Character-level Language Model Research Articles

Improving Deep Learning based Automatic Speech Recognition for Gujarati

We present a novel approach for improving the performance of an End-to-End speech recognition system for the Gujarati language. We follow a deep learning-based approach that includes Convolutional Neural Network, Bi-directional Long Short Term Memory layers, Dense layers, and Connectionist Temporal Classification as a loss function. To improve the performance of the system with the limited size of the dataset, we present a combined language model (Word-level language Model and Character-level language model)-based prefix decoding technique and Bidirectional Encoder Representations from Transformers-based post-processing technique. To gain key insights from our Automatic Speech Recognition (ASR) system, we used the inferences from the system and proposed different analysis methods. These insights help us in understanding and improving the ASR system as well as provide intuition into the language used for the ASR system. We have trained the model on the Microsoft Speech Corpus, and we observe a 5.87% decrease in Word Error Rate (WER) with respect to base-model WER.

GapPredict - A Language Model for Resolving Gaps in Draft Genome Assemblies.

Short-read DNA sequencing instruments can yield over 1012 bases per run, typically composed of reads 150 bases long. Despite this high throughput, de novo assembly algorithms have difficulty reconstructing contiguous genome sequences using short reads due to both repetitive and difficult-to-sequence regions in these genomes. Some of the short read assembly challenges are mitigated by scaffolding assembled sequences using paired-end reads. However, unresolved sequences in these scaffolds appear as “gaps”. Here, we introduce GapPredict – An implementation of a proof of concept that uses a character-level language model to predict unresolved nucleotides in scaffold gaps. We benchmarked GapPredict against the state-of-the-art gap-filling tool Sealer, and observed that the former can fill 65.6% of the sampled gaps that were left unfilled by the latter with high similarity to the reference genome, demonstrating the practical utility of deep learning approaches to the gap-filling problem in genome assembly.

HunFlair: an easy-to-use tool for state-of-the-art biomedical named entity recognition.

SummaryNamed entity recognition (NER) is an important step in biomedical information extraction pipelines. Tools for NER should be easy to use, cover multiple entity types, be highly accurate and be robust toward variations in text genre and style. We present HunFlair, a NER tagger fulfilling these requirements. HunFlair is integrated into the widely used NLP framework Flair, recognizes five biomedical entity types, reaches or overcomes state-of-the-art performance on a wide set of evaluation corpora, and is trained in a cross-corpus setting to avoid corpus-specific bias. Technically, it uses a character-level language model pretrained on roughly 24 million biomedical abstracts and three million full texts. It outperforms other off-the-shelf biomedical NER tools with an average gain of 7.26 pp over the next best tool in a cross-corpus setting and achieves on-par results with state-of-the-art research prototypes in in-corpus experiments. HunFlair can be installed with a single command and is applied with only four lines of code. Furthermore, it is accompanied by harmonized versions of 23 biomedical NER corpora.Availability and implementationHunFlair ist freely available through the Flair NLP framework (https://github.com/flairNLP/flair) under an MIT license and is compatible with all major operating systems.Supplementary informationSupplementary data are available at Bioinformatics online.

Bicleaner at WMT 2020: Universitat d'Alacant-Prompsit's submission to the parallel corpus filtering shared task

This paper describes the joint submission of Universitat d’Alacant and Prompsit Language Engineering to the WMT 2020 shared task on parallel corpus filtering. Our submission, based on the free/open-source tool Bicleaner, enhances it with Extremely Randomised Trees and lexical similarity features that account for the frequency of the words in the parallel sentences to determine if two sentences are parallel. To train this classifier we used the clean corpora provided for the task and synthetic noisy parallel sentences. In addition we re-score the output of Bicleaner using character-level language models and n-gram saturation.

Learning molecular dynamics with simple language model built upon long short-term memory neural network

Recurrent neural networks have led to breakthroughs in natural language processing and speech recognition. Here we show that recurrent networks, specifically long short-term memory networks can also capture the temporal evolution of chemical/biophysical trajectories. Our character-level language model learns a probabilistic model of 1-dimensional stochastic trajectories generated from higher-dimensional dynamics. The model captures Boltzmann statistics and also reproduces kinetics across a spectrum of timescales. We demonstrate how training the long short-term memory network is equivalent to learning a path entropy, and that its embedding layer, instead of representing contextual meaning of characters, here exhibits a nontrivial connectivity between different metastable states in the underlying physical system. We demonstrate our model’s reliability through different benchmark systems and a force spectroscopy trajectory for multi-state riboswitch. We anticipate that our work represents a stepping stone in the understanding and use of recurrent neural networks for understanding the dynamics of complex stochastic molecular systems.

Multi-Zone Unit for Recurrent Neural Networks

Recurrent neural networks (RNNs) have been widely used to deal with sequence learning problems. The input-dependent transition function, which folds new observations into hidden states to sequentially construct fixed-length representations of arbitrary-length sequences, plays a critical role in RNNs. Based on single space composition, transition functions in existing RNNs often have difficulty in capturing complicated long-range dependencies. In this paper, we introduce a new Multi-zone Unit (MZU) for RNNs. The key idea is to design a transition function that is capable of modeling multiple space composition. The MZU consists of three components: zone generation, zone composition, and zone aggregation. Experimental results on multiple datasets of the character-level language modeling task and the aspect-based sentiment analysis task demonstrate the superiority of the MZU.

TabulaNearlyRasa:Probing the Linguistic Knowledge of Character-level Neural Language Models Trained on Unsegmented Text

Recurrent neural networks (RNNs) have reached striking performance in many natural language processing tasks. This has renewed interest in whether these generic sequence processing devices are inducing genuine linguistic knowledge. Nearly all current analytical studies, however, initialize the RNNs with a vocabulary of known words, and feed them tokenized input during training. We present a multi-lingual study of the linguistic knowledge encoded in RNNs trained as character-level language models, on input data with word boundaries removed. These networks face a tougher and more cognitively realistic task, having to discover any useful linguistic unit from scratch based on input statistics. The results show that our “near tabula rasa” RNNs are mostly able to solve morphological, syntactic and semantic tasks that intuitively presuppose word-level knowledge, and indeed they learned, to some extent, to track word boundaries. Our study opens the door to speculations about the necessity of an explicit, rigid word lexicon in language learning and usage.

Character-Level Language Modeling with Deeper Self-Attention

LSTMs and other RNN variants have shown strong performance on character-level language modeling. These models are typically trained using truncated backpropagation through time, and it is common to assume that their success stems from their ability to remember long-term contexts. In this paper, we show that a deep (64-layer) transformer model (Vaswani et al. 2017) with fixed context outperforms RNN variants by a large margin, achieving state of the art on two popular benchmarks: 1.13 bits per character on text8 and 1.06 on enwik8. To get good results at this depth, we show that it is important to add auxiliary losses, both at intermediate network layers and intermediate sequence positions.

Dual Rectified Linear Units (DReLUs): A replacement for tanh activation functions in Quasi-Recurrent Neural Networks

In this paper, we introduce a novel type of Rectified Linear Unit (ReLU), called a Dual Rectified Linear Unit (DReLU). A DReLU, which comes with an unbounded positive and negative image, can be used as a drop-in replacement for a tanh activation function in the recurrent step of Quasi-Recurrent Neural Networks (QRNNs) [1]. Similar to ReLUs, DReLUs are less prone to the vanishing gradient problem, they are noise robust, and they induce sparse activations.We independently reproduce the QRNN experiments of Bradbury et al. [1] and compare our DReLU-based QRNNs with the original tanh-based QRNNs and Long Short-Term Memory networks (LSTMs) on sentiment classification and word-level language modeling. Additionally, we evaluate on character-level language modeling, showing that we are able to stack up to eight QRNN layers with DReLUs, thus making it possible to improve the current state-of-the-art in character-level language modeling over shallow architectures based on LSTMs.

Hierarchical recurrent highway networks

A contemporary approach for acquiring the computational gains of depth in recurrent neural networks (RNNs) is to hierarchically stack multiple recurrent layers. However, such performance gains come with the cost of challenging optimization of hierarchal RNNs (HRNNs) which are deep both hierarchically and temporally. Though the researchers have exclusively highlighted the significance of using highways (or direct shortcuts) for learning deep hierarchical representations and deep temporal dependencies. However, no significant efforts are made to unify these finding into a single framework for learning deep HRNNs. We propose hierarchical recurrent highway network (HRHN) that contains highway within the hierarchical and temporal structure of the network for unimpeded information propagation across both dimensions, thus alleviating gradient vanishing problem. The proposed HRHN contain significantly reduced data-dependent parameters as compared to related methods. The experiments on language modeling (LM) tasks have demonstrated that the proposed architecture leads to design effective models. On character-level LM using Hutter prize dataset, the model achieved entropy of 2.4 bits per character. The attained perplexity on world-level LM using Penn TreeBank is 68.1. The HRHN outperformed other baseline and related models that we tested.

Novel architecture for long short-term memory used in question classification

Abstract Long short-term memory (LSTM) is an extension of the recurrent neural network (RNN) and has achieved excellent performance in various tasks, especially sequential problems. The LSTM is chain-structured, and its architecture is limited by sequential information propagation. In practice, it is hard to solve the problems involving very long term dependencies. Recent studies have indicated that adding recurrent skip connections across multiple timescales may help the RNN improve its performance in long-term dependencies. Moreover, capturing local features can improve the performance of the RNN. In this paper, we propose a novel architecture (Att-LSTM) for the LSTM, which connects continuous hidden states of previous time steps to the current time step and applies an attention mechanism to these hidden states. This architecture can not only capture local features effectively but also help learn very long-distance correlations in an input sequence. We evaluate Att-LSTM in various sequential tasks, such as adding problem, sequence classification, and character-level language modeling. In addition, to prove the generalization and practicality of the novel architecture, we design a character-level hierarchical Att-LSTM and refine the word representation with a highway network. This hierarchical model achieved excellent performance on question classification.

Syllable language models for Mandarin speech recognition: Exploiting character language models

Mandarin Chinese is based on characters which are syllabic in nature and morphological in meaning. All spoken languages have syllabiotactic rules which govern the construction of syllables and their allowed sequences. These constraints are not as restrictive as those learned from word sequences, but they can provide additional useful linguistic information. Hence, it is possible to improve speech recognition performance by appropriately combining these two types of constraints. For the Chinese language considered in this paper, character level language models (LMs) can be used as a first level approximation to allowed syllable sequences. To test this idea, word and character level n-gram LMs were trained on 2.8 billion words (equivalent to 4.3 billion characters) of texts from a wide collection of text sources. Both hypothesis and model based combination techniques were investigated to combine word and character level LMs. Significant character error rate reductions up to 7.3% relative were obtained on a state-of-the-art Mandarin Chinese broadcast audio recognition task using an adapted history dependent multi-level LM that performs a log-linearly combination of character and word level LMs. This supports the hypothesis that character or syllable sequence models are useful for improving Mandarin speech recognition performance.