We will introduce the problem of online model selection where a learner is to select among a set of online algorithms to solve a specific problem instance. We would like to design algorithms that allow such a learner to select in an online fashion the best algorithm without incurring much regret. This problem is challenging because in contrast with for example multi armed bandits, the algorithms' rewards -due to the algorithm's own learning process- may be non-stationary. We will introduce the principle of regret balancing, a simple, practical and effective model selection algorithmic design technique that allows for online selection of the best among multiple (base) algorithms in a fully blackbox fashion. Regret balancing solves the problem of non-stationarity by introducing an elegant misspecification test' that can efficiently detect when a base algorithm is not appropriate for the problem at hand.

Fethiye International Group "Helping Children in our Local Community" The FIG coffee morning will be held the first Thursday of every month at the Meydan Restaurant from 11am to 1pm.

Artificial intelligence (AI) is a topic that's increasingly being explored around the world and on Thursday, Prince Daniel attended the Royal Swedish Academy of Engineering Sciences's (IVA) Science & Society Forum seminar to learn more about the subject. During the event, held at IVA Konferenscenter in Stockholm, the Prince heard lectures from researchers, professors and experts on AI. "AI offers incredible opportunities – everything from better medical diagnostics and self-driving cars to individualised services and efficient warehouse management," IVA said on their website. "The potential for value-creation and efficiency gains in organisations, in society and for individuals is enormous. But increased awareness is needed about the fact that these gains must be weighed against ethical considerations. How can we create an overview and traceability, and steer AI towards more long-term, sustainable ethical principles? How can AI be part of people's daily lives without the social, ethical and legal consequences getting out of control?"

In supervised learning, we leverage a labeled dataset to design methods for function estimation. In many practical situations, we are able to obtain alternative feedback, possibly at a low cost. A broad goal is to understand the usefulness of, and to design algorithms to exploit, this alternative feedback.