The use of large pretrained neural networks to create contextualized word embeddings has drastically improved performance on several natural language processing (NLP) tasks. These computationally expensive models have begun to be applied to domain-specific NLP tasks such as re-hospitalization prediction from clinical notes. This paper demonstrates that using large pretrained models produces excellent results on common learning analytics tasks. Pre-training deep language models using student forum data from a wide array of online courses improves performance beyond the state of the art on three text classification tasks. We also show that a smaller, distilled version of our model produces the best results on two of the three tasks while limiting computational cost. We make both models available to the research community at large.
Although deep neural networks generally have fixed network structures, the concept of dynamic mechanism has drawn more and more attention in recent years. Attention mechanisms compute input-dependent dynamic attention weights for aggregating a sequence of hidden states. Dynamic network configuration in convolutional neural networks (CNNs) selectively activates only part of the network at a time for different inputs. In this paper, we combine the two dynamic mechanisms for text classification tasks. Traditional attention mechanisms attend to the whole sequence of hidden states for an input sentence, while in most cases not all attention is needed especially for long sequences. We propose a novel method called Gated Attention Network (GA-Net) to dynamically select a subset of elements to attend to using an auxiliary network, and compute attention weights to aggregate the selected elements. It avoids a significant amount of unnecessary computation on unattended elements, and allows the model to pay attention to important parts of the sequence. Experiments in various datasets show that the proposed method achieves better performance compared with all baseline models with global or local attention while requiring less computation and achieving better interpretability. It is also promising to extend the idea to more complex attention-based models, such as transformers and seq-to-seq models.
Effective representation learning from text has been an active area of research in the fields of NLP and text mining. Attention mechanisms have been at the forefront in order to learn contextual sentence representations. Current state-of-the-art approaches for many NLP tasks use large pre-trained language models such as BERT, XLNet and so on for learning representations. These models are based on the Transformer architecture that involves recurrent blocks of computation consisting of multi-head self-attention and feedforward networks. One of the major bottlenecks largely contributing to the computational complexity of the Transformer models is the self-attention layer, that is both computationally expensive and parameter intensive. In this work, we introduce a novel multi-head self-attention mechanism operating on GRUs that is shown to be computationally cheaper and more parameter efficient than self-attention mechanism proposed in Transformers for text classification tasks. The efficiency of our approach mainly stems from two optimizations; 1) we use low-rank matrix factorization of the affinity matrix to efficiently get multiple attention distributions instead of having separate parameters for each head 2) attention scores are obtained by querying a global context vector instead of densely querying all the words in the sentence. We evaluate the performance of the proposed model on tasks such as sentiment analysis from movie reviews, predicting business ratings from reviews and classifying news articles into topics. We find that the proposed approach matches or outperforms a series of strong baselines and is more parameter efficient than comparable multi-head approaches. We also perform qualitative analyses to verify that the proposed approach is interpretable and captures context-dependent word importance.
In this paper, we introduce the prior knowledge, multi-scale structure, into self-attention modules. We propose a Multi-Scale Transformer which uses multi-scale multi-head self-attention to capture features from different scales. Based on the linguistic perspective and the analysis of pre-trained Transformer (BERT) on a huge corpus, we further design a strategy to control the scale distribution for each layer. Results of three different kinds of tasks (21 datasets) show our Multi-Scale Transformer outperforms the standard Transformer consistently and significantly on small and moderate size datasets.
In this paper, we transform tag recommendation into a word-based text generation problem and introduce a sequence-to-sequence model. The model inherits the advantages of LSTM-based encoder for sequential modeling and attention-based decoder with local positional encodings for learning relations globally. Experimental results on Zhihu datasets illustrate the proposed model outperforms other state-of-the-art text classification based methods.
Graph convolutional networks (GCNs) have shown the powerful ability in text structure representation and effectively facilitate the task of text classification. However, challenges still exist in adapting GCN on learning discriminative features from texts due to the main issue of graph variants incurred by the textual complexity and diversity. In this paper, we propose a dual-attention GCN to model the structural information of various texts as well as tackle the graph-invariant problem through embedding two types of attention mechanisms, i.e. the connection-attention and hop-attention, into the classic GCN. To encode various connection patterns between neighbour words, connection-attention adaptively imposes different weights specified to neighbourhoods of each word, which captures the short-term dependencies. On the other hand, the hop-attention applies scaled coefficients to different scopes during the graph diffusion process to make the model learn more about the distribution of context, which captures long-term semantics in an adaptive way. Extensive experiments are conducted on five widely used datasets to evaluate our dual-attention GCN, and the achieved state-of-the-art performance verifies the effectiveness of dual-attention mechanisms.
Most studies on text classification are focused on the English language. However, short texts such as SMS are influenced by regional languages. This makes the automatic text classification task challenging due to the multilingual, informal, and noisy nature of language in the text. In this work, we propose a novel multi-cascaded deep learning model called McM for bilingual SMS classification. McM exploits $n$-gram level information as well as long-term dependencies of text for learning. Our approach aims to learn a model without any code-switching indication, lexical normalization, language translation, or language transliteration. The model relies entirely upon the text as no external knowledge base is utilized for learning. For this purpose, a 12 class bilingual text dataset is developed from SMS feedbacks of citizens on public services containing mixed Roman Urdu and English languages. Our model achieves high accuracy for classification on this dataset and outperforms the previous model for multilingual text classification, highlighting language independence of McM.
Pre-trained word embeddings encode general word semantics and lexical regularities of natural language, and have proven useful across many NLP tasks, including word sense disambiguation, machine translation, and sentiment analysis, to name a few. In supervised tasks such as multiclass text classification (the focus of this article) it seems appealing to enhance word representations with ad-hoc embeddings that encode task-specific information. We propose (supervised) word-class embeddings (WCEs), and show that, when concatenated to (unsupervised) pre-trained word embeddings, they substantially facilitate the training of deep-learning models in multiclass classification by topic. We show empirical evidence that WCEs yield a consistent improvement in multiclass classification accuracy, using four popular neural architectures and six widely used and publicly available datasets for multiclass text classification. Our code that implements WCEs is publicly available at https://github.com/AlexMoreo/word-class-embeddings
Credit attribution is the task of associating individual parts in a document with their most appropriate class labels. It is an important task with applications to information retrieval and text summarization. When labeled training data is available, traditional approaches for sequence tagging can be used for credit attribution. However, generating such labeled datasets is expensive and time-consuming. In this paper, we present "Credit Attribution With Attention (CAWA)", a neural-network-based approach, that instead of using sentence-level labeled data, uses the set of class labels that are associated with an entire document as a source of distant-supervision. CAWA combines an attention mechanism with a multilabel classifier into an end-to-end learning framework to perform credit attribution. CAWA labels the individual sentences from the input document using the resultant attention-weights. CAWA improves upon the state-of-the-art credit attribution approach by not constraining a sentence to belong to just one class, but modeling each sentence as a distribution over all classes, leading to better modeling of semantically-similar classes. Experiments on the credit attribution task on a variety of datasets show that the sentence class labels generated by CAWA outperform the competing approaches. Additionally, on the multilabel text classification task, CAWA performs better than the competing credit attribution approaches.
Most recent neural semi-supervised learning algorithms rely on adding small perturbation to either the input vectors or their representations. These methods have been successful on computer vision tasks as the images form a continuous manifold, but are not appropriate for discrete input such as sentence. To adapt these methods to text input, we propose to decompose a neural network $M$ into two components $F$ and $U$ so that $M = U\circ F$. The layers in $F$ are then frozen and only the layers in $U$ will be updated during most time of the training. In this way, $F$ serves as a feature extractor that maps the input to high-level representation and adds systematical noise using dropout. We can then train $U$ using any state-of-the-art SSL algorithms such as $\Pi$-model, temporal ensembling, mean teacher, etc. Furthermore, this gradually unfreezing schedule also prevents a pretrained model from catastrophic forgetting. The experimental results demonstrate that our approach provides improvements when compared to state of the art methods especially on short texts.
Recent results show that deep neural networks using contextual embeddings significantly outperform non-contextual embeddings on a majority of text classification task. We offer precomputed embeddings from popular contextual ELMo model for seven languages: Croatian, Estonian, Finnish, Latvian, Lithuanian, Slovenian, and Swedish. We demonstrate that the quality of embeddings strongly depends on the size of training set and show that existing publicly available ELMo embeddings for listed languages shall be improved. We train new ELMo embeddings on much larger training sets and show their advantage over baseline non-contextual FastText embeddings. In evaluation, we use two benchmarks, the analogy task and the NER task.
Text classification is an important field of research from mid 90s up to now. It has many applications, one of them is in Web-based biosurveillance systems which identify and summarize online disease outbreak reports. In this paper we focus on classifying Vietnamese disease outbreak reports. We investigate important properties of disease outbreak reports, e.g., sentences containing names of outbreak disease, locations. Evaluation on 10-time 10- fold cross-validation using the Support Vector Machine algorithm shows that using sentences containing disease outbreak names with its preceding/following sentences in combination with location features achieve the best F-score with 86.67% - an improvement of 0.38% in comparison to using all raw text. Our results suggest that using important sentences and rich feature can improve performance of Vietnamese disease outbreak text classification.
Identifying articles that relate to infectious diseases is a necessary step for any automatic bio-surveillance system that monitors news articles from the Internet. Unlike scientific articles which are available in a strongly structured form, news articles are usually loosely structured. In this chapter, we investigate the importance of each section and the effect of section weighting on performance of text classification. The experimental results show that (1) classification models using the headline and leading sentence achieve a high performance in terms of F-score compared to other parts of the article; (2) all section with bag-of-word representation (full text) achieves the highest recall; and (3) section weighting information can help to improve accuracy.
Previous works related to automatic personality recognition focus on using traditional classification models with linguistic features. However, attentive neural networks with contextual embeddings, which have achieved huge success in text classification, are rarely explored for this task. In this project, we have two major contributions. First, we create the first dialogue-based personality dataset, FriendsPersona, by annotating 5 personality traits of speakers from Friends TV Show through crowdsourcing. Second, we present a novel approach to automatic personality recognition using pre-trained contextual embeddings (BERT and RoBERTa) and attentive neural networks. Our models largely improve the state-of-art results on the monologue Essays dataset by 2.49%, and establish a solid benchmark on our FriendsPersona. By comparing results in two datasets, we demonstrate the challenges of modeling personality in multi-party dialogue.
Recently, large language models such as GPT-2 have shown themselves to be extremely adept at text generation and have also been able to achieve high-quality results in many downstream NLP tasks such as text classification, sentiment analysis and question answering with the aid of fine-tuning. We present a useful technique for using a large language model to perform the task of paraphrasing on a variety of texts and subjects. Our approach is demonstrated to be capable of generating paraphrases not only at a sentence level but also for longer spans of text such as paragraphs without needing to break the text into smaller chunks.
Recurrent neural networks have been widely used in sequence learning tasks. In previous studies, the performance of the model has always been improved by either wider or deeper structures. However, the former becomes more prone to overfitting, while the latter is difficult to optimize. In this paper, we propose a simple new model named Thick-Net, by expanding the network from another dimension: thickness. Multiple parallel values are obtained via more sets of parameters in each hidden state, and the maximum value is selected as the final output among parallel intermediate outputs. Notably, Thick-Net can efficiently avoid overfitting, and is easier to optimize than the vanilla structures due to the large dropout affiliated with it. Our model is evaluated on four sequential tasks including adding problem, permuted sequential MNIST, text classification and language modeling. The results of these tasks demonstrate that our model can not only improve accuracy with faster convergence but also facilitate a better generalization ability.
Language models are at the core of natural language processing. The ability to represent natural language gives rise to its applications in numerous NLP tasks including text classification, summarization, and translation. Research in this area is very limited in Bangla due to the scarcity of resources, except for some count-based models and very recent neural language models being proposed, which are all based on words and limited in practical tasks due to their high perplexity. This paper attempts to approach this issue of perplexity and proposes a subword level neural language model with the AWD-LSTM architecture and various other techniques suitable for training in Bangla language. The model is trained on a corpus of Bangla newspaper articles of an appreciable size consisting of more than 28.5 million word tokens. The performance comparison with various other models depicts the significant reduction in perplexity the proposed model provides, reaching as low as 39.84, in just 20 epochs.
Recently, a lot of techniques were developed to sparsify the weights of neural networks and to remove networks' structure units, e.g. neurons. We adjust the existing sparsification approaches to the gated recurrent architectures. Specifically, in addition to the sparsification of weights and neurons, we propose sparsifying the preactivations of gates. This makes some gates constant and simplifies LSTM structure. We test our approach on the text classification and language modeling tasks. We observe that the resulting structure of gate sparsity depends on the task and connect the learned structure to the specifics of the particular tasks. Our method also improves neuron-wise compression of the model in most of the tasks.
We introduce a training method for both better word representation and performance, which we call GROVER (Gradual Rumination On the Vector with maskERs). The method is to gradually and iteratively add random noises to word embeddings while training a model. GROVER first starts from conventional training process, and then extracts the fine-tuned representations. Next, we gradually add random noises to the word representations and repeat the training process from scratch, but initialize with the noised word representations. Through the re-training process, we can mitigate some noises to be compensated and utilize other noises to learn better representations. As a result, we can get word representations further fine-tuned and specialized on the task. When we experiment with our method on 5 text classification datasets, our method improves model performances on most of the datasets. Moreover, we show that our method can be combined with other regularization techniques, further improving the model performance.
Leveraging weak or noisy supervision for building effective machine learning models has long been an important research problem. Its importance has further increased recently due to the growing need for large-scale datasets to train deep learning models. Weak or noisy supervision could originate from multiple sources including non-expert annotators or automatic labeling based on heuristics or user interaction signals. There is an extensive amount of previous work focusing on leveraging noisy labels. Most notably, recent work has shown impressive gains by using a meta-learned instance re-weighting approach where a meta-learning framework is used to assign instance weights to noisy labels. In this paper, we extend this approach via posing the problem as label correction problem within a meta-learning framework. We view the label correction procedure as a meta-process and propose a new meta-learning based framework termed MLC (Meta Label Correction) for learning with noisy labels. Specifically, a label correction network is adopted as a meta-model to produce corrected labels for noisy labels while the main model is trained to leverage the corrected labeled. Both models are jointly trained by solving a bi-level optimization problem. We run extensive experiments with different label noise levels and types on both image recognition and text classification tasks. We compare the reweighing and correction approaches showing that the correction framing addresses some of the limitation of reweighting. We also show that the proposed MLC approach achieves large improvements over previous methods in many settings.