Sparsity-aware personalized recommender system via meta-learning
Wang, Junting
This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.
Permalink
https://hdl.handle.net/2142/115612
Description
Title
Sparsity-aware personalized recommender system via meta-learning
Author(s)
Wang, Junting
Issue Date
2022-04-27
Director of Research (if dissertation) or Advisor (if thesis)
Sundaram, Hari
Department of Study
Computer Science
Discipline
Computer Science
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
M.S.
Degree Level
Thesis
Keyword(s)
Recommendation System
Collaborative Filtering
Meta Learning
Few-shot learning
Abstract
With the advancement of neural collaborative filtering methods, deep learning methods have become the backbone of modern recommender systems. Neural recommenders provide significant performance gains over conventional methods. However, the recommendations made by them are not truly personalized. Instead, they tend to suggest popular items due to the popularity bias among items. Popularity bias is a fundamental challenge in recommender systems, and it originates from the heavy-tailed distribution of users’ activity data. Modern neural recommenders lack the resolution to rank long-tail items accurately since the interaction data used in the model training is heavily skewed. The biased training results in a biased recommendation model that only recommends the popular subset of the item inventory.
In this thesis, we propose a meta-learning framework ProtoCF that effectively eliminates the distribution mismatch between items and learns robust prototype representations for long-tail items. Our experimental results demonstrate that ProtoCF consistently outperforms state-of-the-art approaches on overall recommendation (by 5% Recall@50) while achieving significant gains (of 60-80% Recall@50) for tail items with less than 20 interactions.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.