An official website of the United States government
Unsupervised word embeddings have become a popular approach of word representation in NLP tasks. However there are limitations to the semantics represented by unsupervised embeddings, and inadequate fine-tuning of embeddings can lead to suboptimal performance. We propose a novel learning technique called Delta Embedding Learning, which can be applied to general NLP tasks to improve performance by optimized tuning of the word embeddings. A structured regularization is applied to the embeddings to ensure they are tuned in an incremental way. As a result, the tuned word embeddings become better word representations by absorbing semantic information from supervision without “forgetting.” We apply the method to various NLP tasks and see a consistent improvement in performance. Evaluation also confirms the tuned word embeddings have better semantic properties.
more »
« less
embComp: Visual Interactive Comparison of Vector Embeddings
This paper introduces embComp, a novel approach for comparing two embeddings that capture the similarity between objects, such as word and document embeddings. We survey scenarios where comparing these embedding spaces is useful. From those scenarios, we derive common tasks, introduce visual analysis methods that support these tasks, and combine them into a comprehensive system. One of embComp's central features are overview visualizations that are based on metrics for measuring differences in the local structure around objects. Summarizing these local metrics over the embeddings provides global overviews of similarities and differences. Detail views allow comparison of the local structure around selected objects and relating this local information to the global views. Integrating and connecting all of these components, embComp supports a range of analysis workflows that help understand similarities and differences between embedding spaces. We assess our approach by applying it in several use cases, including understanding corpora differences via word vector embeddings, and understanding algorithmic differences in generating embeddings.
more »
« less
- Award ID(s):
- 1841349
- PAR ID:
- 10309253
- Date Published:
- Journal Name:
- IEEE Transactions on Visualization and Computer Graphics
- ISSN:
- 1077-2626
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Transformer-based language models such as BERT and its variants have found widespread use in natural language processing (NLP). A common way of using these models is to fine-tune them to improve their performance on a specific task. However, it is currently unclear how the fine-tuning process affects the underlying structure of the word embeddings from these models. We present TopoBERT, a visual analytics system for interactively exploring the fine-tuning process of various transformer-based models – across multiple fine-tuning batch updates, subsequent layers of the model, and different NLP tasks – from a topological perspective. The system uses the mapper algorithm from topological data analysis (TDA) to generate a graph that approximates the shape of a model’s embedding space for an input dataset. TopoBERT enables its users (e.g. experts in NLP and linguistics) to (1) interactively explore the fine-tuning process across different model-task pairs, (2) visualize the shape of embedding spaces at multiple scales and layers, and (3) connect linguistic and contextual information about the input dataset with the topology of the embedding space. Using TopoBERT, we provide various use cases to exemplify its applications in exploring fine-tuned word embeddings. We further demonstrate the utility of TopoBERT, which enables users to generate insights about the fine-tuning process and provides support for empirical validation of these insights.more » « less
-
Contextualized word embeddings, such as ELMo, provide meaningful representations for words and their contexts. They have been shown to have a great impact on downstream applications. However, we observe that the contextualized embeddings of a word might change drastically when its contexts are paraphrased. As these embeddings are over-sensitive to the context, the downstream model may make different predictions when the input sentence is paraphrased. To address this issue, we propose a post-processing approach to retrofit the embedding with paraphrases. Our method learns an orthogonal transformation on the input space of the contextualized word embedding model, which seeks to minimize the variance of word representations on paraphrased contexts. Experiments show that the proposed method significantly improves ELMo on various sentence classification and inference tasks.more » « less
-
Wasserstein distances form a family of metrics on spaces of probability measures that have recently seen many applications. However, statistical analysis in these spaces is complex due to the nonlinearity of Wasserstein spaces. One potential solution to this problem is Linear Optimal Transport (LOT). This method allows one to find a Euclidean embedding, called {\it LOT embedding}, of measures in some Wasserstein spaces, but some information is lost in this embedding. So, to understand whether statistical analysis relying on LOT embeddings can make valid inferences about original data, it is helpful to quantify how well these embeddings describe that data. To answer this question, we present a decomposition of the {\it Fr\'echet variance} of a set of measures in the 2-Wasserstein space, which allows one to compute the percentage of variance explained by LOT embeddings of those measures. We then extend this decomposition to the Fused Gromov-Wasserstein setting. We also present several experiments that explore the relationship between the dimension of the LOT embedding, the percentage of variance explained by the embedding, and the classification accuracy of machine learning classifiers built on the embedded data. We use the MNIST handwritten digits dataset, IMDB-50000 dataset, and Diffusion Tensor MRI images for these experiments. Our results illustrate the effectiveness of low dimensional LOT embeddings in terms of the percentage of variance explained and the classification accuracy of models built on the embedded data.more » « less
-
The use of word vector embeddings as the basis for many upstream tasks in text processing has lead to large improvements in accuracy. However, the exact reasons for this success largely remain unclear, as the properties and relations that these embeddings encode are often not well understood. Our goal in this ongoing project is to design effective interactive visualizations that help practitioners and researchers understand and compare such spaces better. The initial steps we have taken is to review relevant literature to identify properties and relations of word vectors that are important for various applications. From these, we derive basic tasks to inform the design of adequate and effective interactive visualizations that help users gain deeper insights into the structure of vector spaces. In addition, we present three initial designs to support these tasks.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1109/TVCG.2020.3045918
