No abstract

Abstract. The goal of distributed learning in P2P networks is to achieve results as close as possible to those from centralized approaches. Learning models of classification in a P2P network faces several challenges like scalability, peer dynamism, asynchronism and data privacy preservation. In this paper, we study the feasibility of building SVM classifiers in a P2P network. We show how cascading SVM can be mapped to a P2P network of data propagation. Our proposed P2P SVM provides a method for constructing classifiers in P2P networks with classification accuracy comparable to centralized classifiers and better than other distributed classifiers. The proposed algorithm also satisfies the characteristics of P2P computing and has an upper bound on the communication overhead. Extensive experimental results confirm the feasibility and attractiveness of this approach.

No abstract

In a distributed computing environment, peers collaboratively learn to classify concepts of interest from each other. When external changes happen and their concepts drift, the peers should adapt to avoid increase in misclassification errors. The problem of adaptation becomes more difficult when the changes are asynchronous, i.e., when peers experience drifts at different times. We address this problem by developing an ensemble approach, PINE, that combines reactive adaptation via drift detection, and proactive handling of upcoming changes via early warning and adaptation across the peers. With empirical study on simulated and real-world data sets, we show that PINE handles asynchronous concept drifts better and faster than current state-of-the-art approaches, which have been designed to work in less challenging environments. In addition, PINE is parameter insensitive and incurs less communication cost while achieving better accuracy.

Distributed classification aims to learn with accuracy comparable to that of centralized approaches but at far lesser communication and computation costs. By nature, P2P networks provide an excellent environment for performing a distributed classification task due to the high availability of shared resources, such as bandwidth, storage space, and rich computational power. However, learning in P2P networks is faced with many challenging issues; viz., scalability, peer dynamism, asynchronism and fault-tolerance. In this paper, we address these challenges by presenting CEMPaR-a communication-efficient framework based on cascading SVMs that exploits the characteristics of DHT-based lookup protocols. CEMPaR is designed to be robust to parameters such as the number of peers in the network, imbalanced data sizes and class distribution while incurring extremely low communication cost yet maintaining accuracy comparable to the best-in-the-class approaches. Feasibility and effectiveness of our approach are demonstrated with extensive experimental studies on real and synthetic datasets.