project_ Machine unlearning

Please read it First Big project- Implementation of Machine Unlearning and an Evaluation Against the Naïve Baseline. Machine Learning is the field of study that gives computers the ability to learn without being explicitly programmed. For example, supervised classification models input training data to predict the likelihood that subsequent data will fall into one of the predetermined categories e.g. logistic regression. In some cases, the model learns the training data too well, and as a consequence, the ML models can become indirect stores of the training data. In the case that the training data contains personal information, the model may expose personal data and successfully infer individual records on which they were trained on (Shokri et al., 2017). This breach of privacy has led to the motivation to find ways to delete certain points of training data in an already deployed ML model. The aim is to undertake an experimental study to implement a supervised machine learning model and compare the efficacy of retraining the model without certain records of the training data, against an approximate machine unlearning solution. the aim is to investigate the important qualitative and quantitative metrics that are required for the evaluation (e.g. privacy and efficiency metrics). Project dataset - [login to view URL]