Distributed representations of words (aka word embedding) have proven helpful in solving natural language processing (NLP) tasks. Training distributed representations of words with neural networks has lately been a major focus of researchers in the field. Recent work on word embedding, the Continuous Bag-of-Words (CBOW) model and the Continuous Skip-gram (Skip-gram) model, have produced particularly impressive results, significantly speeding up the training process to enable word representation learning from large-scale data. However, both CBOW and Skip-gram do not pay enough attention to word proximity in terms of model or word ambiguity in terms of linguistics. In this paper, we propose Proximity-Ambiguity Sensitive (PAS) models (i.e. PAS CBOW and PAS Skip-gram) to produce high quality distributed representations of words considering both word proximity and ambiguity. From the model perspective, we introduce proximity weights as parameters to be learned in PAS CBOW and used in PAS Skip-gram. By better modeling word proximity, we reveal the strength of pooling-structured neural networks in word representation learning. The proximity-sensitive pooling layer can also be applied to other neural network applications that employ pooling layers. From the linguistics perspective, we train multiple representation vectors per word. Each representation vector corresponds to a particular group of POS tags of the word. By using PAS models, we achieved a 16.9% increase in accuracy over state-of-the-art models.

This thesis seeks to address word reasoning problems from a semantic standpoint, proposing a uniform approach for generating solutions while also providing human-understandable explanations. Current state of the art solvers of semantic problems rely on traditional machine learning methods. Therefore their results are not easily reusable by algorithms or interpretable by humans. We propose leveraging web-scale knowledge graphs to determine a semantic frame of interpretation. Semantic knowledge graphs are graphs in which nodes represent concepts and the edges represent the relations between them. Our approach has the following advantages: (1) it reduces the space in which the problem is to be solved; (2) sparse and noisy data can be used without relying only on the relations deducible from the data itself; (3) the output of the inference algorithm is supported by an interpretable justification. We demonstrate our approach in two domains: (1) Topic Modeling: We form topics using connectivity in semantic graphs. We use the same topic models for two very different recommendation systems, one designed for high noise interactive applications and the other for large amounts of web data. (2) Analogy Solving: For humans, analogies are a fundamental reasoning pattern, which relies on abstraction and comparative analysis. In order for an analogy to be understood, precise relations have to be identified and mapped. We introduce graph algorithms to assess the analogy strength in contexts derived from the analogy words. We demonstrate our approach by solving standardized test analogy question.

We deal with embedding a large scale knowledge graph composed of entities and relations into a continuous vector space. TransE is a promising method proposed recently, which is very efficient while achieving state-of-the-art predictive performance. We discuss some mapping properties of relations which should be considered in embedding, such as reflexive, one-to-many, many-to-one, and many-to-many. We note that TransE does not do well in dealing with these properties. Some complex models are capable of preserving these mapping properties but sacrifice efficiency in the process. To make a good trade-off between model capacity and efficiency, in this paper we propose TransH which models a relation as a hyperplane together with a translation operation on it. In this way, we can well preserve the above mapping properties of relations with almost the same model complexity of TransE. Additionally, as a practical knowledge graph is often far from completed, how to construct negative examples to reduce false negative labels in training is very important. Utilizing the one-to-many/many-to-one mapping property of a relation, we propose a simple trick to reduce the possibility of false negative labeling. We conduct extensive experiments on link prediction, triplet classification and fact extraction on benchmark datasets like WordNet and Freebase. Experiments show TransH delivers significant improvements over TransE on predictive accuracy with comparable capability to scale up.

Creating knowledge bases based on the crowd-sourced wikis, like Wikipedia, has attracted significant research interest in the field of intelligent Web. However, the derived taxonomies usually contain many mistakenly imported taxonomic relations due to the difference between the user-generated subsumption relations and the semantic taxonomic relations. Current approaches to solving the problem still suffer the following issues: (i) the heuristic-based methods strongly rely on specific language dependent rules. (ii) the corpus-based methods depend on a large-scale high-quality corpus, which is often unavailable. In this paper, we formulate the cross-lingual taxonomy derivation problem as the problem of cross-lingual taxonomic relation prediction. We investigate different linguistic heuristics and language independent features, and propose a cross-lingual knowledge validation based dynamic adaptive boosting model to iteratively reinforce the performance of taxonomic relation prediction. The proposed approach successfully overcome the above issues, and experiments show that our approach significantly outperforms the designed state-of-the-art comparison methods.

A semantic net can be used to represent a sentence. A sentence in a language contains semantics which are polar in nature, that is, semantics which are positive, neutral and negative. Neutrosophy is a relatively new field of science which can be used to mathematically represent triads of concepts. These triads include truth, indeterminacy and falsehood, and so also positivity, neutrality and negativity. Thus a conventional semantic net has been extended in this paper using neutrosophy into a Polar Fuzzy Neutrosophic Semantic Net. A Polar Fuzzy Neutrosophic Semantic Net has been implemented in MATLAB and has been used to illustrate a polar sentence in English language. The paper demonstrates a method for the representation of polarity in a computers memory. Thus, polar concepts can be applied to imbibe a machine such as a robot, with emotions, making machine emotion representation possible.

Cross-language learning allows us to use training data from one language to build models for a different language. Many approaches to bilingual learning require that we have word-level alignment of sentences from parallel corpora. In this work we explore the use of autoencoder-based methods for cross-language learning of vectorial word representations that are aligned between two languages, while not relying on word-level alignments. We show that by simply learning to reconstruct the bag-of-words representations of aligned sentences, within and between languages, we can in fact learn high-quality representations and do without word alignments. Since training autoencoders on word observations presents certain computational issues, we propose and compare different variations adapted to this setting. We also propose an explicit correlation maximizing regularizer that leads to significant improvement in the performance. We empirically investigate the success of our approach on the problem of cross-language test classification, where a classifier trained on a given language (e.g., English) must learn to generalize to a different language (e.g., German). These experiments demonstrate that our approaches are competitive with the state-of-the-art, achieving up to 10-14 percentage point improvements over the best reported results on this task.

Recent work on learning multilingual word representations usually relies on the use of word-level alignements (e.g. infered with the help of GIZA++) between translated sentences, in order to align the word embeddings in different languages. In this workshop paper, we investigate an autoencoder model for learning multilingual word representations that does without such word-level alignements. The autoencoder is trained to reconstruct the bag-of-word representation of given sentence from an encoded representation extracted from its translation. We evaluate our approach on a multilingual document classification task, where labeled data is available only for one language (e.g. English) while classification must be performed in a different language (e.g. French). In our experiments, we observe that our method compares favorably with a previously proposed method that exploits word-level alignments to learn word representations.

Large-scale information processing systems are able to extract massive collections of interrelated facts, but unfortunately transforming these candidate facts into useful knowledge is a formidable challenge. In this paper, we show how uncertain extractions about entities and their relations can be transformed into a knowledge graph. The extractions form an extraction graph and we refer to the task of removing noise, inferring missing information, and determining which candidate facts should be included into a knowledge graph as knowledge graph identification. In order to perform this task, we must reason jointly about candidate facts and their associated extraction confidences, identify co-referent entities, and incorporate ontological constraints. Our proposed approach uses probabilistic soft logic (PSL), a recently introduced probabilistic modeling framework which easily scales to millions of facts. We demonstrate the power of our method on a real-world set of extractions from the NELL project containing over 1M extractions and 70K ontological relations. We show that compared to existing methods, our approach is able to achieve improved AUC and F1 with significantly lower running time.

One way to obtain large amounts of semantic data is to extract facts from the vast quantities of text that is now available on-line. The relatively low accuracy of current information extraction techniques introduces a need for evaluating the quality of the knowledge bases (KBs) they generate. We frame the problem as comparing KBs generated by different systems from the same documents and show that exploiting provenance leads to more efficient techniques for aligning them and identifying their differences. We describe two types of tools: entity-match focuses on differences in entities found and linked; kbdiff focuses on differences in relations among those entities. Together, these tools support assessment of relative KB accuracy by sampling the parts of two KBs that disagree. We explore the usefulness of the tools through the construction of tens of different KBs built from the same 26,000 Washington Post articles and identifying the differences.

The approach described here allows to use the fuzzy Object Based Representation of imprecise and uncertain knowledge. This representation has a great practical interest due to the possibility to realize reasoning on classification with a fuzzy semantic network based system. For instance, the distinction between necessary, possible and user classes allows to take into account exceptions that may appear on fuzzy knowledge-base and facilitates integration of user's Objects in the base. This approach describes the theoretical aspects of the architecture of the whole experimental A.I. system we built in order to provide effective on-line assistance to users of new technological systems: the understanding of "how it works" and "how to complete tasks" from queries in quite natural languages. In our model, procedural semantic networks are used to describe the knowledge of an "ideal" expert while fuzzy sets are used both to describe the approximative and uncertain knowledge of novice users in fuzzy semantic networks which intervene to match fuzzy labels of a query with categories from our "ideal" expert.