Search Results

Search Constraints

Start Over
Filtering by: Depositor ssim truas@umich.edu Remove constraint Depositor ssim: truas@umich.edu Discipline Other Remove constraint Discipline: Other

Search Results

  • Enhanced word embeddings using multi-semantic representation through lexical chains

    work
    Creator:
    Ruas, Terry, Ferreira, Charles H. P., Grosky, William, França, Fabrício O., and Medeiros, Débora M. R,
    Description:
    The relationship between words in a sentence often tell us more about the underlying semantic content of a document than its actual words, individually. Recent publications in the natural language processing arena, more specifically using word embeddings, try to incorporate semantic aspects into their word vector representation by considering the context of words and how they are distributed in a document collection. In this work, we propose two novel algorithms, called Flexible Lexical Chain II and Fixed Lexical Chain II that combine the semantic relations derived from lexical chains, prior knowledge from lexical databases, and the robustness of the distributional hypothesis in word embeddings into a single decoupled system. In short, our approach has three main contributions: (i) unsupervised techniques that fully integrate word embeddings and lexical chains; (ii) a more solid semantic representation that considers the latent relation between words in a document; and (iii) lightweight word embeddings models that can be extended to any natural language task. Knowledge-based systems that use natural language text can benefit from our approach to mitigate ambiguous semantic representations provided by traditional statistical approaches. The proposed techniques are tested against seven word embeddings algorithms using five different machine learning classifiers over six scenarios in the document classification task. Our results show that the integration between lexical chains and word embeddings representations sustain state-of-the-art results, even against more complex systems. Github:  https://github.com/truas/LexicalChain_Builder
    Keyword:
    document classification, lexical chains, word embeddings, synset embeddings, chain2vec, and natural language processing
    Citation to related publication:
    Terry Ruas, Charles Henrique Porto Ferreira, William Grosky, Fabrício Olivetti de França, Débora Maria Rossi de Medeiros, "Enhanced word embeddings using multi-semantic representation through lexical chains", Information Sciences, 2020,  https://doi.org/10.1016/j.ins.2020.04.048
    Discipline:
    Other, Science, and Engineering

  • Multi-Sense embeddings through a word sense disambiguation process

    work
    Creator:
    Ruas, Terry, Grosky, William, and Aizawa, Akiko
    Description:
    This data set is a collection of word similarity benchmarks (RG65, MEN3K, Wordsim 353, simlex999, SCWS, yp130, simverb3500) in their original format and converted into a cosine similarity scale. In addition, we have two Wikpedia Dumps from 2010 (April) and 2018 (January) in which we provide the original format (raw words), converted using the techniques described in the paper (MSSA, MSSA-D and MSSA-NR) (title in this repository), and also the word embeddings models for 300d and 1000d using a word2vec implementation. A readme.txt is provided with more details for each file.
    Keyword:
    multi-sense embeddings, MSSA, word2vec, wikipedia dump, synset, and natural language processing
    Citation to related publication:
     https://doi.org/10.1016/j.eswa.2019.06.026
    Discipline:
    Other