More Like this

1. Exploring Sentence Vector Spaces through Automatic Summarization

   Templeton, Adly; Kalita, Jugal (December 2018, International Conference on Machine Learning Applications (ICMLA), Orlando.)

   Given vector representations for individual words, it is necessary to compute vector representations of sentences for many applications in a compositional manner, often using artificial neural networks. Relatively little work has explored the internal structure and properties of such sentence vectors. In this paper, we explore the properties of sentence vectors in the context of automatic summarization. In particular, we show that cosine similarity between sentence vectors and document vectors is strongly correlated with sentence importance and that vector semantics can identify and correct gaps between the sentences chosen so far and the document. In addition, we identify specific dimensions which are linked to effective summaries. To our knowledge, this is the first time specific dimensions of sentence embeddings have been connected to sentence properties. We also compare the features of different methods of sentence embeddings. Many of these insights have applications in uses of sentence embeddings far beyond summarization. 

   more » « less
2. How to encode arbitrarily complex morphology in word embeddings, no corpus needed

   Schwartz, Lane; Haley, Coleman; Tyers, Francis (October 2022, Proceedings of the First Workshop on NLP Applications to Field Linguistics)

   Serikov, Oleg; Voloshina, Ekaterina; Postnikova, Anna; Klyachko, Elena; Neminova, Ekaterina; Vylomova, Ekaterina; Shavrina, Tatiana; Le Ferrand, Eric; Malykh, Valentin; Tyers, Francis (Ed.)

   In this paper, we present a straightforward technique for constructing interpretable word embeddings from morphologically analyzed examples (such as interlinear glosses) for all of the world’s languages. Currently, fewer than 300-400 languages out of approximately 7000 have have more than a trivial amount of digitized texts; of those, between 100-200 languages (most in the Indo-European language family) have enough text data for BERT embeddings of reasonable quality to be trained. The word embeddings in this paper are explicitly designed to be both linguistically interpretable and fully capable of handling the broad variety found in the world’s diverse set of 7000 languages, regardless of corpus size or morphological characteristics. We demonstrate the applicability of our representation through examples drawn from a typologically diverse set of languages whose morphology includes prefixes, suffixes, infixes, circumfixes, templatic morphemes, derivational morphemes, inflectional morphemes, and reduplication. 

   more » « less
3. Capacity and Bias of Learned Geometric Embeddings for Directed Graphs

   Boratko, Michael; Zhang, Dongxu; Monath, Nicholas; Vilnis, Luke; Clarkson, Kenneth L; McCallum, Andrew (January 2021, NeurIPS 2021)

   A wide variety of machine learning tasks such as knowledge base completion, ontology alignment, and multi-label classification can benefit from incorporating into learning differentiable representations of graphs or taxonomies. While vectors in Euclidean space can theoretically represent any graph, much recent work shows that alternatives such as complex, hyperbolic, order, or box embeddings have geometric properties better suited to modeling real-world graphs. Experimentally these gains are seen only in lower dimensions, however, with performance benefits diminishing in higher dimensions. In this work, we introduce a novel variant of box embeddings that uses a learned smoothing parameter to achieve better representational capacity than vector models in low dimensions, while also avoiding performance saturation common to other geometric models in high dimensions. Further, we present theoretical results that prove box embeddings can represent any DAG. We perform rigorous empirical evaluations of vector, hyperbolic, and region-based geometric representations on several families of synthetic and real-world directed graphs. Analysis of these results exposes correlations between different families of graphs, graph characteristics, model size, and embedding geometry, providing useful insights into the inductive biases of various differentiable graph representations. 

   more » « less
4. Predicting Human Judgments of Relational Similarity: A Comparison of Computational Models Based on Vector Representations of Meaning

   Snefjella, B.; Ichien, N.; Holyoak, K. J.; Lu, H. (January 2022, Proceedings of the 44th Annual Meeting of the Cognitive Science Society)

   Computational models of verbal analogy and relational similarity judgments can employ different types of vector representations of word meanings (embeddings) generated by machine-learning algorithms. An important question is whether human-like relational processing depends on explicit representations of relations (i.e., representations separable from those of the concepts being related), or whether implicit relation representations suffice. Earlier machine-learning models produced static embeddings for individual words, identical across all contexts. However, more recent Large Language Models (LLMs), which use transformer architectures applied to much larger training corpora, are able to produce contextualized embeddings that have the potential to capture implicit knowledge of semantic relations. Here we compare multiple models based on different types of embeddings to human data concerning judgments of relational similarity and solutions of verbal analogy problems. For two datasets, a model that learns explicit representations of relations, Bayesian Analogy with Relational Transformations (BART), captured human performance more successfully than either a model using static embeddings (Word2vec) or models using contextualized embeddings created by LLMs (BERT, RoBERTa, and GPT-2). These findings support the proposal that human thinking depends on representations that separate relations from the concepts they relate. 

   more » « less
5. Interactive Visualizations of Word Embeddings for K-12 Students

   Bandyopadhyay, Saptarashmi; Xu, Jason; Pawar, Neel; Touretzky, David (June 2022, Proceedings of the AAAI Conference on Artificial Intelligence)

   Word embeddings, which represent words as dense feature vectors, are widely used in natural language processing. In their seminal paper on word2vec, Mikolov and colleagues showed that a feature space created by training a word prediction network on a large text corpus will encode semantic information that supports analogy by vector arithmetic, e.g., "king" minus "man" plus "woman" equals "queen". To help novices appreciate this idea, people have sought effective graphical representations of word embeddings.We describe a new interactive tool for visually exploring word embeddings. Our tool allows users to define semantic dimensions by specifying opposed word pairs, e.g., gender is defined by pairs such as boy/girl and father/mother, and age by pairs such as father/son and mother/daughter. Words are plotted as points in a zoomable and rotatable 3D space, where the third ”residual” dimension encodes distance from the hyperplane defined by all the opposed word vectors with age and gender subtracted out. Our tool allows users to visualize vector analogies, drawing the vector from “king” to “man” and a parallel vector from “woman” to “king-man+woman”, which is closest to “queen”. Visually browsing the embedding space and experimenting with this tool can make word embeddings more intuitive. We include a series of experiments teachers can use to help K-12 students appreciate the strengths and limitations of this representation. 

   more » « less