Home Artificial Intelligence AI Ethics Researchers From China Introduce ‘FedPerGNN’: A New Federated Graph Neural Network (GNN)...

Researchers From China Introduce ‘FedPerGNN’: A New Federated Graph Neural Network (GNN) Framework For Both Effective And Privacy-Preserving Personalization

Source: https://www.nature.com/articles/s41467-022-30714-9.pdf
This Article is written as a summay by Marktechpost Staff based on the paper 'A federated graph neural network framework for privacy-preserving personalization'. All Credit For This Research Goes To The Researchers of This Project. Check out the paper and github.

Please Don't Forget To Join Our ML Subreddit

Machine Learning is increasingly being integrated into our daily lives without understanding. Our data is based on all the customization we have today, whether adverts, keyboard word suggestions, or other forms of individualized digital content. Personalization is an essential direction in Web development. It can alleviate the stress of information overload by offering various services to different users depending on their preferences and traits to meet their requirements better. Current technology users are concerned about their data privacy and are hesitant to provide it to train these algorithms.

Storing raw data locally on user devices and developing local GNN models based on it is an appealing solution to address the privacy issue of these systems. However, in most circumstances, consumer device data volume is insufficient to train accurate GNN models locally. A new machine-learning algorithm, Federated Learning, was developed to address this issue. Federated learning is a privacy-preserving machine learning paradigm that can develop intelligent models jointly from data dispersed across many user clients while maintaining privacy.

The graph neural network (GNN) helps simulate high-order interactions and is commonly utilized in customized applications such as recommendations. However, due to the sensitive nature of user data, popular customization solutions rely on centralized GNN learning on global graphs, which poses significant privacy problems. Researchers from China developed FedPerGNN, a federated GNN system for effective and privacy-preserving personalization.

Source: https://www.nature.com/articles/s41467-022-30714-9.pdf
Free-2 Min AI NewsletterJoin 500,000+ AI Folks

Training of GNN models is done cooperatively based on decentralized graphs inferred from local data using a privacy-preserving model updating technique. For expanding the use of graph information beyond local interactions, a privacy-preserving graph expansion technique is proposed that incorporates high-order information while maintaining privacy.

FedPerGNN offers a viable approach to mining decentralized graph data while protecting privacy for ethical and intelligent customization. FedPerGNN yields 4.0% – 9.6% reduced errors than state-of-the-art federated customization algorithms under adequate privacy protection, according to experimental results on six datasets for personalization in diverse circumstances. Each client sends the locally calculated gradients to a server for aggregation, and the aggregated gradients are then delivered to user clients for local updates.

The paper presents a privacy-preserving model update mechanism to ensure user privacy in model training since the conveyed model gradients may contain sensitive user information. Unlike previous solutions, which can safeguard private user ratings, this method can protect ratings and interaction histories, resulting in more extensive privacy preservation. Furthermore, the solution does not need communication or local memory of the global item set, and its communication overhead is generally reasonable for current portable devices. FedPerGNN may be more easily applied in real-world customization services because inferred local user-item graphs only store low-order interaction information. They present a privacy-preserving user-item graph extension protocol to expand local graphs and convey high-order information while maintaining privacy. During this process, each client gets anonymous user embeddings to grow the local subgraph, which aids in the propagation of high-order information on the user-item graph while maintaining privacy to improve the performance of the GNN model. The high-order information on the user-item graph may be successfully utilized without incurring substantial communication costs after only a few rounds of privacy-preserving graph growth.

Furthermore, this method is not restricted to the customization scenario. It may be used as a fundamental strategy for privacy-preserving data mining on decentralized graph data, thus facilitating research in various domains involving graph-structured data. The code to this algorithm is available on GitHub.

With ever-rising concerns over data privacy, algorithms like these help maintain data secrecy while maintaining the accuracy of the models.