Using Dynamic Embeddings to Improve Static Embeddings

Recent methods for learning vector space representations of words have succeeded in capturing fine-grained semantic and syntactic regularities using vector arithmetic , but the origin of these regularities has remained opaque. We analyze and make explicit the model properties needed for such regularities to emerge in word vectors. The result is a new global log-bilinear regression model that combines the advantages of the two major model families in the literature: global matrix factorization and local context window methods. Our model efficiently leverages statistical information by training only on the nonzero elements in a word-word co-occurrence matrix, rather than on the entire sparse matrix or on individual context windows in a large corpus. The model produces a vector space with meaningful sub-structure, as evidenced by its performance of 75% on a recent word analogy task. It also outperforms related models on similarity tasks and named entity recognition.

Analysis and Application of Natural Language and Speech Processing, 2023

Vector semantics has slightly become a key tool for Natural Language Processing, especially concerning text analysis. This kind of vector representation is usually encoded through embeddings that can be used to encode semantic information at different levels of granularity. In fact, through the years, not only models for word embeddings have been developed, but also for sentence and documents. With this work we address sentence embeddings, in particular the non-parametric ones, which offer a good trade off between performance and inference speed. We present Static Fuzzy Bag-of-Word (SFBoW) model, a refinement of the Fuzzy Bag-of-Words approach yielding fixed-dimension sentence embeddings. We targeted fixed size embeddings to promote caching a re-usability, speeding the inference of a system that relies on our model. In this paper we explore various approaches for the construction of a static universe matrix, fundamental to make the sentence embeddings of fixed size. To show the validity of our approach, we benchmarked our model on a semantic similarity task, obtaining competitive performances.

Northern European Journal of Language Technology

A variety of NLP applications use word2vec skip-gram, GloVe, and fastText word embeddings. These models learn two sets of embedding vectors, but most practitioners use only one of them, or alternately an unweighted sum of both. This is the first study to systematically explore a range of linear combinations between the first and second embedding sets. We evaluate these combinations on a set of six NLP benchmarks including IR, POS-tagging, and sentence similarity. We show that the default embedding combinations are often suboptimal and demonstrate 1.0-8.0% improvements. Notably, GloVe’s default unweighted sum is its least effective combination across tasks. We provide a theoretical basis for weighting one set of embeddings more than the other according to the algorithm and task. We apply our findings to improve accuracy in applications of cross-lingual alignment and navigational knowledge by up to 15.2%.

ArXiv, 2021

The current dominance of deep neural networks in natural language processing is based on contextual embeddings such as ELMo, BERT, and BERT derivatives. Most existing work focuses on English; in contrast, we present here the first multilingual empirical comparison of two ELMo and several monolingual and multilingual BERT models using 14 tasks in nine languages. In monolingual settings, our analysis shows that monolingual BERT models generally dominate, with a few exceptions such as the dependency parsing task, where they are not competitive with ELMo models trained on large corpora. In cross-lingual settings, BERT models trained on only a few languages mostly do best, closely followed by massively multilingual BERT models.

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 2019

Word embeddings have been widely adopted across several NLP applications. Most existing word embedding methods utilize sequential context of a word to learn its embedding. While there have been some attempts at utilizing syntactic context of a word, such methods result in an explosion of the vocabulary size. In this paper, we overcome this problem by proposing SynGCN, a flexible Graph Convolution based method for learning word embeddings. SynGCN utilizes the dependency context of a word without increasing the vocabulary size. Word embeddings learned by SynGCN outperform existing methods on various intrinsic and extrinsic tasks and provide an advantage when used with ELMo. We also propose SemGCN, an effective framework for incorporating diverse semantic knowledge for further enhancing learned word representations. We make the source code of both models available to encourage reproducible research.

Learning representations for semantic relations is important for various tasks such as analogy detection , relational search, and relation classification. Although there have been several proposals for learning representations for individual words, learning word representations that explicitly capture the semantic relations between words remains under developed. We propose an unsupervised method for learning vector representations for words such that the learnt representations are sensitive to the semantic relations that exist between two words. First, we extract lexical patterns from the co-occurrence contexts of two words in a corpus to represent the semantic relations that exist between those two words. Second, we represent a lexical pattern as the weighted sum of the representations of the words that co-occur with that lexical pattern. Third, we train a binary classifier to detect relationally similar vs. non-similar lexical pattern pairs. The proposed method is unsupervised in ...

Proceedings of the 2019 Conference of the North, 2019

Pre-trained word vectors are ubiquitous in Natural Language Processing applications. In this paper, we show how training word embeddings jointly with bigram and even trigram embeddings, results in improved unigram embeddings. We claim that training word embeddings along with higher n-gram embeddings helps in the removal of the contextual information from the unigrams, resulting in better stand-alone word embeddings. We empirically show the validity of our hypothesis by outperforming other competing word representation models by a significant margin on a wide variety of tasks. We make our models publicly available.

Most word embedding algorithms such as word2vec or fastText construct two sort of vectors: for words and for contexts. Naive use of vectors of only one sort leads to poor results. We suggest using indefinite inner product in skip-gram negative sampling algorithm. This allows us to use only one sort of vectors without loss of quality. Our "context-free" cf algorithm performs on par with SGNS on word similarity datasets.

We propose two novel model architectures for computing continuous vector representations of words from very large data sets. The quality of these representations is measured in a word similarity task, and the results are compared to the previously best performing techniques based on different types of neural networks. We observe large improvements in accuracy at much lower computational cost, i.e. it takes less than a day to learn high quality word vectors from a 1.6 billion words data set. Furthermore, we show that these vectors provide state-of-the-art performance on our test set for measuring syntactic and semantic word similarities. arXiv:1301.3781v3 [cs.CL] 7 Sep 2013 1 The test set is available at www.fit.vutbr.cz/˜imikolov/rnnlm/word-test.v1.txt 2

Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence

As time goes by, language evolves with word semantics changing. Unfortunately, traditional word embedding methods neglect the evolution of language and assume that word representations are static. Although contextualized word embedding models can capture the diverse representations of polysemous words, they ignore temporal information as well. To tackle the aforementioned challenges, we propose a graph-based dynamic word embedding (GDWE) model, which focuses on capturing the semantic drift of words continually. We introduce word-level knowledge graphs (WKGs) to store short-term and long-term knowledge. WKGs can provide rich structural information as supplement of lexical information, which help enhance the word embedding quality and capture semantic drift quickly. Theoretical analysis and extensive experiments validate the effectiveness of our GDWE on dynamic word embedding learning.