5 Hits in 1.2 sec

embComp: Visual Interactive Comparison of Vector Embeddings [article]

Florian Heimerl, Christoph Kralj, Torsten Möller, Michael Gleicher
2019 arXiv   pre-print
This work introduces embComp, a novel approach for comparing two embeddings that capture the similarity between objects, such as word and document embeddings.  ...  Detail views allow comparison of the local structure around selected objects and relating this local information to the global views.  ...  VISUAL INTERACTIVE EMBEDDING COMPARISON The previously described overviews build the core of our approach for embedding comparison.  ... 
arXiv:1911.01542v1 fatcat:5bkrti4bfbhwdm7pavpokypgcu

VERB: Visualizing and Interpreting Bias Mitigation Techniques for Word Representations [article]

Archit Rathore, Sunipa Dev, Jeff M. Phillips, Vivek Srikumar, Yan Zheng, Chin-Chia Michael Yeh, Junpeng Wang, Wei Zhang, Bei Wang
2021 arXiv   pre-print
on the word vectors using dimensionality reduction and interactive visual exploration.  ...  In this paper, we utilize interactive visualization to increase the interpretability and accessibility of a collection of state-of-the-art debiasing techniques.  ...  CONCLUSIONS AND DISCUSSIONS This paper presents VERB, a new tool for visualizing, interacting with, and teaching embedded representations of data.  ... 
arXiv:2104.02797v1 fatcat:htizaqvn2vdyzhxaef6umscz7a

Embedding Comparator: Visualizing Differences in Global Structure and Local Neighborhoods via Small Multiples [article]

Angie Boggust, Brandon Carter, Arvind Satyanarayan
2022 arXiv   pre-print
In response, we present the Embedding Comparator, an interactive system that presents a global comparison of embedding spaces alongside fine-grained inspection of local neighborhoods.  ...  It systematically surfaces points of comparison by computing the similarity of the k-nearest neighbors of every embedded object between a pair of spaces.  ...  [23] developed the embedding comparison system empComp. embComp facilitates top-down comparison of embeddings by displaying summary visualizations of embedding space differences and employing interaction  ... 
arXiv:1912.04853v3 fatcat:zfool6r745hlpef4nu7tl3gxi4

Emblaze: Illuminating Machine Learning Representations through Interactive Comparison of Embedding Spaces [article]

Venkatesh Sivaraman, Yiwei Wu, Adam Perer
2022 arXiv   pre-print
Emblaze uses an animated, interactive scatter plot with a novel Star Trail augmentation to enable visual comparison.  ...  Through a series of case studies with ML experts, we demonstrate how interactive comparison with Emblaze can help gain new insights into embedding space structure.  ...  We thank Carolyn Rosé, Jill Lehman, Denis Newman-Griffis, and Dominik Moritz for their valuable feedback throughout the development of Emblaze.  ... 
arXiv:2202.02641v2 fatcat:sndccrscgzcmlh2hag7fqt6yfu

Comparing and Exploring High-Dimensional Data with Dimensionality Reduction Algorithms and Matrix Visualizations

Rene Cutura, Michaël Aupetit, Jean-Daniel Fekete, Michael Sedlmair
2020 Proceedings of the International Conference on Advanced Visual Interfaces  
Figure 1 : The scatterplots in (a) and (c) visualize the dimensionality reduced representations of two distinct subspaces of a high-dimensional dataset.  ...  ABSTRACT We propose Compadre, a tool for visual analysis for comparing distances of high-dimensional (HD) data and their low-dimensional projections.  ...  EmbComp [20] is a recent approach that seeks to bridge the usage of metrics and interactive visual analysis for analyzing different word embeddings.  ... 
doi:10.1145/3399715.3399875 dblp:conf/avi/CuturaAFS20 fatcat:oehwih6ugncv7pkujwbw2vk5we