Federated Collaborative Filtering for Privacy-Preserving Personalized Recommendation System

By Muhammad Ammad-ud-din, Elena Ivannikova, Suleiman A. Khan, Were Oyomno, Qiang Fu, Kuan Eeik Tan, Adrian Flanagan. 2019

374 citations recorded by the source corpus.

The increasing interest in user privacy is leading to new privacy preserving machine learning paradigms. In the Federated Learning paradigm, a master machine learning model is distributed to user clients, the clients use their locally stored data and model for both inference and calculating model updates. The model updates are sent back and aggregated on the server to update the master model then redistributed to the clients. In this paradigm, the user data never leaves the client, greatly enhancing the user' privacy, in contrast to the traditional paradigm of collecting, storing and processing user data on a backend server beyond the user's control. In this paper we introduce, as far as we are aware, the first federated implementation of a Collaborative Filter. The federated updates to the model are based on a stochastic gradient approach. As a classical case study in machine learning, we explore a personalized recommendation system based on users' implicit feedback and demonstrate the method's applicability to both the MovieLens and an in-house dataset. Empirical validation confirms a collaborative filter can be federated without a loss of accuracy compared to a standard implementation, hence enhancing the user's privacy in a widely used recommender application while maintaining recommender performance.

Subfields: Artificial Intelligence, Information Systems, Recommender Systems, Federated & Distributed Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Recommender Systems

In Federated & Distributed Learning

Citations in the corpus

FedML: A Research Library and Benchmark for Federated Machine Learning (2020, background): “Promising application domains include, but are not limited to, computer vision [85, 86], natural language processing [87, 88, 89, 90, 91], finance [48, 51, 92], transportation [93, 94, 95, 96, 97, 98, 99, 100, 101, 102], digital health [103, 104, 105, 106, 107, 108, 109], recommendation [110, 111, 1”
Federated Learning and Differential Privacy: Software tools analysis, the Sherpa.ai FL framework and methodological guidelines for preserving data privacy (2020, background): “Another notable example is represented by the federated version of matrix factorization-based ranking algorithms for recommendation systems [39, 40].”
Robust Aggregation for Federated Learning (2019, background): “The paradigm has found widespread applications ranging from mobile apps deployed on millions of devices [6, 87], to sensitive healthcare applications [40, 73].”
Privacy Preserving Machine Learning with Homomorphic Encryption and Federated Learning (2021, background): “Besides this, natural language processing [42] and recommendation systems [43] are also applicable for federated learning as well.”
Federated Learning with Personalization Layers (2019, background): “For personalization via recommendation, ud din et al. (2019) present the ﬁrst attempt to extend collaborative ﬁltering to a federated setup and Chen et al. (2018) design a meta-learning approach for federated training.”
Robust Federated Recommendation System (2020, methodology): “The results demonstrate that our proposed A-FRS (red line) outperforms FMF (blue line) on all datasets.”
A Simple and Efficient Federated Recommender System (2019, background): “A federated, collaborative filtering paper [2] was proposed where the item vectors are distributed by a central server and the user vectors are updated locally on each device.”
A federated graph neural network framework for privacy-preserving personalization (2021, methodology): “PMF SVD++ GRALS sRGCNN GC-MC PinSage NGCF FCF FedMF FedGNN High-order user-item interaction Cen.”
Variational Federated Multi-Task Learning (2019, methodology): “This paradigm has been applied successfully in many real world cases, e.g to train smart keyboards in commercial mobile devices [46] and to train privacy-preserving recommendation systems [2].”
Federated User Representation Learning (2019, background): “Collaborative ﬁltering [2] can be seen as a speciﬁc instance of the generalized approach in FURL.”
FedRec: Privacy-Preserving News Recommendation with Federated Learning (2020, methodology): “…and user representations; (8) NRMS (Wu et al., 2019d), learning representations of news and users via multi-head self-attention network; (9) FCF (Ammad et al., 2019), a federated collaborative ﬁltering method for recommendation; (10) CenNewsRec , which has the same news recommendation model…”
A Survey on Federated Learning: The Journey From Centralized to Distributed On-Site Learning and Beyond (2021, background): “Moreover, FL has as well found its way into many other areas, such as edge computing [74], networking [75], robotics [12], grid-world [76], FL enhancement [77], recommender system [78], cybersecurity [79], online retailers [80], wireless communication [18], and electric vehicles [81].”
Federated Multi-view Matrix Factorization for Personalized Recommendations (2020, methodology): “We used Federated Collaborative Filter (FCF) [13, 20, 21] as a baseline method.”