The Michigan Student Artificial Intelligence Lab (MSAIL) is a student organization for discussion of artificial intelligence and machine learning. Andrew Ng said:
“ ...if you read research papers consistently, if you seriously study half a dozen papers a week and you do that for two years, after those two years you will have learned a lot... But that sort of investment, if you spend a whole Saturday studying rather than watching TV, there's no one there to pat you on the back or tell you you did a good job. ” — Andrew NgMSAIL is a community in which motivated students can read and discuss modern machine learning literature together. We welcome students of all backgrounds and ability. To join MSAIL and stay up to date, simply join our Slack team! Also be sure to check out our sister organization: the Michigan Data Science Team! We are both graciously sponsored by the Michigan Institute for Data Science.
A major goal of unsupervised learning is to discover data representations that are useful for subsequent tasks, without access to supervised labels during training. Typically, this involves minimizing a surrogate objective with the hope that representations useful for subsequent tasks will arise as a side effect. This week we discussed a paper that proposes directly targeting later desired tasks by meta-learning an unsupervised learning rule that will later be useful for those tasks.
In the past decade, we have seen rapid progress in the field of Artificial Intelligence and Machine Learning. However, the advent of AI tools being used in production environments raise many ethical questions that have yet to be answered. Dr. Kuipers will first give a 30 min lecture on how we can ensure agents that autonomously make decisions will make ethical ones and then open up for discussion on this topic.
Privacy preserving machine learning is a technique that is being developed to train machine learning models on data that is decentralized and kept private to the model creator. We strive to keep the training data, inputs and outputs to the model, and the model parameters themselves visible only to their owners. Data privacy is one of the important problems of the next decade and is becoming extremely important to building trust between AI agents and people. We will discuss the various technical concepts that are necessary for a fully privacy preserving machine learning system: homomorphic encryption, multi-party computation, federated learning, and differential privacy.