BERT Goes Shopping: Comparing Distributional Models for Product
Representations
F. Bianchi, B. Yu, and J. Tagliabue. (2020)cite arxiv:2012.09807Comment: Updated version. Published as a workshop paper at ECNLP 4 at ACL-IJCNLP 2021.
Abstract
Word embeddings (e.g., word2vec) have been applied successfully to eCommerce
products through~prod2vec. Inspired by the recent performance
improvements on several NLP tasks brought by contextualized embeddings, we
propose to transfer BERT-like architectures to eCommerce: our model --
~Prod2BERT -- is trained to generate representations of products
through masked session modeling. Through extensive experiments over multiple
shops, different tasks, and a range of design choices, we systematically
compare the accuracy of~Prod2BERT and~prod2vec embeddings:
while~Prod2BERT is found to be superior in several scenarios, we
highlight the importance of resources and hyperparameters in the best
performing models. Finally, we provide guidelines to practitioners for training
embeddings under a variety of computational and data constraints.
Description
BERT Goes Shopping: Comparing Distributional Models for Product Representations
%0 Generic
%1 bianchi2020shopping
%A Bianchi, Federico
%A Yu, Bingqing
%A Tagliabue, Jacopo
%D 2020
%K prod2vec recommendation
%T BERT Goes Shopping: Comparing Distributional Models for Product
Representations
%U http://arxiv.org/abs/2012.09807
%X Word embeddings (e.g., word2vec) have been applied successfully to eCommerce
products through~prod2vec. Inspired by the recent performance
improvements on several NLP tasks brought by contextualized embeddings, we
propose to transfer BERT-like architectures to eCommerce: our model --
~Prod2BERT -- is trained to generate representations of products
through masked session modeling. Through extensive experiments over multiple
shops, different tasks, and a range of design choices, we systematically
compare the accuracy of~Prod2BERT and~prod2vec embeddings:
while~Prod2BERT is found to be superior in several scenarios, we
highlight the importance of resources and hyperparameters in the best
performing models. Finally, we provide guidelines to practitioners for training
embeddings under a variety of computational and data constraints.
@misc{bianchi2020shopping,
abstract = {Word embeddings (e.g., word2vec) have been applied successfully to eCommerce
products through~\textit{prod2vec}. Inspired by the recent performance
improvements on several NLP tasks brought by contextualized embeddings, we
propose to transfer BERT-like architectures to eCommerce: our model --
~\textit{Prod2BERT} -- is trained to generate representations of products
through masked session modeling. Through extensive experiments over multiple
shops, different tasks, and a range of design choices, we systematically
compare the accuracy of~\textit{Prod2BERT} and~\textit{prod2vec} embeddings:
while~\textit{Prod2BERT} is found to be superior in several scenarios, we
highlight the importance of resources and hyperparameters in the best
performing models. Finally, we provide guidelines to practitioners for training
embeddings under a variety of computational and data constraints.},
added-at = {2021-07-08T17:18:56.000+0200},
author = {Bianchi, Federico and Yu, Bingqing and Tagliabue, Jacopo},
biburl = {https://www.bibsonomy.org/bibtex/2496828b45c05e7b41fb845778bbcbe4e/e.fischer},
description = {BERT Goes Shopping: Comparing Distributional Models for Product Representations},
interhash = {7971596e5653ffd338a10f6d48114e31},
intrahash = {496828b45c05e7b41fb845778bbcbe4e},
keywords = {prod2vec recommendation},
note = {cite arxiv:2012.09807Comment: Updated version. Published as a workshop paper at ECNLP 4 at ACL-IJCNLP 2021},
timestamp = {2021-07-08T17:18:56.000+0200},
title = {BERT Goes Shopping: Comparing Distributional Models for Product
Representations},
url = {http://arxiv.org/abs/2012.09807},
year = 2020
}