PyTorchKR Papers

최신 논문

총 9,642편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2011.02809

Semi-supervised Learning for Singing Synthesis Timbre

We propose a semi-supervised singing synthesizer, which is able to learn new voices from audio data only, without any annotations such as phonetic segmentation. Our system is an encoder-decoder model with two encoders, linguistic and acoustic, and one (acoustic) decoder. In a first step, the system is trained in a supervised manner, using a labelled multi-singer dataset. Here, we ensure that the embeddings produced by both encoders are similar, so that we can later use the model with either acoustic or linguistic input features. To learn a new voice in an unsupervised manner, the pretrained acoustic encoder is used to train a decoder for the target singer. Finally, at inference, the pretrained linguistic encoder is used together with the decoder of the new voice, to produce acoustic features from linguistic input. We evaluate our system with a listening test and show that the results are comparable to those obtained with an equivalent supervised approach.

arXiv:2011.02678

BW-EDA-EEND: Streaming End-to-End Neural Speaker Diarization for a Variable Number of Speakers

We present a novel online end-to-end neural diarization system, BW-EDA-EEND, that processes data incrementally for a variable number of speakers. The system is based on the Encoder-Decoder-Attractor (EDA) architecture of Horiguchi et al., but utilizes the incremental Transformer encoder, attending only to its left contexts and using block-level recurrence in the hidden states to carry information from block to block, making the algorithm complexity linear in time. We propose two variants: For unlimited-latency BW-EDA-EEND, which processes inputs in linear time, we show only moderate degradation for up to two speakers using a context size of 10 seconds compared to offline EDA-EEND. With more than two speakers, the accuracy gap between online and offline grows, but the algorithm still outperforms a baseline offline clustering diarization system for one to four speakers with unlimited context size, and shows comparable accuracy with context size of 10 seconds. For limited-latency BW-EDA-EEND, which produces diarization outputs block-by-block as audio arrives, we show accuracy comparable to the offline clustering-based system.

arXiv:2011.02307

FDRN: A Fast Deformable Registration Network for Medical Images

Deformable image registration is a fundamental task in medical imaging. Due to the large computational complexity of deformable registration of volumetric images, conventional iterative methods usually face the tradeoff between the registration accuracy and the computation time in practice. In order to boost the registration performance in both accuracy and runtime, we propose a fast convolutional neural network. Specially, to efficiently utilize the memory resources and enlarge the model capacity, we adopt additive forwarding instead of channel concatenation and deepen the network in each encoder and decoder stage. To facilitate the learning efficiency, we leverage skip connection within the encoder and decoder stages to enable residual learning and employ an auxiliary loss at the bottom layer with lowest resolution to involve deep supervision. Particularly, the low-resolution auxiliary loss is weighted by an exponentially decayed parameter during the training phase. In conjunction with the main loss in high-resolution grid, a coarse-to-fine learning strategy is achieved. Last but not least, we introduce an auxiliary loss based on the segmentation prior to improve the registration performance in Dice score. Comparing to the auxiliary loss using average Dice score, the proposed multi-label segmentation loss does not induce additional memory cost in the training phase and can be employed on images with arbitrary amount of categories. In the experiments, we show FDRN outperforms the existing state-of-the-art registration methods for brain MR images by resorting to the compact network structure and efficient learning. Besides, FDRN is a generalized framework for image registration which is not confined to a particular type of medical images or anatomy.

arXiv:2011.02314

VAW-GAN for Disentanglement and Recomposition of Emotional Elements in Speech

Emotional voice conversion (EVC) aims to convert the emotion of speech from one state to another while preserving the linguistic content and speaker identity. In this paper, we study the disentanglement and recomposition of emotional elements in speech through variational autoencoding Wasserstein generative adversarial network (VAW-GAN). We propose a speaker-dependent EVC framework based on VAW-GAN, that includes two VAW-GAN pipelines, one for spectrum conversion, and another for prosody conversion. We train a spectral encoder that disentangles emotion and prosody (F0) information from spectral features; we also train a prosodic encoder that disentangles emotion modulation of prosody (affective prosody) from linguistic prosody. At run-time, the decoder of spectral VAW-GAN is conditioned on the output of prosodic VAW-GAN. The vocoder takes the converted spectral and prosodic features to generate the target emotional speech. Experiments validate the effectiveness of our proposed method in both objective and subjective evaluations.

arXiv:2011.02127

Sequence-to-Sequence Learning via Attention Transfer for Incremental Speech Recognition

Attention-based sequence-to-sequence automatic speech recognition (ASR) requires a significant delay to recognize long utterances because the output is generated after receiving entire input sequences. Although several studies recently proposed sequence mechanisms for incremental speech recognition (ISR), using different frameworks and learning algorithms is more complicated than the standard ASR model. One main reason is because the model needs to decide the incremental steps and learn the transcription that aligns with the current short speech segment. In this work, we investigate whether it is possible to employ the original architecture of attention-based ASR for ISR tasks by treating a full-utterance ASR as the teacher model and the ISR as the student model. We design an alternative student network that, instead of using a thinner or a shallower model, keeps the original architecture of the teacher model but with shorter sequences (few encoder and decoder states). Using attention transfer, the student network learns to mimic the same alignment between the current input short speech segments and the transcription. Our experiments show that by delaying the starting time of recognition process with about 1.7 sec, we can achieve comparable performance to one that needs to wait until the end.

arXiv:2011.02093

Chinese Grammatical Correction Using BERT-based Pre-trained Model

In recent years, pre-trained models have been extensively studied, and several downstream tasks have benefited from their utilization. In this study, we verify the effectiveness of two methods that incorporate a BERT-based pre-trained model developed by Cui et al. (2020) into an encoder-decoder model on Chinese grammatical error correction tasks. We also analyze the error type and conclude that sentence-level errors are yet to be addressed.

arXiv:2011.02081공식 코드 · Papers with Code

A deep learning classifier for local ancestry inference

Local ancestry inference (LAI) identifies the ancestry of each segment of an individual's genome and is an important step in medical and population genetic studies of diverse cohorts. Several techniques have been used for LAI, including Hidden Markov Models and Random Forests. Here, we formulate the LAI task as an image segmentation problem and develop a new LAI tool using a deep convolutional neural network with an encoder-decoder architecture. We train our model using complete genome sequences from 982 unadmixed individuals from each of five continental ancestry groups, and we evaluate it using simulated admixed data derived from an additional 279 individuals selected from the same populations. We show that our model is able to learn admixture as a zero-shot task, yielding ancestry assignments that are nearly as accurate as those from the existing gold standard tool, RFMix.

arXiv:2011.01991

Internal Language Model Estimation for Domain-Adaptive End-to-End Speech Recognition

The external language models (LM) integration remains a challenging task for end-to-end (E2E) automatic speech recognition (ASR) which has no clear division between acoustic and language models. In this work, we propose an internal LM estimation (ILME) method to facilitate a more effective integration of the external LM with all pre-existing E2E models with no additional model training, including the most popular recurrent neural network transducer (RNN-T) and attention-based encoder-decoder (AED) models. Trained with audio-transcript pairs, an E2E model implicitly learns an internal LM that characterizes the training data in the source domain. With ILME, the internal LM scores of an E2E model are estimated and subtracted from the log-linear interpolation between the scores of the E2E model and the external LM. The internal LM scores are approximated as the output of an E2E model when eliminating its acoustic components. ILME can alleviate the domain mismatch between training and testing, or improve the multi-domain E2E ASR. Experimented with 30K-hour trained RNN-T and AED models, ILME achieves up to 15.5% and 6.8% relative word error rate reductions from Shallow Fusion on out-of-domain LibriSpeech and in-domain Microsoft production test sets, respectively.

arXiv:2011.02293

Pixel-wise Dense Detector for Image Inpainting

Recent GAN-based image inpainting approaches adopt an average strategy to discriminate the generated image and output a scalar, which inevitably lose the position information of visual artifacts. Moreover, the adversarial loss and reconstruction loss (e.g., l1 loss) are combined with tradeoff weights, which are also difficult to tune. In this paper, we propose a novel detection-based generative framework for image inpainting, which adopts the min-max strategy in an adversarial process. The generator follows an encoder-decoder architecture to fill the missing regions, and the detector using weakly supervised learning localizes the position of artifacts in a pixel-wise manner. Such position information makes the generator pay attention to artifacts and further enhance them. More importantly, we explicitly insert the output of the detector into the reconstruction loss with a weighting criterion, which balances the weight of the adversarial loss and reconstruction loss automatically rather than manual operation. Experiments on multiple public datasets show the superior performance of the proposed framework. The source code is available at https://github.com/Evergrow/GDN_Inpainting.

arXiv:2011.02164

An Improved Attention for Visual Question Answering

We consider the problem of Visual Question Answering (VQA). Given an image and a free-form, open-ended, question, expressed in natural language, the goal of VQA system is to provide accurate answer to this question with respect to the image. The task is challenging because it requires simultaneous and intricate understanding of both visual and textual information. Attention, which captures intra- and inter-modal dependencies, has emerged as perhaps the most widely used mechanism for addressing these challenges. In this paper, we propose an improved attention-based architecture to solve VQA. We incorporate an Attention on Attention (AoA) module within encoder-decoder framework, which is able to determine the relation between attention results and queries. Attention module generates weighted average for each query. On the other hand, AoA module first generates an information vector and an attention gate using attention results and current context; and then adds another attention to generate final attended information by multiplying the two. We also propose multimodal fusion module to combine both visual and textual information. The goal of this fusion module is to dynamically decide how much information should be considered from each modality. Extensive experiments on VQA-v2 benchmark dataset show that our method achieves the state-of-the-art performance.

arXiv:2011.01519

SelfPose: 3D Egocentric Pose Estimation from a Headset Mounted Camera

We present a solution to egocentric 3D body pose estimation from monocular images captured from downward looking fish-eye cameras installed on the rim of a head mounted VR device. This unusual viewpoint leads to images with unique visual appearance, with severe self-occlusions and perspective distortions that result in drastic differences in resolution between lower and upper body. We propose an encoder-decoder architecture with a novel multi-branch decoder designed to account for the varying uncertainty in 2D predictions. The quantitative evaluation, on synthetic and real-world datasets, shows that our strategy leads to substantial improvements in accuracy over state of the art egocentric approaches. To tackle the lack of labelled data we also introduced a large photo-realistic synthetic dataset. xR-EgoPose offers high quality renderings of people with diverse skintones, body shapes and clothing, performing a range of actions. Our experiments show that the high variability in our new synthetic training corpus leads to good generalization to real world footage and to state of theart results on real world datasets with ground truth. Moreover, an evaluation on the Human3.6M benchmark shows that the performance of our method is on par with top performing approaches on the more classic problem of 3D human pose from a third person viewpoint.

arXiv:2011.01753

Attention Beam: An Image Captioning Approach

The aim of image captioning is to generate textual description of a given image. Though seemingly an easy task for humans, it is challenging for machines as it requires the ability to comprehend the image (computer vision) and consequently generate a human-like description for the image (natural language understanding). In recent times, encoder-decoder based architectures have achieved state-of-the-art results for image captioning. Here, we present a heuristic of beam search on top of the encoder-decoder based architecture that gives better quality captions on three benchmark datasets: Flickr8k, Flickr30k and MS COCO.

arXiv:2011.01482

Layer-Wise Multi-View Learning for Neural Machine Translation

Traditional neural machine translation is limited to the topmost encoder layer's context representation and cannot directly perceive the lower encoder layers. Existing solutions usually rely on the adjustment of network architecture, making the calculation more complicated or introducing additional structural restrictions. In this work, we propose layer-wise multi-view learning to solve this problem, circumventing the necessity to change the model structure. We regard each encoder layer's off-the-shelf output, a by-product in layer-by-layer encoding, as the redundant view for the input sentence. In this way, in addition to the topmost encoder layer (referred to as the primary view), we also incorporate an intermediate encoder layer as the auxiliary view. We feed the two views to a partially shared decoder to maintain independent prediction. Consistency regularization based on KL divergence is used to encourage the two views to learn from each other. Extensive experimental results on five translation tasks show that our approach yields stable improvements over multiple strong baselines. As another bonus, our method is agnostic to network architectures and can maintain the same inference speed as the original model.

arXiv:2011.01210

Focus on the present: a regularization method for the ASR source-target attention layer

This paper introduces a novel method to diagnose the source-target attention in state-of-the-art end-to-end speech recognition models with joint connectionist temporal classification (CTC) and attention training. Our method is based on the fact that both, CTC and source-target attention, are acting on the same encoder representations. To understand the functionality of the attention, CTC is applied to compute the token posteriors given the attention outputs. We found that the source-target attention heads are able to predict several tokens ahead of the current one. Inspired by the observation, a new regularization method is proposed which leverages CTC to make source-target attention more focused on the frames corresponding to the output token being predicted by the decoder. Experiments reveal stable improvements up to 7\% and 13\% relatively with the proposed regularization on TED-LIUM 2 and LibriSpeech.

arXiv:2011.01151

Optimize what matters: Training DNN-HMM Keyword Spotting Model Using End Metric

Deep Neural Network--Hidden Markov Model (DNN-HMM) based methods have been successfully used for many always-on keyword spotting algorithms that detect a wake word to trigger a device. The DNN predicts the state probabilities of a given speech frame, while HMM decoder combines the DNN predictions of multiple speech frames to compute the keyword detection score. The DNN, in prior methods, is trained independent of the HMM parameters to minimize the cross-entropy loss between the predicted and the ground-truth state probabilities. The mis-match between the DNN training loss (cross-entropy) and the end metric (detection score) is the main source of sub-optimal performance for the keyword spotting task. We address this loss-metric mismatch with a novel end-to-end training strategy that learns the DNN parameters by optimizing for the detection score. To this end, we make the HMM decoder (dynamic programming) differentiable and back-propagate through it to maximize the score for the keyword and minimize the scores for non-keyword speech segments. Our method does not require any change in the model architecture or the inference framework; therefore, there is no overhead in run-time memory or compute requirements. Moreover, we show significant reduction in false rejection rate (FRR) at the same false trigger experience (> 70% over independent DNN training).

arXiv:2011.00770

Context-Aware Cross-Attention for Non-Autoregressive Translation

Non-autoregressive translation (NAT) significantly accelerates the inference process by predicting the entire target sequence. However, due to the lack of target dependency modelling in the decoder, the conditional generation process heavily depends on the cross-attention. In this paper, we reveal a localness perception problem in NAT cross-attention, for which it is difficult to adequately capture source context. To alleviate this problem, we propose to enhance signals of neighbour source tokens into conventional cross-attention. Experimental results on several representative datasets show that our approach can consistently improve translation quality over strong NAT baselines. Extensive analyses demonstrate that the enhanced cross-attention achieves better exploitation of source contexts by leveraging both local and global information.

arXiv:2011.00747공식 코드 · Papers with Code

Dual-decoder Transformer for Joint Automatic Speech Recognition and Multilingual Speech Translation

We introduce dual-decoder Transformer, a new model architecture that jointly performs automatic speech recognition (ASR) and multilingual speech translation (ST). Our models are based on the original Transformer architecture (Vaswani et al., 2017) but consist of two decoders, each responsible for one task (ASR or ST). Our major contribution lies in how these decoders interact with each other: one decoder can attend to different information sources from the other via a dual-attention mechanism. We propose two variants of these architectures corresponding to two different levels of dependencies between the decoders, called the parallel and cross dual-decoder Transformers, respectively. Extensive experiments on the MuST-C dataset show that our models outperform the previously-reported highest translation performance in the multilingual settings, and outperform as well bilingual one-to-one results. Furthermore, our parallel models demonstrate no trade-off between ASR and ST compared to the vanilla multi-task architecture. Our code and pre-trained models are available at https://github.com/formiel/speech-translation.

arXiv:2011.00704

Semi-supervised Autoencoding Projective Dependency Parsing

We describe two end-to-end autoencoding models for semi-supervised graph-based projective dependency parsing. The first model is a Locally Autoencoding Parser (LAP) encoding the input using continuous latent variables in a sequential manner; The second model is a Globally Autoencoding Parser (GAP) encoding the input into dependency trees as latent variables, with exact inference. Both models consist of two parts: an encoder enhanced by deep neural networks (DNN) that can utilize the contextual information to encode the input into latent variables, and a decoder which is a generative model able to reconstruct the input. Both LAP and GAP admit a unified structure with different loss functions for labeled and unlabeled data with shared parameters. We conducted experiments on WSJ and UD dependency parsing data sets, showing that our models can exploit the unlabeled data to improve the performance given a limited amount of labeled data, and outperform a previously proposed semi-supervised model.

arXiv:2011.00631

Bifurcated Autoencoder for Segmentation of COVID-19 Infected Regions in CT Images

The new coronavirus infection has shocked the world since early 2020 with its aggressive outbreak. Rapid detection of the disease saves lives, and relying on medical imaging (Computed Tomography and X-ray) to detect infected lungs has shown to be effective. Deep learning and convolutional neural networks have been used for image analysis in this context. However, accurate identification of infected regions has proven challenging for two main reasons. Firstly, the characteristics of infected areas differ in different images. Secondly, insufficient training data makes it challenging to train various machine learning algorithms, including deep-learning models. This paper proposes an approach to segment lung regions infected by COVID-19 to help cardiologists diagnose the disease more accurately, faster, and more manageable. We propose a bifurcated 2-D model for two types of segmentation. This model uses a shared encoder and a bifurcated connection to two separate decoders. One decoder is for segmentation of the healthy region of the lungs, while the other is for the segmentation of the infected regions. Experiments on publically available images show that the bifurcated structure segments infected regions of the lungs better than state of the art.

arXiv:2011.00538공식 코드 · Papers with Code

Deep Diacritization: Efficient Hierarchical Recurrence for Improved Arabic Diacritization

We propose a novel architecture for labelling character sequences that achieves state-of-the-art results on the Tashkeela Arabic diacritization benchmark. The core is a two-level recurrence hierarchy that operates on the word and character levels separately---enabling faster training and inference than comparable traditional models. A cross-level attention module further connects the two, and opens the door for network interpretability. The task module is a softmax classifier that enumerates valid combinations of diacritics. This architecture can be extended with a recurrent decoder that optionally accepts priors from partially diacritized text, which improves results. We employ extra tricks such as sentence dropout and majority voting to further boost the final result. Our best model achieves a WER of 5.34%, outperforming the previous state-of-the-art with a 30.56% relative error reduction.