PROJECT TITLE :

RASP-Boost: Confidential Boosting-Model Learning with Perturbed Data in the Cloud - 2018

ABSTRACT:

Mining massive information needs intensive computing resources and knowledge mining experience, which would possibly be unavailable for many users. With widely out there cloud computing resources, information mining tasks can now be moved to the cloud or outsourced to third parties to avoid wasting prices. In this new paradigm, knowledge and model confidentiality becomes the key concern to the information owner. Knowledge homeowners should perceive the potential trade-offs among shopper-aspect costs, model quality, and confidentiality to justify outsourcing solutions. During this Project, we propose the RASP-Boost framework to deal with these issues in confidential cloud-based learning. The RASP-Boost approach works with our previous developed RAndom Area Perturbation (RASP) method to protect knowledge confidentiality and uses the boosting framework to overcome the problem of learning high-quality classifiers from RASP perturbed knowledge. We have a tendency to develop many cloudclient collaborative boosting algorithms. These algorithms need low consumer-side computation and communication costs. The client does not want to stay online in the process of learning models. We have a tendency to have totally studied the confidentiality of information, model, and learning process under a practical security model. Experiments on public datasets show that the RASP-Boost approach can provide high-quality classifiers, while preserving high information and model confidentiality and requiring low consumer-aspect prices.