Search

Wearable Artificial Intelligence-of-Things (AIoT) devices exhibit the need to be resource and energy-efficient. In this paper, we introduced a quantized multilayer perceptron (qMLP) for converting ECG signals to binary image, which can be combined with binary convolutional neural network (bCNN) for classification. We deploy our model into a low-power and low-resource field programmable gate array (FPGA) fabric. The model requires 5.8× lesser multiply and accumulate (MAC) operations than known wearable CNN models. Our model also achieves a classification accuracy of 98.5%, sensitivity of 85.4%, specificity of 99.5%, precision of 93.3%, and F1-score of 89.2%, along with dynamic power dissipation of 34.9 μW. [ABSTRACT FROM AUTHOR]

Meta-learning offers an effective solution to learn new concepts under scarce supervision through an episodic-training scheme: a series of target-like tasks sampled from base classes are sequentially fed into a meta-learner to extract cross-task knowledge, which can facilitate the quick acquisition of task-specific knowledge of the target task with few samples. Despite its noticeable improvements, the episodic-training strategy samples tasks randomly and uniformly, without considering their hardness and quality, which may not progressively improve the meta-leaner’s generalization. In this paper, we propose Progressive Meta-learning using tasks from easy to hard. First, based on a predefined curriculum, we develop a Curriculum-Based Meta-learning (CubMeta) method. CubMeta is in a stepwise manner, and in each step, we design a BrotherNet module to establish harder tasks and an effective learning scheme for obtaining an ensemble of stronger meta-learners. Then we move a step further to propose an end-to-end Self-Paced Meta-learning (SepMeta) method. The curriculum in SepMeta is effectively integrated as a regularization term into the objective so that the meta-learner can measure the hardness of tasks adaptively, according to what the model has already learned. Extensive experiments on benchmark datasets demonstrate the effectiveness of the proposed methods. Our code is available at https://github.com/nobody-777. [ABSTRACT FROM AUTHOR]

Driving is a highly dynamic task requiring extensive and fast processing of visuospatial information in brief time intervals, while experienced drivers can efficiently perform this task. Therefore, investigating the human-like visual fixation behavior shows much potential in developing the advanced driving assistance systems and efficient autonomous vehicles. Literature shows that driver’s visual fixation behavior is guided by a combination of top-down and bottom-up mechanisms. Currently, existing excellent computational frameworks rely heavily on the deep neural network’s strong feature extraction ability, which is inherently task-unaware. These black-box models are limited in strengthening the guidance of task-aware top-down factors in governing the fixation behavior. Furthermore, using the black-box model to fit the training data directly decreases its interpretability. In this paper, we encourage to adopt the white-box method. By researching the findings in the intersection of neuroscience and psychology, we propose to identify two white-box representations for driving separately: the information selection representation including salience, effort, expectancy, and value, and the driving task-aware representation highlighting sub-task and speed. We present a novel computational framework to incorporate all the above factors to predict driver’s visual fixation systematically. On the BDDA dataset, we have conducted the extensive quantitative evaluation and qualitative analysis to validate the effectiveness. Its performance surpasses the previous state-of-the-art by +11.3% in the Pearson’s Correlation Coefficient, especially by a large margin of +24.5% and +28.8% in the turning operations. The ablation study further demonstrates the influence of each component in our framework. Moreover, the trained model has been deployed on the DR(eye)VE dataset to verify the generalization ability, demonstrating the potential of domain adaptation. [ABSTRACT FROM AUTHOR]

Deep Neural Networks (DNNs) often rely on vast datasets for training. Given the large size of such datasets, it is conceivable that they contain specific samples that either do not contribute or negatively impact the DNN’s optimization. Modifying the training distribution to exclude such samples could provide an effective solution to improve performance and reduce training time. This paper proposes to scale up ensemble Active Learning (AL) methods to perform acquisition at a large scale (10k to 500k samples at a time). We do this with ensembles of hundreds of models, obtained at a minimal computational cost by reusing intermediate training checkpoints. This allows us to automatically and efficiently perform a training data subset search for large labeled datasets. We observe that our approach obtains favorable subsets of training data, which can be used to train more accurate DNNs than training with the entire dataset. We perform an extensive experimental study of this phenomenon on three image classification benchmarks (CIFAR-10, CIFAR-100, and ImageNet), as well as an internal object detection benchmark for prototyping perception models for autonomous driving. Unlike existing studies, our experiments on object detection are at the scale required for production-ready autonomous driving systems. We provide insights on the impact of different initialization schemes, acquisition functions, and ensemble configurations at this scale. Our results provide strong empirical evidence that optimizing the training data distribution can significantly benefit large-scale vision tasks. [ABSTRACT FROM AUTHOR]

Traffic flow modeling for traffic state estimation is a vital component in many traffic management and operation systems. To leverage both machine learning (ML) methods and classical traffic flow models, the previous study has developed a hybrid framework for encoding traffic flow into multivariant Gaussian Process. However, the computational efficiency is low due to multiple inputs, outputs and equations. To improve the efficiency of the previous method, this paper presents a new modeling framework, named gradual physics regularized learning, to incrementally encode complex traffic flow models into the ML process. More specifically, the method starts with the involvement of traffic flow models from the lower-order version, such as the fundamental diagram and the kinetic wave models. Then the learned parameters and hyperparameters can be further fine-tuned with the high-order models. A field test based on real-world freeway measurements indicates the proposed model can leverage the additional physical equations to achieve better performance in estimation accuracy and robustness. Meanwhile, the gradual learning method can significantly reduce the computational efforts and further enables its application to scenarios with either larger datasets or more complex traffic flow models. [ABSTRACT FROM AUTHOR]

Accurate short-term traffic flow prediction plays an important role in the field of modern Intelligent Transportation Systems. Since various uncontrollable factors (e.g. weather, traffic jams or accidents), collected traffic data inevitably contain outliers. This makes it challenge to achieve satisfactory results for traffic flow prediction. Least Squares Twin Support Vector Regression (LSTSVR) has been shown to provide a powerful potential in nonlinear prediction problems. This is especially true when using appropriate heuristic algorithms to determine the parameters of nonlinear LSTSVR. In view of this, a novel LSTSVR model based on the robust $\text{L}_{2,\mathrm {p}}$ -norm ($0< p\le 2$) distance is proposed to alleviate the negative effect of traffic data with outliers, called PLSTSVR. An iterative algorithm is designed to solve the optimization problem of PLSTSVR, which has great potential for solving other relevant optimization problems. To search the parameters of constructed PLSTSVR, this paper constructs two traffic flow prediction models based on PLSTSVR and heuristic algorithms (Fruit Fly Optimization Algorithm and Particle Swarm Optimization), called PLSTSVR-FOA and PLSTSVR-PSO. Extensive experiments demonstrate that the constructed models are more effective and robust than other competing models in various experimental settings. [ABSTRACT FROM AUTHOR]

Few-shot learning under the $N$ N -way $K$ K -shot setting (i.e., $K$ K annotated samples for each of $N$ N classes) has been widely studied in relation extraction (e.g., FewRel) and image classification (e.g., Mini-ImageNet). Named entity recognition (NER) is typically framed as a sequence labeling problem where the entity classes are inherently entangled together because the entity number and classes in a sentence are not known in advance, leaving the $N$ N -way $K$ K -shot NER problem so far unexplored. In this paper, we first formally define a more suitable $N$ N -way $K$ K -shot setting for NER. Then we propose FewNER, a novel meta-learning approach for few-shot NER. FewNER separates the entire network into a task-independent part and a task-specific part. During training in FewNER, the task-independent part is meta-learned across multiple tasks and the task-specific part is learned for each individual task in a low-dimensional space. At test time, FewNER keeps the task-independent part fixed and adapts to a new task via gradient descent by updating only the task-specific part, resulting in it being less prone to overfitting and more computationally efficient. Compared with pre-trained language models (e.g., BERT and ELMo) which obtain the transferability in an implicit manner (i.e., relying on large-scale corpora), FewNER explicitly optimizes the capability of “learning to adapt quickly” through meta-learning. The results demonstrate that FewNER achieves state-of-the-art performance against nine baseline methods by significant margins on three adaptation experiments (i.e., intra-domain cross-type, cross-domain intra-type and cross-domain cross-type). [ABSTRACT FROM AUTHOR]

Wide models such as generalized linear models and factorization-based models have been extensively used in various predictive applications, e.g., recommendation, CTR prediction, and image recognition. Due to the memory bounded property of the models, the performance improvement on CPU is reaching the limitation. GPU is known to have many computation units and high memory bandwidth, and becomes a promising platform for training machine learning models. However, the GPU training for the wide models is far from optimal due to the sparsity and irregularity in wide models. The existing GPU-based wide models are even slower than the ones using CPU. The classical training schema of the wide models does not optimized for the GPU architecture, which suffers from large amount of random memory accesses and redundant read/write of intermediate values. In this paper, we propose an efficient GPU-training framework for the large-scale wide models, named cuWide. To fully benefit from the memory hierarchy of GPU, cuWide applies a new flow-based schema for training, which leverages the spatial and temporal locality of wide models to drastically reduce the amount of communication with GPU global memory. To do so, we adopt a bigraph computation model to efficiently realize the flow-based schema and exploit three flexible interfaces for programming. Further, we use the 2D partition of mini-batch (in sample and feature dimensions) with proposed graph abstraction to optimize GPU memory access for sparse data, and apply several spatial-temporal caching mechanisms (importance-based model caching and cross-stage accumulation caching mechanisms) to achieve a high performance kernel. To efficiently implement cuWide, we also propose several GPU-oriented optimizations, including feature-oriented data layout to enhance the data locality, replication mechanism to reduce update conflicts in shared memory, and multi-stream scheduling to overlap data transferring and kernel computing. We show that cuWide can be up to more than 20× faster than the state-of-the-art GPU solutions and multi-core CPU solutions. [ABSTRACT FROM AUTHOR]

The conventional federated learning (FL) framework usually assumes synchronous reception and fusion of all the local models at the central aggregator and synchronous updating and training of the global model at all the agents as well. However, in a wireless network, due to limited radio resource, inevitable transmission failures and heterogeneous computing capacity, it is very hard to realize strict synchronization among all the involved user equipments (UEs). In this paper, we propose a novel asynchronous FL framework, which well adapts to the heterogeneity of users, communication environments and learning tasks, by considering both the possible delays in training and uploading the local models and the resultant staleness among the received models that has heavy impact on the global model fusion. A novel centralized fusion algorithm is designed to determine the fusion weight during the global update, which aims to make full use of the fresh information contained in the uploaded local models while avoiding the biased convergence by enforcing the impact of each UE’s local dataset to be proportional to its sample share. Numerical experiments validate that the proposed asynchronous FL framework can achieve fast and smooth convergence and enhance the training efficiency significantly. [ABSTRACT FROM AUTHOR]

The paper presents several novel residual recurrent neural network (RNN) models for digital predistortion of radio-frequency power amplifiers. Then, unstructured and structured neural network (NN) pruning algorithms are proposed to reduce the per-sample computational complexity and the required memory, without compromising their modeling performance. The mathematical equations of the complexity and the memory are derived to analyze the hardware cost and design low cost models. An experimental platform is developed to measure the modeling performance in terms of adjacent channel leakage ratio (ACLR), normalized mean square error, error vector magnitude (EVM) and bit error rate. The platform uses an envelope tracking PA excited by 80 MHz IEEE 802.11ac signals. Compared with the state-of-the-art RNN model in (Sun et al., 2019), the proposed models reduce the per-sample complexity and required memory by $94\hbox{--}99$ %, they also improve the ACLR and EVM by at least 3 dB and 1 dB at 0.45 dB and 3.15 dB gain compression, respectively. Furthermore, an observation path with time alignment is used to remove the transmit latency in the state-of-the-art approach. At 3.15 dB gain compression, the proposed models outperform the state-of-the-art feedforward and convolutional NN models by at least 1.3 dB in EVM, outperform the general memory polynomial model by at least 3.3 dB in EVM. The above evidences indicate that the proposed RNN models can be implemented with relatively low hardware cost and achieve the best EVM in high power efficiency scenario. [ABSTRACT FROM AUTHOR]

Federated Learning (FL) has been recognized as a privacy-preserving machine learning (ML) technology that enables collaborative training and learning of a global ML model based on the aggregation of distributed local model updates. However, security and privacy guarantees could be compromised due to malicious participants and the centralized aggregation manner. Possessing attractive features like decentralization, immutability and auditability, Blockchain is promising to enable a tamper-proof and trust-free framework to enhance performance and security in IoT based FL systems. However, directly integrating blockchains into the large scale IoT-based FL scenarios still faces many limitations, such as high computation and storage demands, low transactions throughput, poor scalability and challenges in privacy preservation. This paper proposes $\mu $ DFL, a novel hierarchical IoT network fabric for decentralized federated learning (DFL) atop of a lightweight blockchain called microchain. Following the hierarchical infrastructure of FL, participants in $\mu $ DFL are fragmented into multiple small scale microchains. Each microchain network relies on a hybrid Proof of Credit (PoC) block generation and Voting-based Chain Finality (VCF) consensus protocol to ensure efficiency and privacy-preservation at the network of edge. Meanwhile, microchains are federated vie a high-level inter-chain network, which adopts an efficient Byzantine Fault Tolerance (BFT) consensus protocol to achieve scalability and security. A proof-of-concept prototype is implemented, and the experimental results verify the feasibility of the proposed $\mu $ DFL solution in cross-devices FL settings with efficiency, security and privacy guarantees. [ABSTRACT FROM AUTHOR]

We study network pruning which aims to remove redundant channels/kernels and hence speed up the inference of deep networks. Existing pruning methods either train from scratch with sparsity constraints or minimize the reconstruction error between the feature maps of the pre-trained models and the compressed ones. Both strategies suffer from some limitations: the former kind is computationally expensive and difficult to converge, while the latter kind optimizes the reconstruction error but ignores the discriminative power of channels. In this paper, we propose a simple-yet-effective method called discrimination-aware channel pruning (DCP) to choose the channels that actually contribute to the discriminative power. To this end, we first introduce additional discrimination-aware losses into the network to increase the discriminative power of the intermediate layers. Next, we select the most discriminative channels for each layer by considering the discrimination-aware loss and the reconstruction error, simultaneously. We then formulate channel pruning as a sparsity-inducing optimization problem with a convex objective and propose a greedy algorithm to solve the resultant problem. Note that a channel (3D tensor) often consists of a set of kernels (each with a 2D matrix). Besides the redundancy in channels, some kernels in a channel may also be redundant and fail to contribute to the discriminative power of the network, resulting in kernel level redundancy. To solve this issue, we propose a discrimination-aware kernel pruning (DKP) method to further compress deep networks by removing redundant kernels. To avoid manually determining the pruning rate for each layer, we propose two adaptive stopping conditions to automatically determine the number of selected channels/kernels. The proposed adaptive stopping conditions tend to yield more efficient models with better performance in practice. Extensive experiments on both image classification and face recognition demonstrate the effectiveness of our methods. For example, on ILSVRC-12, the resultant ResNet-50 model with 30 percent reduction of channels even outperforms the baseline model by 0.36 percent in terms of Top-1 accuracy. We also deploy the pruned models on a smartphone (equipped with a Qualcomm Snapdragon 845 processor). The pruned MobileNetV1 and MobileNetV2 achieve 1.93× and 1.42× inference acceleration on the mobile device, respectively, with negligible performance degradation. The source code and the pre-trained models are available at https://github.com/SCUT-AILab/DCP. [ABSTRACT FROM AUTHOR]

Remarkable achievements have been obtained by deep neural networks in the last several years. However, the breakthrough in neural networks accuracy is always accompanied by explosive growth of computation and parameters, which leads to a severe limitation of model deployment. In this paper, we propose a novel knowledge distillation technique named self-distillation to address this problem. Self-distillation attaches several attention modules and shallow classifiers at different depths of neural networks and distills knowledge from the deepest classifier to the shallower classifiers. Different from the conventional knowledge distillation methods where the knowledge of the teacher model is transferred to another student model, self-distillation can be considered as knowledge transfer in the same model - from the deeper layers to the shallow layers. Moreover, the additional classifiers in self-distillation allow the neural network to work in a dynamic manner, which leads to a much higher acceleration. Experiments demonstrate that self-distillation has consistent and significant effectiveness on various neural networks and datasets. On average, 3.49 and 2.32 percent accuracy boost are observed on CIFAR100 and ImageNet. Besides, experiments show that self-distillation can be combined with other model compression methods, including knowledge distillation, pruning and lightweight model design. [ABSTRACT FROM AUTHOR]

Linear estimation of signals is often based on covariance matrices estimated from training, which can perform poorly if the training data are limited and the estimated covariance matrices are ill-conditioned. Shrinking the covariance matrix toward a scaled identity matrix can improve the robustness against the model uncertainty provided the shrinkage factor is appropriately chosen. This paper introduces several cross-validation schemes for choosing the shrinkage factors in applications where the covariance matrices are replaced with sample covariance matrices or constructed from least squares estimates of the linear model parameters. For cases where the training and out-of-training data are identically distributed, we derive leave-one-out cross-validation (LOOCV) schemes that repeatedly split the training data with respect to time to determine the optimal shrinkage factors for either model fitting or signal estimation. For cases where they are distributed differently, we develop alternative LOOCV schemes that repeatedly split the out-of-training observations with respect to space. We derive computationally efficient implementations of those schemes and provide examples to demonstrate their performance. [ABSTRACT FROM AUTHOR]

In recent years there has been a noticeable increase in the number of available Big Data infrastructures. This fact has promoted the adaptation of traditional machine learning techniques to be capable of addressing large scale problems in distributed environments. Kernel methods like support vector machines (SVMs) suffer from scalability problems due to their nonparametric nature and the complexity of their training procedures. In this paper, we propose a new and efficient distributed implementation of a training procedure for nonlinear semiparametric (budgeted) SVMs called distributed iterative reweighted least squares (IRWLS). This algorithm uses $ {k}$ -means to select the centroids of the semiparametric model and a new distributed algorithmic implementation of the IRWLS optimization procedure to find the weights of the model. We have implemented the proposed algorithm in Apache Spark and we have benchmarked it against other state-of-the-art methods, either full SVM ($ {p}$ -pack SVM) or budgeted (budgeted stochastic gradient descent). Experimental results show that the proposed algorithm achieves higher accuracy while controlling the size of the final model, and also offers high performance in terms of run time and efficiency, when processing very large datasets (the computation time grows linear with the number of training patterns). [ABSTRACT FROM AUTHOR]

This study proposes a simple but strong baseline for deep person re-identification (ReID). Deep person ReID has achieved great progress and high performance in recent years. However, many state-of-the-art methods design complex network structures and concatenate multi-branch features. In the literature, some effective training tricks briefly appear in several papers or source codes. The present study collects and evaluates these effective training tricks in person ReID. By combining these tricks, the model achieves 94.5% rank-1 and 85.9% mean average precision on Market1501 with only using the global features of ResNet50. The performance surpasses all existing global- and part-based baselines in person ReID. We propose a novel neck structure named as batch normalization neck (BNNeck). BNNeck adds a batch normalization layer after global pooling layer to separate metric and classification losses into two different feature spaces because we observe they are inconsistent in one embedding space. Extended experiments show that BNNeck can boost the baseline, and our baseline can improve the performance of existing state-of-the-art methods. Our codes and models are available at: https://github.com/michuanhaohao/reid-strong-baseline [ABSTRACT FROM AUTHOR]

Outlier detection is an important topic in machine learning and has been used in a wide range of applications. In this paper, we approach outlier detection as a binary-classification issue by sampling potential outliers from a uniform reference distribution. However, due to the sparsity of data in high-dimensional space, a limited number of potential outliers may fail to provide sufficient information to assist the classifier in describing a boundary that can separate outliers from normal data effectively. To address this, we propose a novel Single-Objective Generative Adversarial Active Learning (SO-GAAL) method for outlier detection, which can directly generate informative potential outliers based on the mini-max game between a generator and a discriminator. Moreover, to prevent the generator from falling into the mode collapsing problem, the stop node of training should be determined when SO-GAAL is able to provide sufficient information. But without any prior information, it is extremely difficult for SO-GAAL. Therefore, we expand the network structure of SO-GAAL from a single generator to multiple generators with different objectives (MO-GAAL), which can generate a reasonable reference distribution for the whole dataset. We empirically compare the proposed approach with several state-of-the-art outlier detection methods on both synthetic and real-world datasets. The results show that MO-GAAL outperforms its competitors in the majority of cases, especially for datasets with various cluster types or high irrelevant variable ratio. The experiment codes are available at: https://github.com/leibinghe/GAAL-based-outlier-detection. [ABSTRACT FROM AUTHOR]

Emergency responders are typically notified immediately after a major earthquake strikes. However, a time delay, called the travel time, is usually experienced between the notification and the arrival of the responders on the scene. The reparative work necessary after the responders arrive takes an additional amount of time, called the response time, depending on the nature of the damage and the volume of resources available. In a smart city, the restoration time, which is the sum of the travel and response times, should be minimized. A new discrete-event-based simulation (DEBS) model is presented in this paper to estimate the restoration time needed to bring the situation under control after notifying the response center. The DEBS model not only relaxes restrictive assumptions on travel time made by the Markov chain models from the existing literature, but it can also quantify the impact of resource volumes on restoration times. Additionally, the DEBS model is very useful for training purposes. The DEBS model was employed on a case study from the state of Missouri (U.S.). The experiments demonstrate that numerical results with the model take a short amount of computational time and that it can be implemented on a real-time basis in a smart-city infrastructure. [ABSTRACT FROM AUTHOR]

Researchers have devoted great efforts to image dehazing with prior assumptions in the past decade. Recently developed example-based approaches typically lack elegant models for the hazy process and meanwhile demand synthetic hazy images by manual selection. The priors from observations, and those trained from synthetic images cannot always reflect true structural information of natural images in practice. In this paper, we present a learning model for haze removal by using two-layer Gaussian process regression (GPR). By using training examples, the two-layer GPR establishes a direct relationship from the input image to the depth-dependent transmission, and learns local image priors to further improve the estimation. We also provide a systematic scheme to automatically collect suitable training pairs, which works for both simulated examples and images of natural scenes. Both qualitative and quantitative comparisons on real-world and synthetic hazy images demonstrate the effectiveness of the proposed approach, especially for white or bright objects and heavy haze regions in which traditional methods may fail. [ABSTRACT FROM PUBLISHER]

Machine-learning algorithms have shown outstanding image recognition/classification performance for computer vision applications. However, the compute and energy requirement for implementing such classifier models for large-scale problems is quite high. In this paper, we propose feature driven selective classification (FALCON) inspired by the biological visual attention mechanism in the brain to optimize the energy-efficiency of machine-learning classifiers. We use the consensus in the characteristic features (color/texture) across images in a dataset to decompose the original classification problem and construct a tree of classifiers (nodes) with a generic-to-specific transition in the classification hierarchy. The initial nodes of the tree separate the instances based on feature information and selectively enable the latter nodes to perform object specific classification. The proposed methodology allows selective activation of only those branches and nodes of the classification tree that are relevant to the input while keeping the remaining nodes idle. Additionally, we propose a programmable and scalable neuromorphic engine (NeuE) that utilizes arrays of specialized neural computational elements to execute the FALCON-based classifier models for diverse datasets. The structure of FALCON facilitates the reuse of nodes while scaling up from small classification problems to larger ones thus allowing us to construct classifier implementations that are significantly more efficient. We evaluate our approach for a 12-object classification task on the Caltech101 dataset and ten-object task on CIFAR-10 dataset by constructing FALCON models on the NeuE platform in 45-nm technology. Our results demonstrate up to $3.66\boldsymbol \times $ improvement in energy-efficiency for no loss in output quality, and even higher improvements of up to $5.91\boldsymbol \times $ with 3.9% accuracy loss compared to an optimized baseline network. In addition, FALCON shows an improvement in training time of up to $1.96\boldsymbol \times $ as compared to the traditional classification approach. [ABSTRACT FROM AUTHOR]

The use of local features for image representation has become popular in recent years. Local features are often used in the bag-of-visual-words scheme. Although proven effective, this method still has two drawbacks. First, local regions from which local features are extracted are not discriminative enough for visual tasks. Hence, the combination of local features is necessary. Second, the semantic gap between visual features and human perception also hinders the performance. To address these two problems, in this paper, we propose a novel contextual exemplar classifier-based method for image representation and apply it for classification tasks. Each exemplar classifier is trained to separate one training image from the other images of different classes. We partition each image into a number of regions and use the responses of these exemplar classifiers as the image region’s representation. The contextual relationship is then modeled using mixture Dirichlet distributions. A bilayer model is used to predict image classes with L2 constraints. Experimental results on the Natural Scene, Caltech-101/256, Flower-17/102, and SUN-397 data sets show that the proposed method is able to outperform the state-of-the-art local feature-based methods for image classification. [ABSTRACT FROM PUBLISHER]

Domain adaptation is to transfer the shared knowledge learned from the source domain to a new environment, i.e., target domain. One common practice is to train the model on both labeled source-domain data and unlabeled target-domain data. Yet the learned models are usually biased due to the strong supervision of the source domain. Most researchers adopt the early-stopping strategy to prevent over-fitting, but when to stop training remains a challenging problem since the lack of the target-domain validation set. In this paper, we propose one efficient bootstrapping method, called Adaboost Student, explicitly learning complementary models during training and liberating users from empirical early stopping. Adaboost Student combines deep model learning with the conventional training strategy, i.e., adaptive boosting, and enables interactions between learned models and the data sampler. We adopt one adaptive data sampler to progressively facilitate learning on hard samples and aggregate “weak” models to prevent over-fitting. Extensive experiments show that (1) Without the need to worry about the stopping time, AdaBoost Student provides one robust solution by efficient complementary model learning during training. (2) AdaBoost Student is orthogonal to most domain adaptation methods, which can be combined with existing approaches to further improve the state-of-the-art performance. We have achieved competitive results on three widely-used scene segmentation domain adaptation benchmarks. [ABSTRACT FROM AUTHOR]

We consider the problem of obtaining image quality representations in a self-supervised manner. We use prediction of distortion type and degree as an auxiliary task to learn features from an unlabeled image dataset containing a mixture of synthetic and realistic distortions. We then train a deep Convolutional Neural Network (CNN) using a contrastive pairwise objective to solve the auxiliary problem. We refer to the proposed training framework and resulting deep IQA model as the CONTRastive Image QUality Evaluator (CONTRIQUE). During evaluation, the CNN weights are frozen and a linear regressor maps the learned representations to quality scores in a No-Reference (NR) setting. We show through extensive experiments that CONTRIQUE achieves competitive performance when compared to state-of-the-art NR image quality models, even without any additional fine-tuning of the CNN backbone. The learned representations are highly robust and generalize well across images afflicted by either synthetic or authentic distortions. Our results suggest that powerful quality representations with perceptual relevance can be obtained without requiring large labeled subjective image quality datasets. The implementations used in this paper are available at https://github.com/pavancm/CONTRIQUE. [ABSTRACT FROM AUTHOR]

Most video-based person re-identification (re-id) methods only focus on appearance features but neglect motion features. In fact, motion features can help to distinguish the target persons that are hard to be identified only by appearance features. However, most existing temporal information modeling methods cannot extract motion features effectively or efficiently for v ideo-based re-id. In this paper, we propose a more efficient Motion Feature Aggregation (MFA) method to model and aggregate motion information in the feature map level for video-based re-id. The proposed MFA consists of (i) a coarse-grained motion learning module, which extracts coarse-grained motion features based on the position changes of body parts over time, and (ii) a fine-grained motion learning module, which extracts fine-grained motion features based on the appearance changes of body parts over time. These two modules can model motion information from different granularities and are complementary to each other. It is easy to combine the proposed method with existing network architectures for end-to-end training. Extensive experiments on four widely used datasets demonstrate that the motion features extracted by MFA are crucial complements to appearance features for video-based re-id, especially for the scenario with large appearance changes. Besides, the results on LS-VID, the current largest publicly available video-based re-id dataset, surpass the state-of-the-art methods by a large margin. The code is available at: https://github.com/guxinqian/Simple-ReID. [ABSTRACT FROM AUTHOR]

Recent studies have pointed out that many well-developed Visual Question Answering (VQA) models are heavily affected by the language prior problem. It refers to making predictions based on the co-occurrence pattern between textual questions and answers instead of reasoning upon visual contents. To tackle this problem, most existing methods focus on strengthening the visual feature learning capability to reduce this text shortcut influence on model decisions. However, few efforts have been devoted to analyzing its inherent cause and providing an explicit interpretation. It thus lacks a good guidance for the research community to move forward in a purposeful way, resulting in model construction perplexity towards overcoming this non-trivial problem. In this paper, we propose to interpret the language prior problem in VQA from a class-imbalance view. Concretely, we design a novel interpretation scheme whereby the loss of mis-predicted frequent and sparse answers from the same question type is distinctly exhibited during the late training phase. It explicitly reveals why the VQA model tends to produce a frequent yet obviously wrong answer, to a given question whose right answer is sparse in the training set. Based upon this observation, we further propose a novel loss re-scaling approach to assign different weights to each answer according to the training data statistics for estimating the final loss. We apply our approach into six strong baselines and the experimental results on two VQA-CP benchmark datasets evidently demonstrate its effectiveness. In addition, we also justify the validity of the class imbalance interpretation scheme on other computer vision tasks, such as face recognition and image classification. [ABSTRACT FROM AUTHOR]

Kernel methods have achieved tremendous success in the past two decades. In the current big data era, data collection has grown tremendously. However, existing kernel methods are not scalable enough both at the training and predicting steps. To address this challenge, in this paper, we first introduce a general sparse kernel learning formulation based on the random feature approximation, where the loss functions are possibly non-convex. In order to reduce the scale of random features required in experiment, we also use that formulation based on the orthogonal random feature approximation. Then we propose a new asynchronous parallel doubly stochastic algorithm for large scale sparse kernel learning (AsyDSSKL). To the best our knowledge, AsyDSSKL is the first algorithm with the techniques of asynchronous parallel computation and doubly stochastic optimization. We also provide a comprehensive convergence guarantee to AsyDSSKL. Importantly, the experimental results on various large-scale real-world datasets show that, our AsyDSSKL method has the significant superiority on the computational efficiency at the training and predicting steps over the existing kernel methods. [ABSTRACT FROM AUTHOR]

Federated learning enables clients to train a machine learning model jointly without sharing their local data. However, due to the centrality of federated learning framework and the untrustworthiness of clients, traditional federated learning solutions are vulnerable to poisoning attacks from malicious clients and servers. In this paper, we aim to mitigate the impact of the central server and malicious clients by designing a Privacy-preserving Byzantine-robust Federated Learning (PBFL) scheme based on blockchain. Specifically, we use cosine similarity to judge the malicious gradients uploaded by malicious clients. Then, we adopt fully homomorphic encryption to provide secure aggregation. Finally, we use blockchain system to facilitate transparent processes and implementation of regulations. Our formal analysis proves that our scheme achieves convergence and provides privacy protection. Our extensive experiments on different datasets demonstrate that our scheme is robust and efficient. Even if the root dataset is small, our scheme can achieve the same efficiency as FedSGD. [ABSTRACT FROM AUTHOR]

Over the past years, the increasingly severe data island problem has spawned an emerging distributed deep learning framework—federated learning, in which the global model can be constructed over multiple participants without directly sharing their raw data. Despite its promising prospect, there are still many security challenges in federated learning, such as privacy preservation and integrity verification. Furthermore, federated learning is usually performed with the assistance of a center, which is prone to cause trust worries and communicational bottlenecks. To tackle these challenges, in this paper, we propose a privacy-preserving and verifiable decentralized federated learning framework, named PVD-FL, which can achieve secure deep learning model training under a decentralized architecture. Specifically, we first design an efficient and verifiable cipher-based matrix multiplication (EVCM) algorithm to execute the most basic calculation in deep learning. Then, by employing EVCM, we design a suite of decentralized algorithms to construct the PVD-FL framework, which ensures the confidentiality of both global model and local update and the verification of every training step. Detailed security analysis shows that PVD-FL can well protect privacy against various inference attacks and guarantee training integrity. In addition, the extensive experiments on real-world datasets also demonstrate that PVD-FL can achieve lossless accuracy and practical performance. [ABSTRACT FROM AUTHOR]

Federated learning (FL) can protect clients’ privacy from leakage in distributed machine learning. Applying federated learning to edge computing can protect the privacy of edge clients and benefit edge computing. Nevertheless, eavesdroppers can analyze the parameter information to specify clients’ private information and model features. And it is difficult to achieve a high privacy level, convergence, and low communication overhead during the entire process in the FL framework. In this paper, we propose a novel privacy-preserving federated learning framework for edge computing (PFLF). In PFLF, each client and the application server add noise before sending the data. To protect the privacy of clients, we design a flexible arrangement mechanism to count the optimal training times for clients. We prove that PFLF guarantees the privacy of clients and servers during the entire training process. Then, we theoretically prove that PFLF has three main properties: 1) For a given privacy level and model aggregation times, there is an optimal number of participating times for clients; 2) There is an upper and lower bound of convergence; 3) PFLF achieves low communication overhead by designing a flexible participation training mechanism. Simulation experiments confirm the correctness of our theoretical analysis. Therefore, PFLF helps design a framework to balance privacy levels and convergence and achieve low communication overhead when there is a part of clients dropping out of training. [ABSTRACT FROM AUTHOR]

The uncontrollable acquisition process limits the performance of iris recognition. In the acquisition process, various inevitable factors, including eyes, devices, and environment, hinder the iris recognition system from learning a discriminative identity representation. This leads to severe performance degradation. In this paper, we explore uncertain acquisition factors and propose uncertainty embedding (UE) and uncertainty-guided curriculum learning (UGCL) to mitigate the influence of acquisition factors. UE represents an iris image using a probabilistic distribution rather than a deterministic point (binary template or feature vector) that is widely adopted in iris recognition methods. Specifically, UE learns identity and uncertainty features from the input image, and encodes them as two independent components of the distribution, mean and variance. Based on this representation, an input image can be regarded as an instantiated feature sampled from the UE, and we can also generate various virtual features through sampling. UGCL is constructed by imitating the progressive learning process of newborns. Particularly, it selects virtual features to train the model in an easy-to-hard order at different training stages according to their uncertainty. In addition, an instance-level enhancement method is developed by utilizing local and global statistics to mitigate the data uncertainty from image noise and acquisition conditions in the pixel-level space. The experimental results on six benchmark iris datasets verify the effectiveness and generalization ability of the proposed method on same-sensor and cross-sensor recognition. [ABSTRACT FROM AUTHOR]

Model compression is a widely-used approach for reducing the size of deep learning models without much accuracy loss, enabling resource-hungry models to be compressed for use on resource-constrained devices. In this paper, we study the risk that model compression could provide an opportunity for adversaries to inject stealthy backdoors. In a backdoor attack on a machine learning model, an adversary produces a model that performs well on normal inputs but outputs targeted misclassifications on inputs containing a small trigger pattern. We design stealthy backdoor attacks such that the full-sized model released by adversaries appears to be free from backdoors (even when tested using state-of-the-art techniques), but when the model is compressed it exhibits a highly effective backdoor. We show this can be done for two common model compression techniques—model pruning and model quantization—even in settings where the adversary has limited knowledge of how the particular compression will be done. Our findings demonstrate the importance of performing security tests on the models that will actually be deployed not in their precompressed version. Our implementation is available at https://github.com/yulongtzzz/Stealthy-Backdoors-as-Compression-Artifacts. [ABSTRACT FROM AUTHOR]

In recent years, data has become the new oil that fuels various machine learning (ML) applications. Just as the oil refining, providing data to an ML model is a product of massive costs and expertise efforts. However, how to protect the intellectual property (IP) of the training data in ML remains largely open. In this paper, we present MeFA, a novel framework for detecting training data IP embezzlement via Membership Fingerprint Authentication, which is able to determine whether a suspect ML model is trained on the to be protected target data or not. The key observation is that a part of data has a similar influence on the prediction behavior of different ML models. On this basis, MeFA leverages membership inference techniques to extract these data as the fingerprints of the target data and constructs an authentication model to verify the data’s ownership by identifying the obtained membership fingerprints. MeFA has several salient features. It does not assume any knowledge of the suspect model except for its black-box prediction API, through which we can merely get the prediction output of a given input, and also does not require any modification to the dataset or the training process, since it takes advantage of the inherent membership property of the data. As a by-product, MeFA can also serve as a post-protection to verify the ownership of ML models, without modifying the training process of the model. Extensive experiments on three realistic datasets and seven types of ML models validate the effectiveness of MeFA, and demonstrate that it is also robust to scenarios when the training data is partially used or preprocessed with representative membership inference defenses. [ABSTRACT FROM AUTHOR]

Deep learning-based auto-driving systems are vulnerable to adversarial examples attacks which may result in wrong decision making and accidents. An adversarial example can fool the well trained neural networks by adding barely imperceptible perturbations to clean data. In this paper, we explore the mechanism of adversarial examples and adversarial robustness from the perspective of statistical mechanics, and propose an statistical mechanics-based interpretation model of adversarial robustness. The state transition caused by adversarial training based on the theory of fluctuation dissipation disequilibrium in statistical mechanics is formally constructed. Besides, we fully study the adversarial example attacks and training process on system robustness, including the influence of different training processes on network robustness. Our work is helpful to understand and explain the adversarial examples problems and improve the robustness of deep learning-based auto-driving systems. [ABSTRACT FROM AUTHOR]

There is considerable evidence that evaluating the subjective risk level of driving decisions can improve the safety of Autonomous Driving Systems (ADS) in both typical and complex driving scenarios. In this paper, we propose a novel data-driven approach that uses scene-graphs as intermediate representations for modeling the subjective risk of driving maneuvers. Our approach includes a Multi-Relation Graph Convolution Network, a Long-Short Term Memory Network, and attention layers. To train our model, we formulate subjective risk assessment as a supervised scene classification problem. We evaluate our model on both synthetic lane-changing datasets and real-driving datasets with various driving maneuvers. We show that our approach achieves a higher classification accuracy than the state-of-the-art approach on both large (96.4% vs. 91.2%) and small (91.8% vs. 71.2%) lane-changing synthesized datasets, illustrating that our approach can learn effectively even from small datasets. We also show that our model trained on a lane-changing synthesized dataset achieves an average accuracy of 87.8% when tested on a real-driving lane-changing dataset. In comparison, the state-of-the-art model trained on the same synthesized dataset only achieved 70.3% accuracy when tested on the real-driving dataset, showing that our approach can transfer knowledge more effectively. Moreover, we demonstrate that the addition of spatial and temporal attention layers improves our model’s performance and explainability. Finally, our results illustrate that our model can assess the risk of various driving maneuvers more accurately than the state-of-the-art model (86.5% vs. 58.4%, respectively). [ABSTRACT FROM AUTHOR]

An emerging computational paradigm, named federated edge learning (FEL), enables intelligent computing at the network edge with the feature of preserving data privacy for edge devices. Given their constrained resources, it becomes a great challenge to achieve high execution performance for FEL. Most of the state-of-the-arts concentrate on enhancing FEL from the perspective of system operation procedures, taking few precautions during the composition step of the FEL system. Though a few recent studies recognize the importance of FEL formation and propose server-centric device selection schemes, the impact of data sizes is largely overlooked. In this paper, we take advantage of game theory to depict the decision dilemma among edge devices regarding whether to participate in FEL or not given their heterogeneous sizes of local datasets. For realizing both the individual and global optimization, the server is employed to solve the participation dilemma, which requires accurate information collection for devices’ local datasets. Hence, we utilize mechanism design to enable truthful information solicitation. With the help of correlated equilibrium, we derive a decision making strategy for devices from the global perspective, which can achieve the long-term stability and efficacy of FEL. For scalability consideration, we optimize the computational complexity of the basic solution to the polynomial level. Lastly, extensive experiments based on both real and synthetic data are conducted to evaluate our proposed mechanisms, with experimental results demonstrating the performance advantages. [ABSTRACT FROM AUTHOR]

Actuated by the growing attention to personal healthcare and the pandemic, the popularity of E-health is proliferating. Nowadays, enhancement on medical diagnosis via machine learning models has been highly effective in many aspects of e-health analytics. Nevertheless, in the classic cloud-based/centralized e-health paradigms, all the data will be centrally stored on the server to facilitate model training, which inevitably incurs privacy concerns and high time delay. Distributed solutions like Decentralized Stochastic Gradient Descent (D-SGD) are proposed to provide safe and timely diagnostic results based on personal devices. However, methods like D-SGD are subject to the gradient vanishing issue and usually proceed slowly at the early training stage, thereby impeding the effectiveness and efficiency of training. In addition, existing methods are prone to learning models that are biased towards users with dense data, compromising the fairness when providing E-health analytics for minority groups. In this paper, we propose a Decentralized Block Coordinate Descent (D-BCD) learning framework that can better optimize deep neural network-based models distributed on decentralized devices for E-health analytics. As a gradient-free optimization method, Block Coordinate Descent (BCD) mitigates the gradient vanishing issue and converges faster at the early stage compared with the conventional gradient-based optimization. To overcome the potential data scarcity issues for users’ local data, we propose similarity-based model aggregation that allows each on-device model to leverage knowledge from similar neighbor models, so as to achieve both personalization and high accuracy for the learned models. Benchmarking experiments on three real-world datasets illustrate the effectiveness and practicality of our proposed D-BCD, where additional simulation study showcases the strong applicability of D-BCD in real-life E-health scenarios. [ABSTRACT FROM AUTHOR]

Hard sample selection can effectively improve model convergence by extracting the most representative samples from a training set. However, due to the large capacity of medical images, existing sampling strategies suffer from insufficient exploitation for hard samples or high time cost for sample selection when adopted by 3D patch-based models in the field of multi-organ segmentation. In this paper, we present a novel and effective online hard patch mining (OHPM) algorithm. In our method, an average shape model that can be mapped with all training images is constructed to guide the exploration of hard patches and aggregate feedback from predicted patches. The process of hard mining is formalized as a multi-armed bandit problem and solved with bandit algorithms. With the shape model, OHPM requires negligible time consumption and can intuitively locate difficult anatomical areas during training. The employment of bandit algorithms ensures online and sufficient hard mining. We integrate OHPM with advanced segmentation networks and evaluate them on two datasets containing different anatomical structures. Comparative experiments with other sampling strategies demonstrate the superiority of OHPM in boosting segmentation performance and improving model convergence. The results in each dataset with each network suggest that OHPM significantly outperforms other sampling strategies by nearly 2% average Dice score. [ABSTRACT FROM AUTHOR]

Deep neural models, in recent years, have been successful in almost every field, even solving the most complex problem statements. However, these models are huge in size with millions (and even billions) of parameters, demanding heavy computation power and failing to be deployed on edge devices. Besides, the performance boost is highly dependent on redundant labeled data. To achieve faster speeds and to handle the problems caused by the lack of labeled data, knowledge distillation (KD) has been proposed to transfer information learned from one model to another. KD is often characterized by the so-called ‘Student-Teacher’ (S-T) learning framework and has been broadly applied in model compression and knowledge transfer. This paper is about KD and S-T learning, which are being actively studied in recent years. First, we aim to provide explanations of what KD is and how/why it works. Then, we provide a comprehensive survey on the recent progress of KD methods together with S-T frameworks typically used for vision tasks. In general, we investigate some fundamental questions that have been driving this research area and thoroughly generalize the research progress and technical details. Additionally, we systematically analyze the research status of KD in vision applications. Finally, we discuss the potentials and open challenges of existing methods and prospect the future directions of KD and S-T learning. [ABSTRACT FROM AUTHOR]

The ability to predict, anticipate and reason about future outcomes is a key component of intelligent decision-making systems. In light of the success of deep learning in computer vision, deep-learning-based video prediction emerged as a promising research direction. Defined as a self-supervised learning task, video prediction represents a suitable framework for representation learning, as it demonstrated potential capabilities for extracting meaningful representations of the underlying patterns in natural videos. Motivated by the increasing interest in this task, we provide a review on the deep learning methods for prediction in video sequences. We first define the video prediction fundamentals, as well as mandatory background concepts and the most used datasets. Next, we carefully analyze existing video prediction models organized according to a proposed taxonomy, highlighting their contributions and their significance in the field. The summary of the datasets and methods is accompanied with experimental results that facilitate the assessment of the state of the art on a quantitative basis. The paper is summarized by drawing some general conclusions, identifying open research challenges and by pointing out future research directions. [ABSTRACT FROM AUTHOR]

Communication limitation at the edge is widely recognized as a major bottleneck for federated learning (FL). Multi-hop wireless networking provides a cost-effective solution to enhance service coverage and spectrum efficiency at the edge, which could facilitate large-scale and efficient machine learning (ML) model aggregation. However, FL over multi-hop wireless networks has rarely been investigated. In this paper, we optimize FL over wireless mesh networks by taking into account the heterogeneity in communication and computing resources at mesh routers and clients. We present a framework that each intermediate router performs in-network model aggregation before sending the data to the next hop, so as to reduce the outgoing data traffic and hence aggregate more models under limited communication resources. To accelerate model training, we formulate our optimization problem by jointly considering model aggregation, routing, and spectrum allocation. Although the problem is a non-convex mixed-integer nonlinear programming, we transform it into a mixed-integer linear programming (MILP), and develop a coarse-grained fixing procedure to solve it efficiently. Simulation results demonstrate the effectiveness of the solution approach, and the superiority of the in-network aggregation scheme over the counterpart without in-network aggregation. [ABSTRACT FROM AUTHOR]

Machine learning (ML) has attracted a great research interest for physical layer design problems, such as channel estimation, thanks to its low complexity and robustness. Channel estimation via ML requires model training on a dataset, which usually includes the received pilot signals as input and channel data as output. In previous works, model training is mostly done via centralized learning (CL), where the whole training dataset is collected from the users at the base station (BS). This approach introduces huge communication overhead for data collection. In this paper, to address this challenge, we propose a federated learning (FL) framework for channel estimation. We design a convolutional neural network (CNN) trained on the local datasets of the users without sending them to the BS. We develop FL-based channel estimation schemes for both conventional and RIS (intelligent reflecting surface) assisted massive MIMO (multiple-input multiple-output) systems, where a single CNN is trained for two different datasets for both scenarios. We evaluate the performance for noisy and quantized model transmission and show that the proposed approach provides approximately 16 times lower overhead than CL, while maintaining satisfactory performance close to CL. Furthermore, the proposed architecture exhibits lower estimation error than the state-of-the-art ML-based schemes. [ABSTRACT FROM AUTHOR]

In this paper, the content popularity prediction problem in fog radio access networks (F-RANs) is investigated. In order to obtain accurate prediction with low complexity, we propose a novel context-aware popularity prediction policy based on federated learning (FL). Firstly, user preference learning is applied by considering that users prefer to request the contents they are interested in. Then, users’ context information is utilized to cluster users efficiently by adaptive context space partitioning. After that, we formulate a popularity prediction optimization problem to learn the local model parameters by using the stochastic variance reduced gradient (SVRG) algorithm. Finally, FL based model integration is proposed to learn the global popularity prediction model based on local models using the distributed approximate Newton (DANE) algorithm with SVRG. Our proposed popularity prediction policy not only can predict content popularity accurately, but also can significantly reduce computational complexity. Moreover, we theoretically analyze the convergence bound of our proposed FL based model integration algorithm. Simulation results show that our proposed policy increases the cache hit rate by up to 21.5 % compared to existing policies. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI)-powered autonomous vehicles (AVs) can integrate different machine learning (ML) techniques to build up a complex autonomous driving system. However, single AV intelligence is not enough to cope with ever-changing driving environments. The underlying reason is that, with current neural network design and training algorithms, it is challenging for the driving model to generalize to diverse driving environments all at once due to sample inefficiency and the curse of dimensionality. Powerful computing resources and massive amount of data can be used to train a good driving model offline. However, the driving model obtained offline might fail in corner case scenarios. In this paper, we propose an intelligence networking framework among AVs assisted by multi-access edge computing (MEC) with end-to-end learning for demonstration. In this framework, driving road is divided into segments and data is collected for each road segment separately. Assisted by MEC networks, a continuously updated driving model is produced in near real-time for each road segment when the environment changes. By dividing the road into segments, we aim to reduce the burden of generalization since a single model only needs to adapt to a specific road segment. Simulation results show that our solutions can produce updated driving model for each road segment to adapt to environmental changes better than the existing scheme. Upcoming AVs can then adapt to changing environments by downloading the updated driving model. [ABSTRACT FROM AUTHOR]

Temporal and spatial contexts, characterizing target appearance variations and target-background differences, respectively, are crucial for improving the online adaptive ability and instance-level discriminative ability of object tracking. However, most existing trackers focus on either the temporal context or the spatial context during tracking and have not exploited these contexts simultaneously and effectively. In this paper, we propose a Spatial-TEmporal Memory (STEM) network to exploit these contexts jointly for object tracking. Specifically, we develop a key-value structured memory model equipped with a key-value index-based memory reading mechanism to model the spatial and temporal contexts simultaneously. To update the memory with new target states and ensure the diversity of the memory, we introduce a similarity-aware memory update scheme. In addition, we construct an entropy-guided ensemble strategy to fuse the prediction models based on these two contexts, such that these two contexts can be exploited to estimate the target state jointly. Extensive experimental results on eight challenging datasets, including OTB2015, TC128, UAV123, VOT2018, LaSOT, TrackingNet, GOT-10k, and OxUvA, demonstrate that the proposed method performs favorably against state-of-the-art trackers. [ABSTRACT FROM AUTHOR]

In an Internet of Vehicle (IoV) system, federated learning (FL) is a new approach to process real-time vehicle data in a distributed way, which can improve the driving experience and service quality. However, due to the high mobility and uncertainty of vehicles, the existing federated learning protocols are difficult to meet the full requirements from an IoV system, such as efficient resource allocation, high precision learning, and fast convergence of learning algorithm. To solve the above problems, in this paper, we propose a semi-synchronous federated learning (Semi-SynFed) protocol, to improve the performance of machine learning at Internet of Vehicles. In Semi-SynFed, we first select appropriate nodes to participate the aggregation by its computing capacity, network capacity and learning value of training samples. Meanwhile, the dynamic waiting time technique is designed to adjust the server waiting time at each round dynamically, which makes federated learning process work much more efficient. Finally, a dynamic aggregation scheme is designed to aggregate model parameters in an asynchronous way. To evaluate the reliability of Semi-SynFed, we establish a simulation of IoVs based on the real world data. In the comparative experiments, the results show that the Semi-SynFed outperforms the existing federated learning protocols in terms of convergence speed and resource consumption. [ABSTRACT FROM AUTHOR]

Federated learning (FL) is a popular edge learning approach that utilizes local data and computing resources of network edge devices to train machine learning (ML) models while preserving users’ privacy. Nevertheless, performing efficient learning tasks on the devices and achieving longer battery life are primary challenges faced by federated learning. In this paper, we are the first to study the application of heterogeneous computing (HC) and wireless power transfer (WPT) to federated learning to address these challenges. Especially, we propose a heterogeneous computation and resource allocation framework based on a heterogeneous mobile architecture to achieve effective implementation of FL. To minimize the energy consumption of smart devices and maximize their harvesting energy simultaneously, we formulate an optimization problem featuring multidimensional control, which jointly considers time splitting for WPT, dataset size allocation, transmit power allocation and subcarrier assignment during communications, and processor frequency of processing units (central processing unit (CPU) and graphics processing unit (GPU)). However, the major obstacle is how to design a proper algorithm to solve this optimization problem efficiently. For this purpose, we decouple the optimization variables so as to achieve high efficiency in deriving its solution. Particularly, we first compute the optimal processor frequency and dataset size allocation via employing the Lagrangian dual method, followed by finding the closed-form solution to the optimal time splitting allocation, and finally attain the optimal subcarrier assignment as well as transmit power for transmissions through an iteration algorithm. To evaluate the performance of our proposed scheme, we set up four baseline schemes as comparison, and simulation results show that the proposed scheme converges quite fast and better enhance the energy efficiency of the wireless powered FL system compared with the baseline schemes. [ABSTRACT FROM AUTHOR]

Edge federated learning (FL) is an emerging paradigm that trains a global parametric model from distributed datasets based on wireless communications. This paper proposes a unit-modulus over-the-air computation (UMAirComp) framework to facilitate efficient edge federated learning, which simultaneously uploads local model parameters and updates global model parameters via analog beamforming. The proposed framework avoids sophisticated baseband signal processing, leading to low communication delays and implementation costs. Training loss bounds of UMAirComp FL systems are derived and two low-complexity large-scale optimization algorithms, termed penalty alternating minimization (PAM) and accelerated gradient projection (AGP), are proposed to minimize the nonconvex nonsmooth loss bound. Simulation results show that the proposed UMAirComp framework with PAM algorithm achieves a smaller mean square error of model parameters’ estimation, training loss, and test error compared with other benchmark schemes. Moreover, the proposed UMAirComp framework with AGP algorithm achieves satisfactory performance while reduces the computational complexity by orders of magnitude compared with existing optimization algorithms. Finally, we demonstrate the implementation of UMAirComp in a vehicle-to-everything autonomous driving simulation platform. It is found that autonomous driving tasks are more sensitive to model parameter errors than other tasks since the neural networks for autonomous driving contain sparser model parameters. [ABSTRACT FROM AUTHOR]

As an emerging distributed machine learning (ML) method, federated learning (FL) can protect data privacy through collaborative learning of artificial intelligence (AI) models across a large number of devices. However, inefficiency and vulnerability to poisoning attacks have slowed FL performance. Therefore, a blockchain-based asynchronous federated learning (BAFL) framework is proposed to ensure the security and efficiency required by FL. The blockchain ensures that the model data cannot be tampered with while asynchronous learning speeds up global aggregation. A novel entropy weight method is used to evaluate the participating rank and proportion of the local model trained in BAFL of the devices. The energy consumption and local model update efficiency are balanced by adjusting the local training and communication delay and optimizing the block generation rate. The extensive evaluation results show that the proposed BAFL framework has higher efficiency and higher performance for preventing poisoning attacks than other distributed ML methods. [ABSTRACT FROM AUTHOR]