Version 2 2022-04-07, 08:03Version 2 2022-04-07, 08:03
Version 1 2022-03-31, 09:43Version 1 2022-03-31, 09:43
conference contribution
posted on 2022-04-07, 08:02authored byHG Noushahr, J Levesley, S Ahmadi, E Mirkes
Multivariate Gaussian probability distributions have been used as distributed representations for text. In comparison with traditional vector representations, these density-based representations are able to model uncertainty, inclusion and entailment. We present a model to learn such representations for products based on a public e-commerce dataset. We qualitatively analyse the properties of the proposed model and how the learned representations capture semantic relatedness, similarity and entailment between products and text.
History
Citation
GaussianProductAttributes: Density-Based Distributed Representations for Products. In: Bramer, M., Ellis, R. (eds) Artificial Intelligence XXXVIII. SGAI-AI 2021. Lecture Notes in Computer Science(), vol 13101. Springer, Cham. https://doi.org/10.1007/978-3-030-91100-3_11
Author affiliation
Department of Mathematics
Version
AM (Accepted Manuscript)
Published in
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)