Search

In recent years many amazing deep learning models have been developed, but in the process of practical applications, people often find that these deep learning models have high requirements for hardware storage space and computing power. In Artificial Intelligent of Things (AIoT) scenario, the computing power of the edge or terminal side are relatively limited, therefore, most conventional deep learning models are difficult to be deployed into AIoT devices. It is significant to explore the different performance under different scales of deep learning models. In this paper, we mainly propose a method to analyze the impact of deep learning models with various sizes through various experiments. We employ slimmable network as a Neural Archtecture Search (NAS) tool to realize various model size freely, and evaluate them on the indicators of flops, robustness and accuracy. The experimental results show the variation of flops, robustness and accuracy with the various model sizes, which help understand the impact on performance of deep learning models with different scales in AIoT systems. • Interpreting robustness of deep learning models with different model complexity. • Employing slimmable network to compress neural network. • Understanding the relevance between robustness and accuracy of deep learning models. [ABSTRACT FROM AUTHOR]

The latest advances in artificial intelligence deep learning models are unprecedented. A wide spectrum of application areas is now thriving thanks to available massive training datasets and gigantic complex neural network models. Those two characteristics demand outstanding computing power that only advanced computing platforms can provide. Therefore, distributed deep learning has become a necessity in capitalizing on the potential of cutting-edge artificial intelligence. Two basic schemes have emerged in distributed learning. First, the data-parallel approach, which aims at dividing the training dataset into multiple computing nodes. Second, the model-parallel approach, which splits layers of a model into several computing nodes. Each scheme has its upsides and downsides, particularly when running on large machines that are susceptible to soft errors. Those errors occur as a consequence of several factors involved in the manufacturing process of current electronic components of supercomputers. On many occasions, those errors are expressed as bit flips that do not cause the whole system to crash, but generate wrong numerical results in computations. To study the effect of soft error on different approaches for distributed learning, we leverage checkpoint alteration, a technique that injects bit flips on checkpoint files. It allows researchers to understand the effect of soft errors on applications that produce checkpoint files in HDF5 format. This paper uses the popular deep learning PyTorch tool on two distributed-learning platforms: one for data-parallel training and one for model-parallel training. We use well-known deep learning models with popular training datasets to provide a picture of how soft errors challenge the training phase of a deep learning model. [ABSTRACT FROM AUTHOR]

• A novel approach for single channel speech enhancement is presented, using the colored spectrograms. • Using deep neural network for spectrogram denoising and regression neural network for translating the colors of denoised spectrograms to short time Fouier transform (STFT) magnitudes. • Output is comparable to the state-of-the-art baseline models at a much reduced computational cost. Speech enhancement concerns the processes required to remove unwanted background sounds from the target speech to improve its quality and intelligibility. In this paper, a novel approach for single-channel speech enhancement is presented using colored spectrograms. We propose the use of a deep neural network (DNN) architecture adapted from the pix2pix generative adversarial network (GAN) and train it over colored spectrograms of speech to denoise them. After denoising, the colors of spectrograms are translated to magnitudes of short-time Fourier transform (STFT) using a shallow regression neural network. These estimated STFT magnitudes are later combined with the noisy phases to obtain an enhanced speech. The results show an improvement of almost 0.84 points in the perceptual evaluation of speech quality (PESQ) and 1 % in the short-term objective intelligibility (STOI) over the unprocessed noisy data. The gain in quality and intelligibility over the unprocessed signal is almost equal to the gain achieved by the baseline methods used for comparison with the proposed model, but at a much reduced computational cost. The proposed solution offers a comparative PESQ score at almost 10 times reduced computational cost than a similar baseline model that has generated the highest PESQ score trained on grayscaled spectrograms, while it provides only a 1 % deficit in STOI at 28 times reduced computational cost when compared to another baseline system based on convolutional neural network-GAN (CNN-GAN) that produces the most intelligible speech. [ABSTRACT FROM AUTHOR]

Conventional automatic speaker verification systems can usually be decomposed into a front-end model such as time delay neural network (TDNN) for extracting speaker embeddings and a back-end model such as statistics-based probabilistic linear discriminant analysis (PLDA) or neural network-based neural PLDA (NPLDA) for similarity scoring. However, the sequential optimization of the front-end and back-end models may lead to a local minimum, which theoretically prevents the whole system from achieving the best optimization. Although some methods have been proposed for jointly optimizing the two models, such as the generalized end-to-end (GE2E) model and NPLDA E2E model, most of these methods have not fully investigated how to model the intra-relationship between multiple enrollment utterances. In this paper, we propose a new E2E joint method for speaker verification especially designed for the practical scenario of multiple enrollment utterances. To leverage the intra-relationship among multiple enrollment utterances, our model comes equipped with frame-level and utterance-level attention mechanisms. Additionally, focal loss is utilized to balance the importance of positive and negative samples within a mini-batch and focus on the difficult samples during the training process. We also utilize several data augmentation techniques, including conventional noise augmentation using MUSAN and RIRs datasets and a unique speaker embedding-level mixup strategy for better optimization. • Extend attention back-end to end-to-end for multi-enrollment speaker verification. • Introduce focal loss, suggest speaker embedding mixup for fine-tuning end-to-end model. • Analyze the effect of the number of enrollment utterances. • Analyze the embedding-level attention weights for speech genres. [ABSTRACT FROM AUTHOR]

Speech difficulties are common in children and teenagers, but they can also occur in adults as a result of physical problems. A speech disorder is a situation in which an individual struggles to produce or construct the spoken sounds necessary for interpersonal communication. As a result, it could be challenging to comprehend the person's speech. Articulation abnormalities are typical speech problems. In this situation, automatic speech recognition (ASR) technology may be used to detect and further rectify such deficiencies. The first attempts to detect speech abnormalities were made in the early 1970s, and they appear to have followed the same path as those on the ASR. These early experiments did rely heavily on signal processing techniques. As time goes on, more ideas from ASR technology are being incorporated into systems that deal with speech impairments. Many traditional techniques are executed in the ASR system. In this paper, we developed an automatic speech recognition technology based on deep learning techniques. In this paper, we research alternative extraction and classification methods of electroencephalography (EEG) to help diagnose speech disorders (SD). The EEG data is prepared before degradation into numerous EEG sub-strands with a discrete wavelet transformation to eliminate unimportant errors. For sharpening signals, the Eigenvector crack curvature wavelet method was used. A hyper-similarity abnormality coder is used for feature extraction in the EEG recording and to detect synchronization between EEG channels, which may show abnormalities in communication. The recovered functions are then categorized using the Deep Residual–encoder–based VGG net CNN Classification Method. Thus, the techniques proposed to produce the most promising outcome aren't the suggested technique attained better classification accuracy when compared to the traditional methodologies. [ABSTRACT FROM AUTHOR]

The performance of deep neural networks is typically associated with their depth and width. As the performance of a model improves, it demands greater computational resources and memory. However, self-distillation offers a pathway to augment this performance by internally transferring knowledge. Traditional self-distillation methods lack guidance from teacher networks, thus lacking the ability to obtain correct and key information. In this paper, we propose a new self-distillation framework—Global key knowledge distillation framework (GK-KD). We implement error correction mechanisms to network logit outputs and assess the importance of feature maps through rank to obtain correct and important knowledge. Experiments demonstrate the effectiveness of this framework: on the CIFAR100 dataset, ResNet18 achieved a 4.56% accuracy improvement over the baseline, ResNet152 also achieved a 5.26% accuracy improvement, and for the VGG series, the average accuracy increased by 5.82%, outperforming previous research methods without adding additional computational costs. Further experiments show that this framework can effectively enhance the robustness and calibration of the network. • The GK-KD framework we propose enhances network performance by distilling accurate and important knowledge. • Experiments show our framework balances accuracy, computational resources, and efficiency. • Effectiveness on three datasets shown; further tests prove method's universality, reliability, and robustness. [ABSTRACT FROM AUTHOR]

• A novel voice conversion technique for body-conducted unvoiced speech is proposed based on Taylor-Boltzmann Machines. The proposed algorithm was the integration of Taylor Series with Boltzmann machine Deep Neural Networks. • A body conductive microphone referred as the non audible murmur (NAM) microphone is used to capture the non-audible murmur from the neck part behind the ear. • The classifier was trained with the features likes "pitch chroma, spectral centroid, and spectral skewness" that are derived from the incoming body conducted low voice verbal communication signal. • In the initial training phase, a Taylor series approximation DNN model is trained from a collection of stereo data which encompasses the NAM speech of same phonemes represented as log-power spectral features. • Moreover, in the recognition phase, a well trained Taylor-DNN model that is based on Boltzmann Machine intakes the speech features and generates the "log-power-spectra features" and it aids in recognizing the speech waveform. • The paper may gain high interest from the readers and experts of Speech processing, Voice Conversion Techniques, body-conducted unvoiced speech. Communication becomes effectual while the verbal communication signal reaches the destination with profound characteristics. An application based on human voice attracts the researcher's attention. It leads the researcher to proceed towards the speech conversion for low voice humans having murmured voices as their attitude. The researchers introduced a technique for improving the body-conducted unvoiced speech for silent speech communication using Taylor-Boltzmann machine-trained optimized Deep Neural Network (DNN). The projected model includes two main stages: the training stage as well as recognition stage. In the training stage, the following functionalities are undergone: (i) pre-processing, (ii) feature extraction, and (iii) deep learning-based speech enhancement. Initially, the collected input Non-Audible Murmur (NAM) speech signal is subjected to the pre-processing stage, wherein the improved wiener filtering is employed for suppressing the noise in the input signal. From the pre-processed speech signal, the most relevant characteristics such as "spectral skewness, spectral chroma, spectral centroid and improved Mel Frequency Cepstral Coefficients (MFCC)" are extracted, and they are fused together. Then, using these fused features, the Taylor series with optimized DNN and Boltzmann Machine (Taylor-DNN-BM) model is trained. The final enhanced speech is acquired from the Taylor- optimized DNN-BM. The recognition phase encapsulates the feature extraction and deep learning-based speech enhancement stage. The input speech data enters into the feature extraction phase, wherein the characteristics such as "spectral skewness, spectral chroma, spectral centroid and improved (MFCC)" are retrieved. After that, these features are fused and fed as input to the Taylor-optimized DNN-BM model. From Taylor- optimized DNN-BM, the enhanced speech signal is acquired. In DNN, the weight function is fine-tuned using the newly projected Self Improved Flower Pollination Algorithm (SI-FPA) algorithm. This SI-FPA model is the conceptual improvement of the Flower Pollination Algorithm (FPA). Finally, the projected model is compared with the existing models to validate its efficiency in terms of speech enhancement. [ABSTRACT FROM AUTHOR]

This paper discusses how automatic speech recognition systems are and could be designed, in order to best exploit the discriminative information encoded in human speech. This contrasts with many recent machine learning approaches that apply general recognition architectures to signals to identify, with little concern for the nature of the input. The implicit assumption has often been that training can automatically discover the useful properties that exist in signals, with minimal manual intervention. These approaches may be suitable for some tasks such as image recognition, where the diversity of visual input is vast; e.g., an image may be any (natural or synthetic) scene that a camera views. We first examine what makes speech special, i.e., a natural signal from a complex tube, driven by a source that is quasi-periodic and/or noisy, aiming to communicate a wide variety of information, using the different vocal systems of human speakers. Then, we view how pertinent features are extracted from speech via efficient means, related to the objectives of communication. We see how to reliably and efficiently identify the different units of oral language. We learn from the history of attempts to do ASR, e.g., why they succeeded and how improved methods exploited the increasing availability of data and computer power (in particular, deep neural networks). Finally, we suggest ways to render ASR both more accurate and efficient. This work is aimed at both newcomers to ASR and experts, in terms of presenting issues broadly, but without mathematical or algorithmic details, which are readily found in the references. [ABSTRACT FROM AUTHOR]

Neural networks activated by the rectified linear unit (ReLU) play a central role in the recent development of deep learning. The topic of approximating functions from Hölder spaces by these networks is crucial for understanding the efficiency of the induced learning algorithms. Although the topic has been well investigated in the setting of deep neural networks with many layers of hidden neurons, it is still open for shallow networks having only one hidden layer. In this paper, we provide rates of uniform approximation by these networks. We show that ReLU shallow neural networks with m hidden neurons can uniformly approximate functions from the Hölder space W ∞ r ([ − 1 , 1 ] d) with rates O ((log ⁡ m) 1 2 + d m − r d d + 2 d + 4 ) when r < d / 2 + 2. Such rates are very close to the optimal one O (m − r d ) in the sense that d + 2 d + 4 is close to 1, when the dimension d is large. [ABSTRACT FROM AUTHOR]

Constructing neural networks for function approximation is a classical and longstanding topic in approximation theory. In this paper, we aim at constructing deep neural networks with three hidden layers using a sigmoidal activation function to approximate smooth and sparse functions. Specifically, we prove that the constructed deep nets with controllable magnitude of free parameters can reach the optimal approximation rate in approximating both smooth and sparse functions. In particular, we prove that neural networks with three hidden layers can avoid the phenomenon of saturation, i.e., the phenomenon that for some neural network architectures, the approximation rate stops improving for functions of very high smoothness. [ABSTRACT FROM AUTHOR]

In this paper, we address the problem of speaker verification in conditions unseen or unknown during development. A standard method for speaker verification consists of extracting speaker embeddings with a deep neural network and processing them through a backend composed of probabilistic linear discriminant analysis (PLDA) and global logistic regression score calibration. This method is known to result in systems that work poorly on conditions different from those used to train the calibration model. We propose to modify the standard backend, introducing an adaptive calibrator that uses duration and other automatically extracted side-information to adapt to the conditions of the inputs. The backend is trained discriminatively to optimize binary cross-entropy. When trained on a number of diverse datasets that are labeled only with respect to speaker, the proposed backend consistently and, in some cases, dramatically improves calibration, compared to the standard PLDA approach, on a number of held-out datasets, some of which are markedly different from the training data. Discrimination performance is also consistently improved. We show that joint training of the PLDA and the adaptive calibrator is essential — the same benefits cannot be achieved when freezing PLDA and fine-tuning the calibrator. To our knowledge, the results in this paper are the first evidence in the literature that it is possible to develop a speaker verification system with robust out-of-the-box performance on a large variety of conditions. • A standard backend for speaker verification is based on PLDA followed by global calibration. • We propose a novel backend that consists of a PLDA-like formulation followed by condition-aware calibration. • The backend is trained discriminatively to optimize binary cross-entropy on a large variety of conditions. • We show that this backend results in significant gains over the standard PLDA approach. • The proposed approach results in a system that can be used out-of-the-box on data from different conditions. [ABSTRACT FROM AUTHOR]

Deep neural networks have powerful capabilities, but they are vulnerable to adversarial attacks. For image classification tasks, if a small disturbance is introduced into the input image, the model is likely to be misled and causes misclassification. In this paper, we develop a robust attention ranking architecture with frequency-domain transform to defend against adversarial samples, which is called RARFTA. We import the discrete cosine transform as the activation layer after the first convolutional layer, which effectively suppresses the attack based on the gradient method. To eliminate the interference of residual adversarial noise, we dynamically select key points from the feature map for classification with the attention mechanism to reduce the impact of other attacked pixels. Experimental results on different datasets show that our method is superior to the existing defense methods in both black-box and white-box attacks and significantly improves the robustness of the deep neural network model. The code for our work is available at https://github.com/lixiaowenaaa/RARTFA/tree/master. • A new network architecture is designed to support dynamic key points selection. • An activation function based on the discrete cosine transform is incorporated into the model. • The model improves the robustness of the deep learning network. [ABSTRACT FROM AUTHOR]

To boost performance, deep neural networks require deeper or wider network structures that involve massive computational and memory costs. To alleviate this issue, the self-knowledge distillation method regularizes the model by distilling the internal knowledge of the model itself. Conventional self-knowledge distillation methods require additional trainable parameters or are dependent on the data. In this paper, we propose a simple and effective self-knowledge distillation method using a dropout (SD-Dropout). SD-Dropout distills the posterior distributions of multiple models through a dropout sampling. Our method does not require any additional trainable modules, does not rely on data, and requires only simple operations. Furthermore, this simple method can be easily combined with various self-knowledge distillation approaches. We provide a theoretical and experimental analysis of the effect of forward and reverse KL-divergences in our work. Extensive experiments on various vision tasks, i.e., image classification, object detection, and distribution shift, demonstrate that the proposed method can effectively improve the generalization of a single network. Further experiments show that the proposed method also improves calibration performance, adversarial robustness, and out-of-distribution detection ability. • We propose a simple self-knowledge distillation methodology using dropout techniques. • Our method can collaborate easily with other knowledge distillation methods. • We describe observations regarding the forward and reverse KL-divergence. • Extensive experiments demonstrate the effectiveness of our methodology. [ABSTRACT FROM AUTHOR]

The deep learning methodologies in state-of-the-art speaker recognition systems are predominantly limited to the extraction of recording level embeddings. This is usually followed by generative modeling of the embeddings to output the verification score. In this paper, we explore a fully neural approach where the neural model outputs the verification score directly, given the acoustic feature inputs. This model, termed as Siamese neural network (SiamNN), combines the embedding extraction and back-end modeling into a single processing pipeline. The back-end modeling is achieved using a neural approach to PLDA modeling, called neural probabilistic linear discriminant analysis (NPLDA). In the NPLDA model, the verification score is computed as a discriminative similarity function. The development of the single neural SiamNN model allows the joint optimization of all the modules using a verification cost. Several speaker recognition experiments are performed using SITW, VOiCES, and NIST SRE datasets where the proposed SiamNN model is shown to significantly improve over the state-of-art x-vector PLDA baseline system (relative improvements of up to 35% in the primary cost metric). We also provide a detailed analysis of the influence of hyper-parameters, choice of loss functions, and data sampling strategies for training the model. In particular, we highlight that the proposed soft detection cost function based optimization improves over other loss functions considered. [ABSTRACT FROM AUTHOR]

3D semantic segmentation of indoor scenes is a popular research topic in the field of computer vision. For many applications, it is very important to know exactly what category each point in the scene belongs to. Benefiting from the development of deep learning, many neural networks based on voxels and points have been proposed to solve these segmentation problems. However, most of them do not fully consider the information of the spatial structure. Current voxel-based sparse convolutional neural networks can effectively extract 3D features in space. However, they assume that the feature in the empty space is zero, causing a loss of information in the spatial structure. In this paper, we propose a system that uses scene point clouds with color information to semantically segment an entire indoor scene. Based on the sparsity of spatial data, we design a novel spatial-aware sparse convolution operation. We encode the spatial information of the object's existence as an additional feature and use the self-attention mechanism to effectively aggregate features. In addition, we introduce a completion network to refine the results from the segmentation network, so that each object in the scene is fitted into a more reasonable and complete shape. Through the above two methods, we build an accurate scene semantic segmentation network to obtain the semantic information of the entire scene. In the experimental part, we use two public datasets to perform quantitative and qualitative analysis. We compare our results with other state-of-the-art methods to prove the superiority of our method. Our models are also examined under different configurations to assure the effectiveness of the proposed method. Finally, the semantic segmentation model was integrated into a real-world application to demonstrate its usefulness. We expect that the proposed 3D scene semantic segmentation system can provide accurate and fast results for practical applications. • Utilizing deep neural networks based on voxels to solve these segmentation problems. • Fully consider the spatial structure of the model. • Using Scene point clouds with color information to semantically segment indoor scenes. • Proposed completion network to refine each object into a more reasonable shape. • Semantic segmentation model integrated into a real-world application. [ABSTRACT FROM AUTHOR]

The sorting of Construction and Demolition (C&D) waste is a critical step to linking the recycling system and to the macro prediction, which helps to promote the development of the circular economy. Moreover, the effective classification and automated separation process will also help to stop the spreading of pathogenic organisms, such as virus and bacteria, by minimizing human intervention in the sorting process, while also helping to prevent further contamination by COVID-19 virus. This study aims to develop an efficient method to sort C&D waste through deep learning combined with knowledge transfer approach. In this paper, CVGGNet models, that is four VGG structures (VGGNet-11, VGGNet-13, VGGNet-16, and VGGNet-19), based on knowledge transfer combined with the technology of data augmentation and cyclical learning rate, are proposed to classify ten types of C&D waste images. Results show that 2.5 × 10−4, 1.8 × 10−4, 0.8 × 10−4, and 1.0 × 10−4 are the optimum learning rate for CVGGNet-11, CVGGNet-13, CVGGNet-16, and CVGGNet-19, respectively. Knowledge transfer helped shorten the training time from 1039.45 s to 991.05 s, and while it improved the performance of the CVGGNet-11 model in training, validation, and test datasets. The average training time increases as the number of the layers in the CVGGNet architecture rises: CVGGNet-11 (991.05 s) ˂ CVGGNet-13 (1025.76 s) ˂ CVGGNet-16 (1090.48 s) ˂ CVGGNet-19 (1337.81 s). Compared to other CVGGNet models, CVGGNet-16 showed an excellent performance in various C&D waste types, in terms of accuracy (76.6%), weighted average precision (76.8%), weighted average recall (76.6%), weighted average F1-score (76.6%) and micro average ROC (87.0%). In addition, the t-distributed Stochastic Neighbor Embedding (t-SNE) approach can reduce the dataset to a lower dimension and distinctly separate each type of C&D waste. This study demonstrates the good performance of CVGGNet models that can be used to automatically sort most of the C&D waste, paving the way for better C&D waste management. [Display omitted] • Dataset of construction and demolition waste classification has been collected. • Transfer learning was applied to sort construction and demolition waste. • Cyclical learning rate was adopted to quickly tune hyperparameters. [ABSTRACT FROM AUTHOR]

The direction-of-arrival (DOA) of sound sources is an essential acoustic parameter used, e.g., for multi-channel speech enhancement or source tracking. Complex acoustic scenarios consisting of sources-of-interest, interfering sources, reverberation, and noise make the estimation of the DOAs corresponding to the sources-of-interest a challenging task. Recently proposed attention mechanisms allow DOA estimators to focus on the sources-of-interest and disregard interference and noise, i.e., they are signal-aware. The attention is typically obtained by a deep neural network (DNN) from a short-time Fourier transform (STFT) based representation of a single microphone signal. Subsequently, attention has been applied as binary or ratio weighting to STFT-based microphone signal representations to reduce the impact of frequency bins dominated by noise, interference, or reverberation. The impact of attention on DOA estimators and different training strategies for attention and DOA DNNs are not yet studied in depth. In this paper, we evaluate systems consisting of different DNNs and signal processing-based methods for DOA estimation when attention is applied. Additionally, we propose training strategies for attention-based DOA estimation optimized via a DOA objective, i.e., end-to-end. The evaluation of the proposed and the baseline systems is performed using data generated with simulated and measured room impulse responses of a uniform-linear microphone array under various acoustic conditions, like reverberation times, noise, and source array distances. The data contains a single source-of-interest, noise, and directional interference. The best-performing systems are also evaluated using measured data. Our experiments show that DNNs used for DOA estimation are biased to the spectral source characteristics and the spectral attention distribution used during training (e.g., spectrally flat/sparse). We also show that this bias in the DOA estimator can be avoided if signal-processing methods are used in combination with attention. Overall, DOA estimation using attention in combination with signal-processing methods exhibits a far lower computational complexity than a fully DNN-based system; however, it yields comparable results. • Attention enables source-selective direction of arrival estimation. • Combining data-driven and signal-processing methods reduces complexity. • Attention can be estimated using masking concepts from source separation. [ABSTRACT FROM AUTHOR]

Multi-source localization is an important and challenging technique for multi-talker conversation analysis. This paper proposes a novel supervised learning method using deep neural networks to estimate the direction of arrival (DOA) of all the speakers simultaneously from the audio mixture. At the heart of the proposal is a source splitting mechanism that creates source-specific intermediate representations inside the network. This allows our model to give source-specific posteriors as the output unlike the traditional multi-label classification approach. Existing deep learning methods perform a frame level prediction, whereas our approach performs an utterance level prediction by incorporating temporal selection and averaging inside the network to avoid post-processing. We also experiment with various loss functions and show that a variant of earth mover distance (EMD) is very effective in classifying DOA at a very high resolution by modeling inter-class relationships. In addition to using the prediction error as a metric for evaluating our localization model, we also establish its potency as a frontend with automatic speech recognition (ASR) as the downstream task. We convert the estimated DOAs into a feature suitable for ASR and pass it as an additional input feature to a strong multi-channel and multi-talker speech recognition baseline. This added input feature drastically improves the ASR performance and gives a word error rate (WER) of 6.3% on the evaluation data of our simulated noisy two speaker mixtures, while the baseline which does not use explicit localization input has a WER of 11.5%. We also perform ASR evaluation on real recordings with the overlapped set of the MC-WSJ-AV corpus in addition to simulated mixtures. • An end-to-end multi-source direction of arrival (DOA) estimation method is proposed. • The method tackles multi-source DOA estimation via single-source DOA estimation. • The DOA classification network is optimized to learn inter-class relationships for a fine-grained angle resolution. • DOA estimation is used as a frontend for multi-talker automatic speech recognition. [ABSTRACT FROM AUTHOR]

• Previous work has demonstrated that individual task-based brain activity can be predicted from resting-state functional connectivity. • We build on recent deep learning methods to create a surface-based fully-convolutional neural network model that works with a representation of the brain's cortical sheet. • The proposed model, BrainSurfCNN, can achieve state of the art predictive accuracy on independent test data from the Human Connectome Project. • BrainSurfCNN yields individual-level predicted maps that are on par with the target-repeat reliability of the measured contrast maps. • We further demonstrate that BrainSurfCNN can generalize well to novel domains with limited training data. Task-based and resting-state represent the two most common experimental paradigms of functional neuroimaging. While resting-state offers a flexible and scalable approach for characterizing brain function, task-based techniques provide superior localization. In this paper, we build on recent deep learning methods to create a model that predicts task-based contrast maps from resting-state fMRI scans. Specifically, we propose BrainSurfCNN, a surface-based fully-convolutional neural network model that works with a representation of the brain's cortical sheet. BrainSurfCNN achieves exceptional predictive accuracy on independent test data from the Human Connectome Project, which is on par with the repeat reliability of the measured subject-level contrast maps. Conversely, our analyses reveal that a previously published benchmark is no better than group-average contrast maps. Finally, we demonstrate that BrainSurfCNN can generalize remarkably well to novel domains with limited training data. [ABSTRACT FROM AUTHOR]

Cross-language article linking (CLAL), the task of generating links between articles in different languages from different encyclopedias, is critical for facilitating sharing among online knowledge bases. Some previous CLAL research has been done on creating links among Wikipedia wikis, but much of this work depends heavily on simple language patterns and encyclopedia format or metadata. In this paper, we propose a new CLAL method based on deep learning paragraph embeddings to link English Wikipedia articles with articles in Baidu Baike, the most popular online encyclopedia in mainland China. To measure article similarity for link prediction, we employ several neural networks with attention mechanisms, such as CNN and LSTM, to train paragraph encoders that create vector representations of the articles' semantics based only on article text, rather than link structure, as input data. Using our "Deep CLAL" method, we compile a data set consisting of Baidu Baike entries and corresponding English Wikipedia entries. Our approach does not rely on linguistic or structural features and can be easily applied to other language pairs by using pre-trained word embeddings, regardless of whether the two languages are on the same encyclopedia platform. • Cross-language article linking helps create a multilingual unified knowledge base. • Using attention-based neural network that learns to attend to the vital part of articles. • The novel method that does not rely on feature engineering and is scalable to large data. [ABSTRACT FROM AUTHOR]