Joint CTC-Attention based End-to-End Speech Recognition using Multi-task Learning

Recently, there has been an increasing interest in end-to-end speech recognition that directly transcribes speech to text without any predefined alignments. One approach is the attention-based encoder-decoder framework that learns a mapping between variable-length input and output sequences in one step using a purely data-driven method. The attention model has often been shown to improve the performance over another end-to-end approach, the Connectionist Temporal Classification (CTC), mainly because it explicitly uses the history of the target character without any conditional independence assumptions. However, we observed that the attention model has shown poor results especially in noisy condition and is hard to be trained in the initial training stage with long input sequences, as compared with CTC. This is because the attention model is too flexible to predict proper alignments in such cases due to the lack of left-to-right constraints as used in CTC. This paper presents a novel method for end-to-end speech recognition to improve robustness and achieve fast convergence by using a joint CTC-attention model within the multi-task learning framework, thereby mitigating the alignment issue. An experiment on the WSJ and CHiME-4 tasks demonstrates its advantages over both the CTC and attention-based encoder-decoder baselines, showing 6.6-10.3% relative improvements in Character Error Rate (CER).


Similar Publications

In this paper, we focus on learning structure-aware document representations from data without recourse to a discourse parser or additional annotations. Drawing inspiration from recent efforts to empower neural networks with a structural bias, we propose a model that can encode a document while automatically inducing rich structural dependencies. Specifically, we embed a differentiable non-projective parsing algorithm into a neural model and use attention mechanisms to incorporate the structural biases. Read More


We introduce a neural network that represents sentences by composing their words according to induced binary parse trees. We use Tree-LSTM as our composition function, applied along a tree structure found by a fully differentiable natural language chart parser. Our model simultaneously optimises both the composition function and the parser, thus eliminating the need for externally-provided parse trees which are normally required for Tree-LSTM. Read More


Recognizing textual entailment is a fundamental task in a variety of text mining or natural language processing applications. This paper proposes a simple neural model for RTE problem. It first matches each word in the hypothesis with its most-similar word in the premise, producing an augmented representation of the hypothesis conditioned on the premise as a sequence of word pairs. Read More


The design of neural architectures for structured objects is typically guided by experimental insights rather than a formal process. In this work, we appeal to kernels over combinatorial structures, such as sequences and graphs, to derive appropriate neural operations. We introduce a class of deep recurrent neural operations and formally characterize their associated kernel spaces. Read More


This paper formulates a novel problem on graphs: find the minimal subset of edges in a fully connected graph, such that the resulting graph contains all spanning trees for a set of specifed sub-graphs. This formulation is motivated by an un-supervised grammar induction problem from computational linguistics. We present a reduction to some known problems and algorithms from graph theory, provide computational complexity results, and describe an approximation algorithm. Read More


We introduce a technique for augmenting neural text-to-speech (TTS) with lowdimensional trainable speaker embeddings to generate different voices from a single model. As a starting point, we show improvements over the two state-ofthe-art approaches for single-speaker neural TTS: Deep Voice 1 and Tacotron. We introduce Deep Voice 2, which is based on a similar pipeline with Deep Voice 1, but constructed with higher performance building blocks and demonstrates a significant audio quality improvement over Deep Voice 1. Read More


The number of word forms in agglutinative languages is theoretically infinite and this variety in word forms introduces sparsity in many natural language processing tasks. Part-of-speech tagging (PoS tagging) is one of these tasks that often suffers from sparsity. In this paper, we present an unsupervised Bayesian model using Hidden Markov Models (HMMs) for joint PoS tagging and stemming for agglutinative languages. Read More


Syntactic parsing is a key task in natural language processing which has been dominated by symbolic, grammar-based syntactic parsers. Neural networks, with their distributed representations, are challenging these methods. In this paper, we want to show that existing parsing algorithms can cross the border and be defined over distributed representations. Read More


This paper is a deep investigation of cross-language plagiarism detection methods on a new recently introduced open dataset, which contains parallel and comparable collections of documents with multiple characteristics (different genres, languages and sizes of texts). We investigate cross-language plagiarism detection methods for 6 language pairs on 2 granularities of text units in order to draw robust conclusions on the best methods while deeply analyzing correlations across document styles and languages. Read More


In this work, we present the Grounded Recurrent Neural Network (GRNN), a recurrent neural network architecture for multi-label prediction which explicitly ties labels to specific dimensions of the recurrent hidden state (we call this process "grounding"). The approach is particularly well-suited for extracting large numbers of concepts from text. We apply the new model to address an important problem in healthcare of understanding what medical concepts are discussed in clinical text. Read More