| Total: 95
In sentence classification tasks, additional contexts, such as the neighboring sentences, may improve the accuracy of the classifier. However, such contexts are domain-dependent and thus cannot be used for another classification task with an inappropriate domain. In contrast, we propose the use of translated sentences as domain-free context that is always available regardless of the domain. We find that naive feature expansion of translations gains only marginal improvements and may decrease the performance of the classifier, due to possible inaccurate translations thus producing noisy sentence vectors. To this end, we present multiple context fixing attachment (MCFA), a series of modules attached to multiple sentence vectors to fix the noise in the vectors using the other sentence vectors as context. We show that our method performs competitively compared to previous models, achieving best classification performance on multiple data sets. We are the first to use translations as domain-free contexts for sentence classification.
The Web and its Semantic extension (i.e. Linked Open Data) contain open global-scale knowledge and make it available to potentially intelligent machines that want to benefit from it. Nevertheless, most of Linked Open Data lack ontological distinctions and have sparse axiomatisation. For example, distinctions such as whether an entity is inherently a class or an individual, or whether it is a physical object or not, are hardly expressed in the data, although they have been largely studied and formalised by foundational ontologies (e.g. DOLCE, SUMO). These distinctions belong to common sense too, which is relevant for many artificial intelligence tasks such as natural language understanding, scene recognition, and the like. There is a gap between foundational ontologies, that often formalise or are inspired by pre-existing philosophical theories and are developed with a top-down approach, and Linked Open Data that mostly derive from existing databases or crowd-based effort (e.g. DBpedia, Wikidata). We investigate whether machines can learn foundational distinctions over Linked Open Data entities, and if they match common sense. We want to answer questions such as “does the DBpedia entity for dog refer to a class or to an instance?”. We report on a set of experiments based on machine learning and crowdsourcing that show promising results.
Distributed word embeddings have shown superior performances in numerous Natural Language Processing (NLP) tasks. However, their performances vary significantly across different tasks, implying that the word embeddings learnt by those methods capture complementary aspects of lexical semantics. Therefore, we believe that it is important to combine the existing word embeddings to produce more accurate and complete meta-embeddings of words. For this purpose, we propose an unsupervised locally linear meta-embedding learning method that takes pre-trained word embeddings as the input, and produces more accurate meta embeddings. Unlike previously proposed meta-embedding learning methods that learn a global projection over all words in a vocabulary, our proposed method is sensitive to the differences in local neighbourhoods of the individual source word embeddings. Moreover, we show that vector concatenation, a previously proposed highly competitive baseline approach for integrating word embeddings, can be derived as a special case of the proposed method. Experimental results on semantic similarity, word analogy, relation classification, and short-text classification tasks show that our meta-embeddings to significantly outperform prior methods in several benchmark datasets, establishing a new state of the art for meta-embeddings.
Machine translation is going through a radical revolution, driven by the explosive development of deep learning techniques using Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN). In this paper, we consider a special case in machine translation problems, targeting to convert natural language into Structured Query Language (SQL) for data retrieval over relational database. Although generic CNN and RNN learn the grammar structure of SQL when trained with sufficient samples, the accuracy and training efficiency of the model could be dramatically improved, when the translation model is deeply integrated with the grammar rules of SQL. We present a new encoder-decoder framework, with a suite of new approaches, including new semantic features fed into the encoder, grammar-aware states injected into the memory of decoder, as well as recursive state management for sub-queries. These techniques help the neural network better focus on understanding semantics of operations in natural language and save the efforts on SQL grammar learning. The empirical evaluation on real world database and queries show that our approach outperform state-of-the-art solution by a significant margin.
Embeddings of medical concepts such as medication, procedure and diagnosis codes in Electronic Medical Records (EMRs) are central to healthcare analytics. Previous work on medical concept embedding takes medical concepts and EMRs as words and documents respectively. Nevertheless, such models miss out the temporal nature of EMR data. On the one hand, two consecutive medical concepts do not indicate they are temporally close, but the correlations between them can be revealed by the time gap. On the other hand, the temporal scopes of medical concepts often vary greatly (e.g., common cold and diabetes). In this paper, we propose to incorporate the temporal information to embed medical codes. Based on the Continuous Bag-of-Words model, we employ the attention mechanism to learn a ``soft'' time-aware context window for each medical concept. Experiments on public and proprietary datasets through clustering and nearest neighbour search tasks demonstrate the effectiveness of our model, showing that it outperforms five state-of-the-art baselines.
Inspired by the observation that word embeddings exhibit isomorphic structure across languages, we propose a novel method to induce a bilingual lexicon from only two sets of word embeddings, which are trained on monolingual source and target data respectively. This is achieved by formulating the task as point set registration which is a more general problem. We show that a transformation from the source to the target embedding space can be learned automatically without any form of cross-lingual supervision. By properly adapting a traditional point set registration model to make it be suitable for processing word embeddings, we achieved state-of-the-art performance on the unsupervised bilingual lexicon induction task. The point set registration problem has been well-studied and can be solved by many elegant models, we thus opened up a new opportunity to capture the universal lexical semantic structure across languages.
Multilingual knowledge graph (KG) embeddings provide latent semantic representations of entities and structured knowledge with cross-lingual inferences, which benefit various knowledge-driven cross-lingual NLP tasks. However, precisely learning such cross-lingual inferences is usually hindered by the low coverage of entity alignment in many KGs. Since many multilingual KGs also provide literal descriptions of entities, in this paper, we introduce an embedding-based approach which leverages a weakly aligned multilingual KG for semi-supervised cross-lingual learning using entity descriptions. Our approach performs co-training of two embedding models, i.e. a multilingual KG embedding model and a multilingual literal description embedding model. The models are trained on a large Wikipedia-based trilingual dataset where most entity alignment is unknown to training. Experimental results show that the performance of the proposed approach on the entity alignment task improves at each iteration of co-training, and eventually reaches a stage at which it significantly surpasses previous approaches. We also show that our approach has promising abilities for zero-shot entity alignment, and cross-lingual KG completion.
Recursive neural network (RvNN) has been proved to be an effective and promising tool to learn sentence representations by explicitly exploiting the sentence structure. However, most existing work can only exploit simple tree structure, e.g., binary trees, or ignore the order of nodes, which yields suboptimal performance. In this paper, we proposed a novel neural network, namely TreeNet, to capture sentences structurally over the raw unconstrained constituency trees, where the number of child nodes can be arbitrary. In TreeNet, each node is learning from its left sibling and right child in a bottom-up left-to-right order, thus enabling the net to learn over any tree. Furthermore, multiple soft gates and a memory cell are employed in implementing the TreeNet to determine to what extent it should learn, remember and output, which proves to be a simple and efficient mechanism for semantic synthesis. Moreover, TreeNet significantly suppresses convolutional neural networks (CNN) and Long Short-Term Memory (LSTM) with fewer parameters. It improves the classification accuracy by 2%-5% with 42% of the best CNN’s parameters or 94% of standard LSTM’s. Extensive experiments demonstrate TreeNet achieves the state-of-the-art performance on all four typical text classification tasks.
We introduce an active learning framework for general sequence learning tasks including sequence labeling and generation. Most existing active learning algorithms mainly rely on an uncertainty measure derived from the probabilistic classifier for query sample selection. However, such approaches suffer from two shortcomings in the context of sequence learning including 1) cold start problem and 2) label sampling dilemma. To overcome these shortcomings, we propose a deep-learning-based active learning framework to directly identify query samples from the perspective of adversarial learning. Our approach intends to offer labeling priorities for sequences whose information content are least covered by existing labeled data. We verify our sequence-based active learning approach on two tasks including sequence labeling and sequence generation.
Spoken language understanding (SLU) systems, such as goal-oriented chatbots or personal assistants, rely on an initial natural language understanding (NLU) module to determine the intent and to extract the relevant information from the user queries they take as input. SLU systems usually help users to solve problems in relatively narrow domains and require a large amount of in-domain training data. This leads to significant data availability issues that inhibit the development of successful systems. To alleviate this problem, we propose a technique of data selection in the low-data regime that enables us to train with fewer labeled sentences, thus smaller labelling costs. We propose a submodularity-inspired data ranking function, the ratio-penalty marginal gain, for selecting data points to label based only on the information extracted from the textual embedding space. We show that the distances in the embedding space are a viable source of information that can be used for data selection. Our method outperforms two known active learning techniques and enables cost-efficient training of the NLU unit. Moreover, our proposed selection technique does not need the model to be retrained in between the selection steps, making it time efficient as well.
Identifying user consumption intention from social media is of great interests to downstream applications. Since such task is domain-dependent, deep neural networks have been applied to learn transferable features for adapting models from a source domain to a target domain. A basic idea to solve this problem is reducing the distribution difference between the source domain and the target domain such that the transfer error can be bounded. However, the feature transferability drops dramatically in higher layers of deep neural networks with increasing domain discrepancy. Hence, previous work has to use a few target domain annotated data to train domain-specific layers. In this paper, we propose a deep transfer learning framework for consumption intention identification, to reduce the data bias and enhance the transferability in domain-specific layers. In our framework, the representation of the domain-specific layer is mapped to a reproducing kernel Hilbert space, where the mean embeddings of different domain distributions can be explicitly matched. By using an optimal tree kernel method for measuring the mean embedding matching, the domain discrepancy can be effectively reduced. The framework can learn transferable features in a completely unsupervised manner with statistical guarantees. Experimental results on five different domain datasets show that our approach dramatically outperforms state-of-the-art baselines, and it is general enough to be applied to more scenarios. The source code and datasets can be found at http://ir.hit.edu.cn/$\scriptsize{\sim}$xding/index\_english.htm.
Natural language inference aims to predict whether a premise sentence can infer another hypothesis sentence. Recent progress on this task only relies on a shallow interaction between sentence pairs, which is insufficient for modeling complex relations. In this paper, we present an attention-fused deep matching network (AF-DMN) for natural language inference. Unlike existing models, AF-DMN takes two sentences as input and iteratively learns the attention-aware representations for each side by multi-level interactions. Moreover, we add a self-attention mechanism to fully exploit local context information within each sentence. Experiment results show that AF-DMN achieves state-of-the-art performance and outperforms strong baselines on Stanford natural language inference (SNLI), multi-genre natural language inference (MultiNLI), and Quora duplicate questions datasets.
Automatically recognising and extracting the reasoning expressed in natural language text is extremely demanding and only very recently has there been significant headway. While such argument mining focuses on logos (the content of what is said) evidence has demonstrated that using ethos (the character of the speaker) can sometimes be an even more powerful tool of influence. We study the UK parliamentary debates which furnish a rich source of ethos with linguistic material signalling the ethotic relationships between politicians. We then develop a novel deep modular recurrent neural network, DMRNN, approach and employ proven methods from argument mining and sentiment analysis to create an ethos mining pipeline. Annotation of ethotic statements is reliable and its extraction is robust (macro-F1 = 0.83), while annotation of polarity is perfect and its extraction is solid (macro-F1 = 0.84). By exploring correspondences between ethos in political discourse and major events in the political landscape through ethos analytics, we uncover tantalising evidence that identifying expressions of positive and negative ethotic sentiment is a powerful instrument for understanding the dynamics of governments.
Visual question generation aims at asking questions about an image automatically. Existing research works on this topic usually generate a single question for each given image without considering the issue of diversity. In this paper, we propose a question type driven framework to produce multiple questions for a given image with different focuses. In our framework, each question is constructed following the guidance of a sampled question type in a sequence-to-sequence fashion. To diversify the generated questions, a novel conditional variational auto-encoder is introduced to generate multiple questions with a specific question type. Moreover, we design a strategy to conduct the question type distribution learning for each image to select the final questions. Experimental results on three benchmark datasets show that our framework outperforms the state-of-the-art approaches in terms of both relevance and diversity.
In this paper we present an efficient and highly accurate algorithm to prune noisy or over-ambiguous knowledge graphs given as input an extensional definition of a domain of interest, namely as a set of instances or concepts. Our method climbs the graph in a bottom-up fashion, iteratively layering the graph and pruning nodes and edges in each layer while not compromising the connectivity of the set of input nodes. Iterative layering and protection of pre-defined nodes allow to extract semantically coherent DAG structures from noisy or over-ambiguous cyclic graphs, without loss of information and without incurring in computational bottlenecks, which are the main problem of state-of-the-art methods for cleaning large, i.e., Web-scale, knowledge graphs. We apply our algorithm to the tasks of pruning automatically acquired taxonomies using benchmarking data from a SemEval evaluation exercise, as well as the extraction of a domain-adapted taxonomy from the Wikipedia category hierarchy. The results show the superiority of our approach over state-of-art algorithms in terms of both output quality and computational efficiency.
Extracting action sequences from texts is challenging, as it requires commonsense inferences based on world knowledge. Although there has been work on extracting action scripts, instructions, navigation actions, etc., they require either the set of candidate actions be provided in advance, or action descriptions are restricted to a specific form, e.g., description templates. In this paper we aim to extract action sequences from texts in \emph{free} natural language, i.e., without any restricted templates, provided the set of actions is unknown. We propose to extract action sequences from texts based on the deep reinforcement learning framework. Specifically, we view ``selecting'' or ``eliminating'' words from texts as ``actions'', and texts associated with actions as ``states''. We build Q-networks to learn policies of extracting actions and extract plans from the labeled texts. We demonstrate the effectiveness of our approach on several datasets with comparison to state-of-the-art approaches.
Neural networks have been widely used for high resource language (e.g. English) named entity recognition (NER) and have shown state-of-the-art results.However, for low resource languages, such as Dutch, Spanish, due to the limitation of resources and lack of annotated data, taggers tend to have lower performances.To narrow this gap, we propose three novel strategies to enrich the semantic representations of low resource languages: we first develop neural networks to improve low resource word representations by knowledge transfer from high resource language using bilingual lexicons. Further, a lexicon extension strategy is designed to address out-of lexicon problem by automatically learning semantic projections.Thirdly, we regard word-level entity type distribution features as an external language-independent knowledge and incorporate them into our neural architecture. Experiments on two low resource languages (including Dutch and Spanish) demonstrate the effectiveness of these additional semantic representations (average 4.8\% improvement). Moreover, on Chinese OntoNotes 4.0 dataset, our approach achieved an F-score of 83.07\% with 2.91\% absolute gain compared to the state-of-the-art results.
We focus on essay generation, which is a challenging task that generates a paragraph-level text with multiple topics.Progress towards understanding different topics and expressing diversity in this task requires more powerful generators and richer training and evaluation resources. To address this, we develop a multi-topic aware long short-term memory (MTA-LSTM) network.In this model, we maintain a novel multi-topic coverage vector, which learns the weight of each topic and is sequentially updated during the decoding process.Afterwards this vector is fed to an attention model to guide the generator.Moreover, we automatically construct two paragraph-level Chinese essay corpora, 305,000 essay paragraphs and 55,000 question-and-answer pairs.Empirical results show that our approach obtains much better BLEU score compared to various baselines.Furthermore, human judgment shows that MTA-LSTM has the ability to generate essays that are not only coherent but also closely related to the input topics.
Many real-world applications require automated data annotation, such as identifying tissue origins based on gene expressions and classifying images into semantic categories. Annotation classes are often numerous and subject to changes over time, and annotating examples has become the major bottleneck for supervised learning methods. In science and other high-value domains, large repositories of data samples are often available, together with two sources of organic supervision: a lexicon for the annotation classes, and text descriptions that accompany some data samples. Distant supervision has emerged as a promising paradigm for exploiting such indirect supervision by automatically annotating examples where the text description contains a class mention in the lexicon. However, due to linguistic variations and ambiguities, such training data is inherently noisy, which limits the accuracy in this approach. In this paper, we introduce an auxiliary natural language processing system for the text modality, and incorporate co-training to reduce noise and augment signal in distant supervision. Without using any manually labeled data, our EZLearn system learned to accurately annotate data samples in functional genomics and scientific figure comprehension, substantially outperforming state-of-the-art supervised methods trained on tens of thousands of annotated examples.
CBOW (Continuous Bag-Of-Words) is one of the most commonly used techniques to generate word embeddings in various NLP tasks. However, it fails to reach the optimal performance due to uniform involvements of positive words and a simple sampling distribution of negative words. To resolve these issues, we propose OptRank to optimize word ranking and approximate negative sampling for bettering word embedding. Specifically, we first formalize word embedding as a ranking problem. Then, we weigh the positive words by their ranks such that highly ranked words have more importance, and adopt a dynamic sampling strategy to select informative negative words. In addition, an approximation method is designed to efficiently compute word ranks. Empirical experiments show that OptRank consistently outperforms its counterparts on a benchmark dataset with different sampling scales, especially when the sampled subset is small. The code and datasets can be obtained from https://github.com/ouououououou/OptRank.
In this paper, we introduce the Reinforced Mnemonic Reader for machine reading comprehension tasks, which enhances previous attentive readers in two aspects. First, a reattention mechanism is proposed to refine current attentions by directly accessing to past attentions that are temporally memorized in a multi-round alignment architecture, so as to avoid the problems of attention redundancy and attention deficiency. Second, a new optimization approach, called dynamic-critical reinforcement learning, is introduced to extend the standard supervised method. It always encourages to predict a more acceptable answer so as to address the convergence suppression problem occurred in traditional reinforcement learning algorithms. Extensive experiments on the Stanford Question Answering Dataset (SQuAD) show that our model achieves state-of-the-art results. Meanwhile, our model outperforms previous systems by over 6% in terms of both Exact Match and F1 metrics on two adversarial SQuAD datasets.
Entity recommendation, providing search users with an improved experience by assisting them in finding related entities for a given query, has become an indispensable feature of today's Web search engine. Existing studies typically only consider the query issued at the current time step while ignoring the in-session preceding queries. Thus, they typically fail to handle the ambiguous queries such as "apple" because the model could not understand which apple (company or fruit) is talked about. In this work, we believe that the in-session contexts convey valuable evidences that could facilitate the semantic modeling of queries, and take that into consideration for entity recommendation. Furthermore, in order to better model the semantics of queries, we learn the model in a multi-task learning setting where the query representation is shared across entity recommendation and context-aware ranking. We evaluate our approach using large-scale, real-world search logs of a widely used commercial Web search engine. The experimental results show that incorporating context information significantly improves entity recommendation, and learning the model in a multi-task learning setting could bring further improvements.
Goal-Oriented (GO) Dialogue Systems, colloquially known as goal oriented chatbots, help users achieve a predefined goal (e.g. book a movie ticket) within a closed domain. A first step is to understand the user's goal by using natural language understanding techniques. Once the goal is known, the bot must manage a dialogue to achieve that goal, which is conducted with respect to a learnt policy. The success of the dialogue system depends on the quality of the policy, which is in turn reliant on the availability of high-quality training data for the policy learning method, for instance Deep Reinforcement Learning. Due to the domain specificity, the amount of available data is typically too low to allow the training of good dialogue policies. In this paper we introduce a transfer learning method to mitigate the effects of the low in-domain data availability. Our transfer learning based approach improves the bot's success rate by 20% in relative terms for distant domains and we more than double it for close domains, compared to the model without transfer learning. Moreover, the transfer learning chatbots learn the policy up to 5 to 10 times faster. Finally, as the transfer learning approach is complementary to additional processing such as warm-starting, we show that their joint application gives the best outcomes.
This paper analyzes the varied performance of Matrix Factorization (MF) on the related tasks of relation extraction and knowledge-base completion, which have been unified recently into a single framework of knowledge-base inference (KBI) [Toutanova et al., 2015]. We first propose a new evaluation protocol that makes comparisons between MF and Tensor Factorization (TF) models fair. We find that this results in a steep drop in MF performance. Our analysis attributes this to the high out-of-vocabulary (OOV) rate of entity pairs in test folds of commonly-used datasets. To alleviate this issue, we propose three extensions to MF. Our best model is a TF-augmented MF model. This hybrid model is robust and obtains strong results across various KBI datasets.
Argument persuasiveness is one of the most important dimensions of argumentative essay quality, yet it is little studied in automated essay scoring research. Using a recently released corpus of essays that are simultaneously annotated with argument components, argument persuasiveness scores, and attributes of argument components that impact an argument’s persuasiveness, we design and train the first set of neural models that predict the persuasiveness of an argument and its attributes in a student essay, enabling useful feedback to be provided to students on why their arguments are (un)persuasive in addition to how persuasive they are.