Distributed Ridge Regression with Feature Partitioning

By Cristiano Gratton, Naveen K. D. Venkategowda, Reza Arablouei, Stefan Werner. 2018

19 citations recorded by the source corpus.

We develop a new distributed algorithm to solve the ridge regression problem with feature partitioning of the observation matrix. The proposed algorithm, named D-Ridge, is based on the alternating direction method of multipliers (ADMM) and estimates the parameters when the observation matrix is distributed among different agents with feature (or vertical) partitioning. We formulate the associated ridge regression problem as a distributed convex optimization problem and utilize the ADMM to obtain an iterative solution. Numerical results demonstrate that D-Ridge converges faster than its diffusion-based contender does.

Subfields: Artificial Intelligence, Computational Mechanics, Computer Networks and Communications

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

1. Yang Liu
2. Tianjian Chen
3. Qiang Yang
4. Yan Kang
5. Yong Cheng
6. Zhi-qian Yi
7. Peng Liu
8. Xinwei Zhang

In Computational Mechanics

1. Yang Liu
2. Xinwei Zhang
3. Tianjian Chen
4. Peng Liu
5. Qiang Yang
6. Mingyi Hong
7. Liping Li
8. Wotao Yin

In Computer Networks and Communications

1. Peng Liu
2. Qiang Yang
3. Di Niu
4. Mingyi Hong
5. Linglong Kong
6. Yaochen Hu
7. Yang Liu

Citations in the corpus

• Backdoor attacks and defenses in feature-partitioned collaborative learning (2020, background): “, 2019), linear and logistic regression (Yang et al., 2019b; Hardy et al., 2017; Gratton et al., 2018; Kikuchi et al., 2017; Hu et al., 2019a), and neural networks (Liu et al.”
• Learning Privately over Distributed Features: An ADMM Sharing Approach (2019, positive): “Gratton et al. (2018) applies ADMM to solve ridge regression.”
• Vertical Federated Learning: Concepts, Advances, and Challenges (2022, methodology): “The local model G k can take various forms including tree [17], linear and logistic regression [3], [16], [18], [19], [20], [21], support vector machine [22], [23], neural network (NN) [24], [25], [26], K-means [27] and EM algorithm [28].”
• FedBCD: A Communication-Efficient Collaborative Learning Framework for Distributed Features (2022, methodology): “Feature-partitioned learning architectures have been developed for models including trees [9], linear and logistic regression [4], [6], [7], [8], and neural networks [5], [10].”
• A Communication Efficient Collaborative Learning Framework for Distributed Features (2019, background): “nvestigated. All the work above consider the sample-partitioned scenario. Feature-partitioned learning architectures have been developed for models including trees [9], linear and logistic regression [4, 8, 5, 7], and neural networks [10, 6]. Distributed Coordinate Descent [13] used balanced partiti”