Bilingual Embeddings with Random Walks over Multilingual Wordnets

Bilingual word embeddings represent words of two languages in the same space, and allow to transfer knowledge from one language to the other without machine translation. The main approach is to train monolingual embeddings first and then map them using bilingual dictionaries. In this work, we present a novel method to learn bilingual embeddings based on multilingual knowledge bases (KB) such as WordNet. Our method extracts bilingual information from multilingual wordnets via random walks and learns a joint embedding space in one go. We further reinforce cross-lingual equivalence adding bilingual con- straints in the loss function of the popular skipgram model. Our experiments involve twelve cross-lingual word similarity and relatedness datasets in six lan- guage pairs covering four languages, and show that: 1) random walks over mul- tilingual wordnets improve results over just using dictionaries; 2) multilingual wordnets on their own improve over text-based systems in similarity datasets; 3) the good results are consistent for large wordnets (e.g. English, Spanish), smaller wordnets (e.g. Basque) or loosely aligned wordnets (e.g. Italian); 4) the combination of wordnets and text yields the best results, above mapping-based approaches. Our method can be applied to richer KBs like DBpedia or Babel- Net, and can be easily extended to multilingual embeddings. All software and resources are open source.

Interpretable Semantic Textual Similarity: Finding and explaining differences between sentences

User acceptance of artificial intelligence agents might depend on their ability to explain their reasoning, which requires adding an interpretability layer that fa- cilitates users to understand their behavior. This paper focuses on adding an in- terpretable layer on top of Semantic Textual Similarity (STS), which measures the degree of semantic equivalence between two sentences. The interpretability layer is formalized as the alignment between pairs of segments across the two sentences, where the relation between the segments is labeled with a relation type and a similarity score. We present a publicly available dataset of sentence pairs annotated following the formalization. We then develop a system trained on this dataset which, given a sentence pair, explains what is similar and different, in the form of graded and typed segment alignments. When evaluated on the dataset, the system performs better than an informed baseline, showing that the dataset and task are well-defined and feasible. Most importantly, two user studies show how the system output can be used to automatically produce explanations in natural language. Users performed better when having access to the explanations, pro- viding preliminary evidence that our dataset and method to automatically produce explanations is useful in real applications.

Learning class-to-class selectional preferences

Selectional preference learning methods have usually focused on word-to-class relations, e.g., a verb selects as its subject a given nominal class. This papers extends previous statistical models to class-to-class preferences, and presents a model that learns selectional preferences for classes of verbs. The motivation is twofold: different senses of a verb may have different preferences, and some classes of verbs can share preferences. The model is tested on a word sense disambiguation task which uses subject-verb and object-verb relationships extracted from a small sense-disambiguated corpus.

Towards a single proposal is spelling correction

The study presented here relies on the integrated use of different kinds of knowledge in order to improve first-guess accuracy in non-word context-sensitive correction for general unrestricted texts. State of the art spelling correction systems, e.g. ispell, apart from detecting spelling errors, also assist the user by offering a set of candidate corrections that are close to the misspelled word. Based on the correction proposals of ispell, we built several guessers, which were combined in different ways. Firstly, we evaluated all possibilities and selected the best ones in a corpus with artificially generated typing errors. Secondly, the best combinations were tested on texts with genuine spelling errors. The results for the latter suggest that we can expect automatic non-word correction for all the errors in a free running text with 80% precision and a single proposal 98% of the times (1.02 proposals on average).

Lexical Knowledge Representation in an Intelligent Dictionary Help System

The frame-based knowledge representation model adopted in IDHS (Intelligent Dictionary Help System) is described in this paper. It is used to represent the lexical knowledge acquired automatically from a conventional dictionary. Moreover, the enrichment processes that have been performed on the Dictionary Knowledge Base and the dynamic exploitation of this knowledge - both based on the exploitation of the properties of lexical semantic relations - are also described.