Deep learning models are often trained on datasets that contain sensitive information such as individuals' shopping transactions, personal contacts, and medical records. An increasingly important line of work therefore has sought to train neural networks subject to privacy constraints that are specified by differential privacy or its divergence-based relaxations. These privacy definitions, however, have weaknesses in handling certain important primitives (composition and subsampling), thereby giving loose or complicated privacy analyses of training neural networks. In this paper, we consider a recently proposed privacy definition termed \textit{$f$-differential privacy} [18] for a refined privacy analysis of training neural networks. Leveraging the appealing properties of $f$-differential privacy in handling composition and subsampling, this paper derives analytically tractable expressions for the privacy guarantees of both stochastic gradient descent and Adam used in training deep neural networks, without the need of developing sophisticated techniques as [3] did. Our results demonstrate that the $f$-differential privacy framework allows for a new privacy analysis that improves on the prior analysis~[3], which in turn suggests tuning certain parameters of neural networks for a better prediction accuracy without violating the privacy budget. These theoretically derived improvements are confirmed by our experiments in a range of tasks in image classification, text classification, and recommender systems. Python code to calculate the privacy cost for these experiments is publicly available in the \texttt{TensorFlow Privacy} library.
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.
Log messages are now widely used in software systems. They are important for classification as millions of logs are generated each day. Most logs are unstructured which makes classification a challenge. In this paper, Deep Learning (DL) methods called Auto-LSTM, Auto-BLSTM and Auto-GRU are developed for anomaly detection and log classification. These models are used to convert unstructured log data to extracted features which is suitable for classification algorithms. They are evaluated using four data sets, namely BGL, Openstack, Thunderbird and IMDB. The first three are popular log data sets while the fourth is a movie review data set which is used for sentiment classification. The results obtained show that Auto-LSTM, Auto-BLSTM and Auto-GRU perform better than other well-known algorithms.
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.
A recently introduced classifier, called SS3, has shown to be well suited to deal with early risk detection (ERD) problems on text streams. It obtained state-of-the-art performance on early depression and anorexia detection on Reddit in the CLEF's eRisk open tasks. SS3 was created to deal with ERD problems naturally since: it supports incremental training and classification over text streams, and it can visually explain its rationale. However, SS3 processes the input using a bag-of-word model lacking the ability to recognize important word sequences. This aspect could negatively affect the classification performance and also reduces the descriptiveness of visual explanations. In the standard document classification field, it is very common to use word n-grams to try to overcome some of these limitations. Unfortunately, when working with text streams, using n-grams is not trivial since the system must learn and recognize which n-grams are important "on the fly". This paper introduces t-SS3, an extension of SS3 that allows it to recognize useful patterns over text streams dynamically. We evaluated our model in the eRisk 2017 and 2018 tasks on early depression and anorexia detection. Experimental results suggest that t-SS3 is able to improve both current results and the richness of visual explanations.
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.
Adversarial training (AT) as a regularization method has proved its effectiveness in various tasks, such as image classification and text classification. Though there are successful applications of AT in many tasks of natural language processing (NLP), the mechanism behind it is still unclear. In this paper, we aim to apply AT on machine reading comprehension (MRC) and study its effects from multiple perspectives. We experiment with three different kinds of RC tasks: span-based RC, span-based RC with unanswerable questions and multi-choice RC. The experimental results show that the proposed method can improve the performance significantly and universally on SQuAD1.1, SQuAD2.0 and RACE. With virtual adversarial training (VAT), we explore the possibility of improving the RC models with semi-supervised learning and prove that examples from a different task are also beneficial. We also find that AT helps little in defending against artificial adversarial examples, but AT helps the model to learn better on examples that contain more low-frequency words.
We propose a novel framework to understand the text by converting sentences or articles into video-like 3-dimensional tensors. Each frame, corresponding to a slice of the tensor, is a word image that is rendered by the word's shape. The length of the tensor equals to the number of words in the sentence or article. The proposed transformation from the text to a 3-dimensional tensor makes it very convenient to implement an $n$-gram model with convolutional neural networks for text analysis. Concretely, we impose a 3-dimensional convolutional kernel on the 3-dimensional text tensor. The first two dimensions of the convolutional kernel size equal the size of the word image and the last dimension of the kernel size is $n$. That is, every time when we slide the 3-dimensional kernel over a word sequence, the convolution covers $n$ word images and outputs a scalar. By iterating this process continuously for each $n$-gram along with the sentence or article with multiple kernels, we obtain a 2-dimensional feature map. A subsequent 1-dimensional max-over-time pooling is applied to this feature map, and three fully-connected layers are used for conducting text classification finally. Experiments of several text classification datasets demonstrate surprisingly superior performances using the proposed model in comparison with existing methods.
The Transformer model is widely successful on many natural language processing tasks. However, the quadratic complexity of self-attention limit its application on long text. In this paper, adopting a fine-to-coarse attention mechanism on multi-scale spans via binary partitioning (BP), we propose BP-Transformer (BPT for short). BPT yields $O(k\cdot n\log (n/k))$ connections where $k$ is a hyperparameter to control the density of attention. BPT has a good balance between computation complexity and model capacity. A series of experiments on text classification, machine translation and language modeling shows BPT has a superior performance for long text than previous self-attention models. Our code, hyperparameters and CUDA kernels for sparse attention are available in PyTorch.
Large pre-trained sentence encoders like BERT start a new chapter in natural language processing. A common practice to apply pre-trained BERT to sequence classification tasks (e.g., classification of sentences or sentence pairs) is by feeding the embedding of [CLS] token (in the last layer) to a task-specific classification layer, and then fine tune the model parameters of BERT and classifier jointly. In this paper, we conduct systematic analysis over several sequence classification datasets to examine the embedding values of [CLS] token before the fine tuning phase, and present the biased embedding distribution issue---i.e., embedding values of [CLS] concentrate on a few dimensions and are non-zero centered. Such biased embedding brings challenge to the optimization process during fine-tuning as gradients of [CLS] embedding may explode and result in degraded model performance. We further propose several simple yet effective normalization methods to modify the [CLS] embedding during the fine-tuning. Compared with the previous practice, neural classification model with the normalized embedding shows improvements on several text classification tasks, demonstrates the effectiveness of our method.