Ivan Vulić

Monolingual and Cross-Lingual Information Retrieval Models Based on (Bilingual) Word Embeddings

We propose a new unified framework for monolingual (MoIR) and cross-lingual information retrieval (CLIR) which relies on the induction of dense real-valued word vectors known as word embeddings (WE) from comparable data. To this end, we make several important contributions: (1) We present a novel word representation learning model called Bilingual Word Embeddings Skip-Gram (BWESG) which is the first model able to learn bilingual word embeddings solely on the basis of document-aligned comparable data; (2) We demonstrate a simple yet effective approach to building document embeddings from single word embeddings by utilizing models from compositional distributional semantics. BWESG induces a shared cross-lingual embedding vector space in which both words, queries, and documents may be presented as dense real-valued vectors; (3) We build novel ad-hoc MoIR and CLIR models which rely on the induced word and document embeddings and the shared cross-lingual embedding space; (4) Experiments for English and Dutch MoIR, as well as for English-to-Dutch and Dutch-to-English CLIR using benchmarking CLEF 2001-2003 collections and queries demonstrate the utility of our WE-based MoIR and CLIR models. The best results on the CLEF collections are obtained by the combination of the WE-based approach and a unigram language model. We also report on significant improvements in ad-hoc IR tasks of our WE-based framework over the state-of-the-art framework for learning text representations from comparable data based on latent Dirichlet allocation (LDA).

Probabilistic topic modeling in multilingual settings: An overview of its methodology and applications

Abstract Probabilistic topic models are unsupervised generative models which model document content as a two-step generation process, that is, documents are observed as mixtures of latent concepts or topics, while topics are probability distributions over vocabulary words. Recently, a significant research effort has been invested into transferring the probabilistic topic modeling concept from monolingual to multilingual settings. Novel topic models have been designed to work with parallel and comparable texts. We define multilingual probabilistic topic modeling (MuPTM) and present the first full overview of the current research, methodology, advantages and limitations in MuPTM. As a representative example, we choose a natural extension of the omnipresent LDA model to multilingual settings called bilingual LDA (BiLDA). We provide a thorough overview of this representative multilingual model from its high-level modeling assumptions down to its mathematical foundations. We demonstrate how to use the data representation by means of output sets of (i) per-topic word distributions and (ii) per-document topic distributions coming from a multilingual probabilistic topic model in various real-life cross-lingual tasks involving different languages, without any external language pair dependent translation resource: (1) cross-lingual event-centered news clustering, (2) cross-lingual document classification, (3) cross-lingual semantic similarity, and (4) cross-lingual information retrieval. We also briefly review several other applications present in the relevant literature, and introduce and illustrate two related modeling concepts: topic smoothing and topic pruning. In summary, this article encompasses the current research in multilingual probabilistic topic modeling. By presenting a series of potential applications, we reveal the importance of the language-independent and language pair independent data representations by means of MuPTM. We provide clear directions for future research in the field by providing a systematic overview of how to link and transfer aspect knowledge across corpora written in different languages via the shared space of latent cross-lingual topics, that is, how to effectively employ learned per-topic word distributions and per-document topic distributions of any multilingual probabilistic topic model in various cross-lingual applications.

Bilingual Word Embeddings from Non-Parallel Document-Aligned Data Applied to Bilingual Lexicon Induction

We propose a simple yet effective approach to learning bilingual word embeddings (BWEs) from non-parallel document-aligned data (based on the omnipresent skip-gram model), and its application to bilingual lexicon induction (BLI). We demonstrate the utility of the induced BWEs in the BLI task by reporting on benchmarking BLI datasets for three language pairs: (1) We show that our BWE-based BLI models significantly outperform the MuPTM-based and context-counting models in this setting, and obtain the best reported BLI results for all three tested language pairs; (2) We also show that our BWE-based BLI models outperform other BLI models based on recently proposed BWEs that require parallel data for bilingual training.

On the role of seed lexicons in learning bilingual word embeddings

A shared bilingual word embedding space (SBWES) is an indispensable resource in a variety of cross-language NLP and IR tasks. A common approach to the SBWES induction is to learn a mapping function between monolingual semantic spaces, where the mapping critically relies on a seed word lexicon used in the learning process. In this work, we analyze the importance and properties of seed lexicons for the SBWES induction across different dimensions (i.e., lexicon source, lexicon size, translation method, translation pair reliability). On the basis of our analysis, we propose a simple but effective hybrid bilingual word embedding (BWE) model. This model (HYBWE) learns the mapping between two monolingual embedding spaces using only highly reliable symmetric translation pairs from a seed document-level embedding space. We perform bilingual lexicon learning (BLL) with 3 language pairs and show that by carefully selecting reliable translation pairs our new HYBWE model outperforms benchmarking BWE learning models, all of which use more expensive bilingual signals. Effectively, we demonstrate that a SBWES may be induced by leveraging only a very weak bilingual signal (document alignments) along with monolingual data.

Bilingual distributed word representations from document-aligned comparable data

We propose a new model for learning bilingual word representations from nonparallel document-aligned data. Following the recent advances in word representation learning, our model learns dense real-valued word vectors, that is, bilingual word embeddings (BWEs). Unlike prior work on inducing BWEs which heavily relied on parallel sentence-aligned corpora and/or readily available translation resources such as dictionaries, the article reveals that BWEs may be learned solely on the basis of document-aligned comparable data without any additional lexical resources nor syntactic information. We present a comparison of our approach with previous state-of-the-art models for learning bilingual word representations from comparable data that rely on the framework of multilingual probabilistic topic modeling (MuPTM), as well as with distributional local context-counting models. We demonstrate the utility of the induced BWEs in two semantic tasks: (1) bilingual lexicon extraction, (2) suggesting word translations in context for polysemous words. Our simple yet effective BWE-based models significantly outperform the MuPTM-based and context-counting representation models from comparable data as well as prior BWE-based models, and acquire the best reported results on both tasks for all three tested language pairs.

SimVerb-3500: A Large-Scale Evaluation Set of Verb Similarity.

Verbs play a critical role in the meaning of sentences, but these ubiquitous words have received little attention in recent distributional semantics research. We introduce SimVerb-3500, an evaluation resource that provides human ratings for the similarity of 3,500 verb pairs. SimVerb-3500 covers all normed verb types from the USF free-association database, providing at least three examples for every VerbNet class. This broad coverage facilitates detailed analyses of how syntactic and semantic phenomena together influence human understanding of verb meaning. Further, with significantly larger development and test sets than existing benchmarks, SimVerb-3500 enables more robust evaluation of representation learning architectures and promotes the development of methods tailored to verbs. We hope that SimVerb-3500 will enable a richer understanding of the diversity and complexity of verb semantics and guide the development of systems that can effectively represent and interpret this meaning.

Exploiting Image Generality for Lexical Entailment Detection

We exploit the visual properties of concepts for lexical entailment detection by examining a concept’s generality. We introduce three unsupervised methods for determining a concept’s generality, based on its related images, and obtain state-ofthe-art performance on two standard semantic evaluation datasets. We also introduce a novel task that combines hypernym detection and directionality, significantly outperforming a competitive frequencybased baseline.

Visual Bilingual Lexicon Induction with Transferred ConvNet Features

This paper is concerned with the task of bilingual lexicon induction using imagebased features. By applying features from a convolutional neural network (CNN), we obtain state-of-the-art performance on a standard dataset, obtaining a 79% relative improvement over previous work which uses bags of visual words based on SIFT features. The CNN image-based approach is also compared with state-of-the-art linguistic approaches to bilingual lexicon induction, even outperforming these for one of three language pairs on another standard dataset. Furthermore, we shed new light on the type of visual similarity metric to use for genuine similarity versus relatedness tasks, and experiment with using multiple layers from the same network in an attempt to improve performance.

Probabilistic Models of Cross-Lingual Semantic Similarity in Context Based on Latent Cross-Lingual Concepts Induced from Comparable Data

We propose the first probabilistic approach to modeling cross-lingual semantic similarity (CLSS) in context which requires only comparable data. The approach relies on an idea of projecting words and sets of words into a shared latent semantic space spanned by language-pair independent latent semantic concepts (e.g., crosslingual topics obtained by a multilingual topic model). These latent cross-lingual concepts are induced from a comparable corpus without any additional lexical resources. Word meaning is represented as a probability distribution over the latent concepts, and a change in meaning is represented as a change in the distribution over these latent concepts. We present new models that modulate the isolated out-ofcontext word representations with contextual knowledge. Results on the task of suggesting word translations in context for 3 language pairs reveal the utility of the proposed contextualized models of crosslingual semantic similarity.

HyperLex: A Large-Scale Evaluation of Graded Lexical Entailment

We introduce HyperLex—a data set and evaluation resource that quantifies the extent of the semantic category membership, that is, type-of relation, also known as hyponymy–hypernymy or lexical entailment (LE) relation between 2,616 concept pairs. Cognitive psychology research has established that typicality and category/class membership are computed in human semantic memory as a gradual rather than binary relation. Nevertheless, most NLP research and existing large-scale inventories of concept category membership (WordNet, DBPedia, etc.) treat category membership and LE as binary. To address this, we asked hundreds of native English speakers to indicate typicality and strength of category membership between a diverse range of concept pairs on a crowdsourcing platform. Our results confirm that category membership and LE are indeed more gradual than binary. We then compare these human judgments with the predictions of automatic systems, which reveals a huge gap between human performance and state-of-the-art LE, distributional and representation learning models, and substantial differences between the models themselves. We discuss a pathway for improving semantic models to overcome this discrepancy, and indicate future application areas for improved graded LE systems.