with flexible channel and layer sizes. The number of the channels/layers is learned by minimizing the loss of the pruned networks. The feature map of the pruned network is an aggregation of K feature map fragments (generated by K networks of different sizes), which are sampled based on the probability distribution.The loss can be back-propagated not only to the network weights, but also to the parameterized distribution to explicitly tune the size of the channels/layers. Specifically, we apply channel-wise interpolation to keep the feature map with different channel sizes aligned in the aggregation procedure. The maximum probability for the size in each distribution serves as the width and depth of the pruned network, whose parameters are learned by knowledge transfer, e.g., knowledge distillation, from the original networks. Experiments on CIFAR-10, CIFAR-100 and ImageNet demonstrate the effectiveness of our new perspective of network pruning compared to traditional network pruning algorithms. Various searching and knowledge transfer approaches are conducted to show the effectiveness of the two components. Code is at: https://github.com/D-X-Y/NAS-Projects.

cs.CV↗

arXiv2019

Searching for A Robust Neural Architecture in Four GPU Hours

Conventional neural architecture search (NAS) approaches are based on reinforcement learning or evolutionary strategy, which take more than 3000 GPU hours to find a good model on CIFAR-10. We propose an efficient NAS approach learning to search by gradient descent. Our approach represents the search space as a directed acyclic graph (DAG). This DAG contains billions of sub-graphs, each of which indicates a kind of neural architecture. To avoid traversing all the possibilities of the sub-graphs, we develop a differentiable sampler over the DAG. This sampler is learnable and optimized by the validation loss after training the sampled architecture. In this way, our approach can be trained in an end-to-end fashion by gradient descent, named Gradient-based search using Differentiable Architecture Sampler (GDAS). In experiments, we can finish one searching procedure in four GPU hours on CIFAR-10, and the discovered model obtains a test error of 2.82\% with only 2.5M parameters, which is on par with the state-of-the-art. Code is publicly available on GitHub: https://github.com/D-X-Y/NAS-Projects.

cs.CV↗