We propose a unified neural network architecture and learning algorithm that can be applied to various natural language processing tasks including part-of-speech tagging, chunking, named entity recognition, and semantic role labeling. This versatility is achieved by trying to avoid task-specific engineering and therefore disregarding a lot of prior knowledge. Instead of exploiting man-made input features carefully optimized for each task, our system learns internal representations on the basis of vast amounts of mostly unlabeled training data. This work is then used as a basis for building a freely available tagging system with good performance and minimal computational requirements.

/pdf/natural-language-processing-almost-from-scratch-osj5vem4f3.pdf

Natural Language Processing (Almost) from Scratch

We introduce a new representation learning approach for domain adaptation, in which data at training and test time come from similar but different distributions. Our approach is directly inspired by the theory on domain adaptation suggesting that, for effective domain transfer to be achieved, predictions must be made based on features that cannot discriminate between the training (source) and test (target) domains.

The approach implements this idea in the context of neural network architectures that are trained on labeled data from the source domain and unlabeled data from the target domain (no labeled target-domain data is necessary). As the training progresses, the approach promotes the emergence of features that are (i) discriminative for the main learning task on the source domain and (ii) indiscriminate with respect to the shift between the domains. We show that this adaptation behaviour can be achieved in almost any feed-forward model by augmenting it with few standard layers and a new gradient reversal layer. The resulting augmented architecture can be trained using standard backpropagation and stochastic gradient descent, and can thus be implemented with little effort using any of the deep learning packages.

We demonstrate the success of our approach for two distinct classification problems (document sentiment analysis and image classification), where state-of-the-art domain adaptation performance on standard benchmarks is achieved. We also validate the approach for descriptor learning task in the context of person re-identification application.

/pdf/domain-adversarial-training-of-neural-networks-53puwk7wkk.pdf

Domain-adversarial training of neural networks

If we take an existing supervised NLP system, a simple and general way to improve accuracy is to use unsupervised word representations as extra word features. We evaluate Brown clusters, Collobert and Weston (2008) embeddings, and HLBL (Mnih & Hinton, 2009) embeddings of words on both NER and chunking. We use near state-of-the-art supervised baselines, and find that each of the three word representations improves the accuracy of these baselines. We find further improvements by combining different word representations. You can download our word features, for off-the-shelf use in existing NLP systems, as well as our code, here: http://metaoptimize.com/projects/wordreprs/

/pdf/word-representations-a-simple-and-general-method-for-semi-1096uffw40.pdf

Word Representations: A Simple and General Method for Semi-Supervised Learning

Two problems arise when using distant supervision for relation extraction. First, in this method, an already existing knowledge base is heuristically aligned to texts, and the alignment results are treated as labeled data. However, the heuristic alignment can fail, resulting in wrong label problem. In addition, in previous approaches, statistical models have typically been applied to ad hoc features. The noise that originates from the feature extraction process can cause poor performance. In this paper, we propose a novel model dubbed the Piecewise Convolutional Neural Networks (PCNNs) with multi-instance learning to address these two problems. To solve the first problem, distant supervised relation extraction is treated as a multi-instance problem in which the uncertainty of instance labels is taken into account. To address the latter problem, we avoid feature engineering and instead adopt convolutional architecture with piecewise max pooling to automatically learn relevant features. Experiments show that our method is effective and outperforms several competitive baseline methods.

/pdf/distant-supervision-for-relation-extraction-via-piecewise-2zx0d5jsk9.pdf

Distant Supervision for Relation Extraction via Piecewise Convolutional Neural Networks

We propose a unified neural network architecture and learning algorithm that can be applied to various natural language processing tasks including: part-of-speech tagging, chunking, named entity recognition, and semantic role labeling. This versatility is achieved by trying to avoid task-specific engineering and therefore disregarding a lot of prior knowledge. Instead of exploiting man-made input features carefully optimized for each task, our system learns internal representations on the basis of vast amounts of mostly unlabeled training data. This work is then used as a basis for building a freely available tagging system with good performance and minimal computational requirements.

Natural Language Processing (almost) from Scratch

Supervised sequence-labeling systems in natural language processing often suffer from data sparsity because they use word types as features in their prediction tasks. Consequently, they have difficulty estimating parameters for types which appear in the test set, but seldom (or never) appear in the training set. We demonstrate that distributional representations of word types, trained on unannotated text, can be used to improve performance on rare words. We incorporate aspects of these representations into the feature space of our sequence-labeling systems. In an experiment on a standard chunking dataset, our best technique improves a chunker from 0.76 F1 to 0.86 F1 on chunks beginning with rare words. On the same dataset, it improves our part-of-speech tagger from 74% to 80% accuracy on rare words. Furthermore, our system improves significantly over a baseline system when applied to text from a different domain, and it reduces the sample complexity of sequence labeling.

Distributional Representations for Handling Sparsity in Supervised Sequence-Labeling

Finding the right representations for words is critical for building accurate NLP systems when domain-specific labeled data for the task is scarce. This article investigates novel techniques for extracting features from n-gram models, Hidden Markov Models, and other statistical language models, including a novel Partial Lattice Markov Random Field model. Experiments on part-of-speech tagging and information extraction, among other tasks, indicate that features taken from statistical language models, in combination with more traditional features, outperform traditional representations alone, and that graphical model representations outperform n-gram models, especially on sparse and polysemous words.

Learning representations for weakly supervised natural language processing tasks

Most supervised language processing systems show a significant drop-off in performance when they are tested on text that comes from a domain significantly different from the domain of the training data. Sequence labeling systems like part-of-speech taggers are typically trained on newswire text, and in tests their error rate on, for example, biomedical data can triple, or worse. We investigate techniques for building open-domain sequence labeling systems that approach the ideal of a system whose accuracy is high and constant across domains. In particular, we investigate unsupervised techniques for representation learning that provide new features which are stable across domains, in that they are predictive in both the training and out-of-domain test data. In experiments, our novel techniques reduce error by as much as 29% relative to the previous state of the art on out-of-domain text.

Exploring Representation-Learning Approaches to Domain Adaptation

Most information extraction research identifies the state of the world in text, including the entities and the relationships that exist between them. Much less attention has been paid to the understanding of dynamics, or how the state of the world changes over time. Because intelligent behavior seeks to change the state of the world in rational and utilitymaximizing ways, common-sense knowledge about dynamics is essential for intelligent agents. In this paper, we describe a novel system, PREPOST, that tackles the problem of extracting the preconditions and effects of actions and events, two important kinds of knowledge for connecting world state and the actions that affect it. In experiments on Web text, PREPOST is able to improve by 79% over a baseline technique for identifying the effects of actions (64% improvement for preconditions).

/pdf/extracting-action-and-event-semantics-from-web-text-3bibwheqg1.pdf

Extracting Action and Event Semantics from Web Text

Most supervised language processing systems show a significant drop-off in performance when they are tested on text that comes from a domain significantly different from the domain of the training data. Semantic role labeling techniques are typically trained on newswire text, and in tests their performance on fiction is as much as 19% worse than their performance on newswire text. We investigate techniques for building open-domain semantic role labeling systems that approach the ideal of a train-once, use-anywhere system. We leverage recently-developed techniques for learning representations of text using latent-variable language models, and extend these techniques to ones that provide the kinds of features that are useful for semantic role labeling. In experiments, our novel system reduces error by 16% relative to the previous state of the art on out-of-domain text.

Fei Huang

Papers

Distributional Representations for Handling Sparsity in Supervised Sequence-Labeling

Learning representations for weakly supervised natural language processing tasks

Exploring Representation-Learning Approaches to Domain Adaptation

Extracting Action and Event Semantics from Web Text

Open-Domain Semantic Role Labeling by Modeling Word Spans