arXiv subjects

Li Shen

Publications and source records attributed to Li Shen.

At least 415 records · Page 23Linked to original sources

arXiv2021

Spatial-Temporal-Fusion BNN: Variational Bayesian Feature Layer

Bayesian neural networks (BNNs) have become a principal approach to alleviate overconfident predictions in deep learning, but they often suffer from scaling issues due to a large number of distribution parameters. In this paper, we discover that the first layer of a deep network possesses multiple disparate optima when solely retrained. This indicates a large posterior variance when the first layer is altered by a Bayesian layer, which motivates us to design a spatial-temporal-fusion BNN (STF-BNN) for efficiently scaling BNNs to large models: (1) first normally train a neural network from scratch to realize fast training; and (2) the first layer is converted to Bayesian and inferred by employing stochastic variational inference, while other layers are fixed. Compared to vanilla BNNs, our approach can greatly reduce the training time and the number of parameters, which contributes to scale BNNs efficiently. We further provide theoretical guarantees on the generalizability and the capability of mitigating overconfidence of STF-BNN. Comprehensive experiments demonstrate that STF-BNN (1) achieves the state-of-the-art performance on prediction and uncertainty quantification; (2) significantly improves adversarial robustness and privacy preservation; and (3) considerably reduces training time and memory costs.

cs.LG↗

arXiv2021

FedCV: A Federated Learning Framework for Diverse Computer Vision Tasks

Federated Learning (FL) is a distributed learning paradigm that can learn a global or personalized model from decentralized datasets on edge devices. However, in the computer vision domain, model performance in FL is far behind centralized training due to the lack of exploration in diverse tasks with a unified FL framework. FL has rarely been demonstrated effectively in advanced computer vision tasks such as object detection and image segmentation. To bridge the gap and facilitate the development of FL for computer vision tasks, in this work, we propose a federated learning library and benchmarking framework, named FedCV, to evaluate FL on the three most representative computer vision tasks: image classification, image segmentation, and object detection. We provide non-I.I.D. benchmarking datasets, models, and various reference FL algorithms. Our benchmark study suggests that there are multiple challenges that deserve future exploration: centralized training tricks may not be directly applied to FL; the non-I.I.D. dataset actually downgrades the model accuracy to some degree in different tasks; improving the system efficiency of federated training is challenging given the huge number of parameters and the per-client memory cost. We believe that such a library and benchmark, along with comparable evaluation settings, is necessary to make meaningful progress in FL on computer vision tasks. FedCV is publicly available: https://github.com/FedML-AI/FedCV.

cs.CV↗

arXiv2021

Off-policy Imitation Learning from Visual Inputs

Recently, various successful applications utilizing expert states in imitation learning (IL) have been witnessed. However, another IL setting -- IL from visual inputs (ILfVI), which has a greater promise to be applied in reality by utilizing online visual resources, suffers from low data-efficiency and poor performance resulted from an on-policy learning manner and high-dimensional visual inputs. We propose OPIfVI (Off-Policy Imitation from Visual Inputs), which is composed of an off-policy learning manner, data augmentation, and encoder techniques, to tackle the mentioned challenges, respectively. More specifically, to improve data-efficiency, OPIfVI conducts IL in an off-policy manner, with which sampled data can be used multiple times. In addition, we enhance the stability of OPIfVI with spectral normalization to mitigate the side-effect of off-policy training. The core factor, contributing to the poor performance of ILfVI, that we think is the agent could not extract meaningful features from visual inputs. Hence, OPIfVI employs data augmentation from computer vision to help train encoders that can better extract features from visual inputs. In addition, a specific structure of gradient backpropagation for the encoder is designed to stabilize the encoder training. At last, we demonstrate that OPIfVI is able to achieve expert-level performance and outperform existing baselines no matter visual demonstrations or visual observations are provided via extensive experiments using DeepMind Control Suite.

cs.LG↗

arXiv2021

Multidimensional representations in late-life depression: convergence in neuroimaging, cognition, clinical symptomatology and genetics

Late-life depression (LLD) is characterized by considerable heterogeneity in clinical manifestation. Unraveling such heterogeneity would aid in elucidating etiological mechanisms and pave the road to precision and individualized medicine. We sought to delineate, cross-sectionally and longitudinally, disease-related heterogeneity in LLD linked to neuroanatomy, cognitive functioning, clinical symptomatology, and genetic profiles. Multimodal data from a multicentre sample (N=996) were analyzed. A semi-supervised clustering method (HYDRA) was applied to regional grey matter (GM) brain volumes to derive dimensional representations. Two dimensions were identified, which accounted for the LLD-related heterogeneity in voxel-wise GM maps, white matter (WM) fractional anisotropy (FA), neurocognitive functioning, clinical phenotype, and genetics. Dimension one (Dim1) demonstrated relatively preserved brain anatomy without WM disruptions relative to healthy controls. In contrast, dimension two (Dim2) showed widespread brain atrophy and WM integrity disruptions, along with cognitive impairment and higher depression severity. Moreover, one de novo independent genetic variant (rs13120336) was significantly associated with Dim 1 but not with Dim 2. Notably, the two dimensions demonstrated significant SNP-based heritability of 18-27% within the general population (N=12,518 in UKBB). Lastly, in a subset of individuals having longitudinal measurements, Dim2 demonstrated a more rapid longitudinal decrease in GM and brain age, and was more likely to progress to Alzheimers disease, compared to Dim1 (N=1,413 participants and 7,225 scans from ADNI, BLSA, and BIOCARD datasets).

q-bio.NC↗

arXiv2021

Enabling variable high spatial resolution retrieval from a long pulse BOTDA sensor

In the field of Internet of Things, there is an urgent need for sensors with large-scale sensing capability for scenarios such as intelligent monitoring of production lines and urban infrastructure. Brillouin optical time domain analysis (BOTDA) sensors, which can monitor thousands of continuous points simultaneously, show great advantages in these applications. We propose a convolutional neural network (CNN) to process the data of conventional Brillouin optical time domain analysis (BOTDA) sensors, which achieves unprecedented performance improvement that allows to directly retrieve higher spatial resolution (SR) from the sensing system that use long pump pulses. By using the simulated Brillouin gain spectrums (BGSs) as the CNN input and the corresponding high SR BFS as the output target, the trained CNN is able to obtain a SR higher than the theoretical value determined by the pump pulse width. In the experiment, the CNN accurately retrieves 0.5-m hotspots from the measured BGS with pump pulses from 20 to 50 ns, and the acquired BFS is in great agreement with 45/40 ns differential pulse-width pair (DPP) measurement results. Compared with the DPP technique, the proposed CNN demonstrates a 2-fold improvement in BFS uncertainty with only half the measurement time. In addition, by changing the training datasets, the proposed CNN can obtain tunable high SR retrieval based on conventional BOTDA sensors that use long pulses without any requirement of hardware modifications. The proposed data post-processing approach paves the way to enable novel high spatial resolution BOTDA sensors, which brings substantial improvement over the state-of-the-art techniques in terms of system complexity, measurement time and reliability, etc.

eess.SP↗

arXiv2021

End-to-End Adaptive Monte Carlo Denoising and Super-Resolution

The classic Monte Carlo path tracing can achieve high quality rendering at the cost of heavy computation. Recent works make use of deep neural networks to accelerate this process, by improving either low-resolution or fewer-sample rendering with super-resolution or denoising neural networks in post-processing. However, denoising and super-resolution have only been considered separately in previous work. We show in this work that Monte Carlo path tracing can be further accelerated by joint super-resolution and denoising (SRD) in post-processing. This new type of joint filtering allows only a low-resolution and fewer-sample (thus noisy) image to be rendered by path tracing, which is then fed into a deep neural network to produce a high-resolution and clean image. The main contribution of this work is a new end-to-end network architecture, specifically designed for the SRD task. It contains two cascaded stages with shared components. We discover that denoising and super-resolution require very different receptive fields, a key insight that leads to the introduction of deformable convolution into the network design. Extensive experiments show that the proposed method outperforms previous methods and their variants adopted for the SRD task.

cs.CV↗

arXiv2021

UniFaceGAN: A Unified Framework for Temporally Consistent Facial Video Editing

Recent research has witnessed advances in facial image editing tasks including face swapping and face reenactment. However, these methods are confined to dealing with one specific task at a time. In addition, for video facial editing, previous methods either simply apply transformations frame by frame or utilize multiple frames in a concatenated or iterative fashion, which leads to noticeable visual flickers. In this paper, we propose a unified temporally consistent facial video editing framework termed UniFaceGAN. Based on a 3D reconstruction model and a simple yet efficient dynamic training sample selection mechanism, our framework is designed to handle face swapping and face reenactment simultaneously. To enforce the temporal consistency, a novel 3D temporal loss constraint is introduced based on the barycentric coordinate interpolation. Besides, we propose a region-aware conditional normalization layer to replace the traditional AdaIN or SPADE to synthesize more context-harmonious results. Compared with the state-of-the-art facial image editing methods, our framework generates video portraits that are more photo-realistic and temporally smooth.

cs.CV↗

arXiv2021

AlphaGAN: Fully Differentiable Architecture Search for Generative Adversarial Networks

Generative Adversarial Networks (GANs) are formulated as minimax game problems, whereby generators attempt to approach real data distributions by virtue of adversarial learning against discriminators. The intrinsic problem complexity poses the challenge to enhance the performance of generative networks. In this work, we aim to boost model learning from the perspective of network architectures, by incorporating recent progress on automated architecture search into GANs. To this end, we propose a fully differentiable search framework for generative adversarial networks, dubbed alphaGAN. The searching process is formalized as solving a bi-level minimax optimization problem, in which the outer-level objective aims for seeking a suitable network architecture towards pure Nash Equilibrium conditioned on the generator and the discriminator network parameters optimized with a traditional GAN loss in the inner level. The entire optimization performs a first-order method by alternately minimizing the two-level objective in a fully differentiable manner, enabling architecture search to be completed in an enormous search space. Extensive experiments on CIFAR-10 and STL-10 datasets show that our algorithm can obtain high-performing architectures only with 3-GPU hours on a single GPU in the search space comprised of approximate 2 ? 1011 possible configurations. We also provide a comprehensive analysis on the behavior of the searching process and the properties of searched architectures, which would benefit further research on architectures for generative models. Pretrained models and codes are available at https://github.com/yuesongtian/AlphaGAN.

cs.CV↗

arXiv2021

MIMO Transmission Under Discrete Input Signal Constraints

In this paper, we propose a multiple-input multipleoutput (MIMO) transmission strategy that is closer to the Shannon limit than the existing strategies. Different from most existing strategies which only consider uniformly distributed discrete input signals, we present a unified framework to optimize the MIMO precoder and the discrete input signal distribution jointly. First, a general model of MIMO transmission under discrete input signals and its equivalent formulation are established. Next, in order to maximize the mutual information between the input and output signals, we provide an algorithm that jointly optimizes the precoder and the input distribution. Finally, we compare our strategy with other existing strategies in the simulation. Numerical results indicate that our strategy narrows the gap between the mutual information and Shannon limit, and shows a lower frame error rate in simulation.

cs.IT↗

arXiv2021

Quantized Adam with Error Feedback

In this paper, we present a distributed variant of adaptive stochastic gradient method for training deep neural networks in the parameter-server model. To reduce the communication cost among the workers and server, we incorporate two types of quantization schemes, i.e., gradient quantization and weight quantization, into the proposed distributed Adam. Besides, to reduce the bias introduced by quantization operations, we propose an error-feedback technique to compensate for the quantized gradient. Theoretically, in the stochastic nonconvex setting, we show that the distributed adaptive gradient method with gradient quantization and error-feedback converges to the first-order stationary point, and that the distributed adaptive gradient method with weight quantization and error-feedback converges to the point related to the quantized level under both the single-worker and multi-worker modes. At last, we apply the proposed distributed adaptive gradient methods to train deep neural networks. Experimental results demonstrate the efficacy of our methods.

cs.LG↗

arXiv2021

Structure-Regularized Attention for Deformable Object Representation

Capturing contextual dependencies has proven useful to improve the representational power of deep neural networks. Recent approaches that focus on modeling global context, such as self-attention and non-local operation, achieve this goal by enabling unconstrained pairwise interactions between elements. In this work, we consider learning representations for deformable objects which can benefit from context exploitation by modeling the structural dependencies that the data intrinsically possesses. To this end, we provide a novel structure-regularized attention mechanism, which formalizes feature interaction as structural factorization through the use of a pair of light-weight operations. The instantiated building blocks can be directly incorporated into modern convolutional neural networks, to boost the representational power in an efficient manner. Comprehensive studies on multiple tasks and empirical comparisons with modern attention mechanisms demonstrate the gains brought by our method in terms of both performance and model complexity. We further investigate its effect on feature representations, showing that our trained models can capture diversified representations characterizing object parts without resorting to extra supervision.

cs.CV↗

arXiv2021

Attacking Adversarial Attacks as A Defense

It is well known that adversarial attacks can fool deep neural networks with imperceptible perturbations. Although adversarial training significantly improves model robustness, failure cases of defense still broadly exist. In this work, we find that the adversarial attacks can also be vulnerable to small perturbations. Namely, on adversarially-trained models, perturbing adversarial examples with a small random noise may invalidate their misled predictions. After carefully examining state-of-the-art attacks of various kinds, we find that all these attacks have this deficiency to different extents. Enlightened by this finding, we propose to counter attacks by crafting more effective defensive perturbations. Our defensive perturbations leverage the advantage that adversarial training endows the ground-truth class with smaller local Lipschitzness. By simultaneously attacking all the classes, the misled predictions with larger Lipschitzness can be flipped into correct ones. We verify our defensive perturbation with both empirical experiments and theoretical analyses on a linear model. On CIFAR10, it boosts the state-of-the-art model from 66.16% to 72.66% against the four attacks of AutoAttack, including 71.76% to 83.30% against the Square attack. On ImageNet, the top-1 robust accuracy of FastAT is improved from 33.18% to 38.54% under the 100-step PGD attack.

cs.LG↗

arXiv2021

Robust Registration of Multimodal Remote Sensing Images Based on Structural Similarity

Automatic registration of multimodal remote sensing data (e.g., optical, LiDAR, SAR) is a challenging task due to the significant non-linear radiometric differences between these data. To address this problem, this paper proposes a novel feature descriptor named the Histogram of Orientated Phase Congruency (HOPC), which is based on the structural properties of images. Furthermore, a similarity metric named HOPCncc is defined, which uses the normalized correlation coefficient (NCC) of the HOPC descriptors for multimodal registration. In the definition of the proposed similarity metric, we first extend the phase congruency model to generate its orientation representation, and use the extended model to build HOPCncc. Then a fast template matching scheme for this metric is designed to detect the control points between images. The proposed HOPCncc aims to capture the structural similarity between images, and has been tested with a variety of optical, LiDAR, SAR and map data. The results show that HOPCncc is robust against complex non-linear radiometric differences and outperforms the state-of-the-art similarities metrics (i.e., NCC and mutual information) in matching performance. Moreover, a robust registration method is also proposed in this paper based on HOPCncc, which is evaluated using six pairs of multimodal remote sensing images. The experimental results demonstrate the effectiveness of the proposed method for multimodal image registration.

cs.CV↗

arXiv2021

Functional Connectome Fingerprint Gradients in Young Adults

The assessment of brain fingerprints has emerged in the recent years as an important tool to study individual differences and to infer quality of neuroimaging datasets. Studies so far have mainly focused on connectivity fingerprints between different brain scans of the same individual. Here, we extend the concept of brain connectivity fingerprints beyond test/retest and assess fingerprint gradients in young adults by developing an extension of the differential identifiability framework. To do so, we look at the similarity between not only the multiple scans of an individual (subject fingerprint), but also between the scans of monozygotic and dizygotic twins (twin fingerprint). We have carried out this analysis on the 8 fMRI conditions present in the Human Connectome Project -- Young Adult dataset, which we processed into functional connectomes (FCs) and timeseries parcellated according to the Schaefer Atlas scheme, which has multiple levels of resolution. Our differential identifiability results show that the fingerprint gradients based on genetic and environmental similarities are indeed present when comparing FCs for all parcellations and fMRI conditions. Importantly, only when assessing optimally reconstructed FCs, we fully uncover fingerprints present in higher resolution atlases. We also study the effect of scanning length on subject fingerprint of resting-state FCs to analyze the effect of scanning length and parcellation. In the pursuit of open science, we have also made available the processed and parcellated FCs and timeseries for all conditions for ~1200 subjects part of the HCP-YA dataset to the scientific community.

q-bio.NC↗

arXiv2020

Adaptive Compact Attention For Few-shot Video-to-video Translation

This paper proposes an adaptive compact attention model for few-shot video-to-video translation. Existing works in this domain only use features from pixel-wise attention without considering the correlations among multiple reference images, which leads to heavy computation but limited performance. Therefore, we introduce a novel adaptive compact attention mechanism to efficiently extract contextual features jointly from multiple reference images, of which encoded view-dependent and motion-dependent information can significantly benefit the synthesis of realistic videos. Our core idea is to extract compact basis sets from all the reference images as higher-level representations. To further improve the reliability, in the inference phase, we also propose a novel method based on the Delaunay Triangulation algorithm to automatically select the resourceful references according to the input label. We extensively evaluate our method on a large-scale talking-head video dataset and a human dancing dataset; the experimental results show the superior performance of our method for producing photorealistic and temporally consistent videos, and considerable improvements over the state-of-the-art method.

cs.CV↗

arXiv2020

Cognitive Biomarker Prioritization in Alzheimer's Disease using Brain Morphometric Data

Background:Cognitive assessments represent the most common clinical routine for the diagnosis of Alzheimer's Disease (AD). Given a large number of cognitive assessment tools and time-limited office visits, it is important to determine a proper set of cognitive tests for different subjects. Most current studies create guidelines of cognitive test selection for a targeted population, but they are not customized for each individual subject. In this manuscript, we develop a machine learning paradigm enabling personalized cognitive assessments prioritization. Method: We adapt a newly developed learning-to-rank approach PLTR to implement our paradigm. This method learns the latent scoring function that pushes the most effective cognitive assessments onto the top of the prioritization list. We also extend PLTR to better separate the most effective cognitive assessments and the less effective ones. Results: Our empirical study on the ADNI data shows that the proposed paradigm outperforms the state-of-the-art baselines on identifying and prioritizing individual-specific cognitive biomarkers. We conduct experiments in cross validation and level-out validation settings. In the two settings, our paradigm significantly outperforms the best baselines with improvement as much as 22.1% and 19.7%, respectively, on prioritizing cognitive features. Conclusions: The proposed paradigm achieves superior performance on prioritizing cognitive biomarkers. The cognitive biomarkers prioritized on top have great potentials to facilitate personalized diagnosis, disease subtyping, and ultimately precision medicine in AD.

q-bio.QM↗

arXiv2020

FedML: A Research Library and Benchmark for Federated Machine Learning

Federated learning (FL) is a rapidly growing research field in machine learning. However, existing FL libraries cannot adequately support diverse algorithmic development; inconsistent dataset and model usage make fair algorithm comparison challenging. In this work, we introduce FedML, an open research library and benchmark to facilitate FL algorithm development and fair performance comparison. FedML supports three computing paradigms: on-device training for edge devices, distributed computing, and single-machine simulation. FedML also promotes diverse algorithmic research with flexible and generic API design and comprehensive reference baseline implementations (optimizer, models, and datasets). We hope FedML could provide an efficient and reproducible means for developing and evaluating FL algorithms that would benefit the FL research community. We maintain the source code, documents, and user community at https://fedml.ai.

cs.LG↗

arXiv2020

A Distributed Training Algorithm of Generative Adversarial Networks with Quantized Gradients

Training generative adversarial networks (GAN) in a distributed fashion is a promising technology since it is contributed to training GAN on a massive of data efficiently in real-world applications. However, GAN is known to be difficult to train by SGD-type methods (may fail to converge) and the distributed SGD-type methods may also suffer from massive amount of communication cost. In this paper, we propose a {distributed GANs training algorithm with quantized gradient, dubbed DQGAN,} which is the first distributed training method with quantized gradient for GANs. The new method trains GANs based on a specific single machine algorithm called Optimistic Mirror Descent (OMD) algorithm, and is applicable to any gradient compression method that satisfies a general $δ$-approximate compressor. The error-feedback operation we designed is used to compensate for the bias caused by the compression, and moreover, ensure the convergence of the new method. Theoretically, we establish the non-asymptotic convergence of {DQGAN} algorithm to first-order stationary point, which shows that the proposed algorithm can achieve a linear speedup in the parameter server model. Empirically, our experiments show that our {DQGAN} algorithm can reduce the communication cost and save the training time with slight performance degradation on both synthetic and real datasets.

cs.LG↗