The Knowledge Discovery and Data Mining (KDD) conference is held in Anchorage, Alaska this year. Leading experts in the world of applied data mining and knowledge discovery will be presenting both theoretical and applied advances during tutorials, workshops and lectures.
SigOpt is excited to be participating in KDD this year. Last year, SigOpt was an exhibitor, and this year we are presenting at the 2nd Workshop on Offline and Online Evaluation of Interactive Systems. We are very grateful to the organizers for their generous invitation and look forward to presenting recent results from the SigOpt research team.
Michael McCourt will be attending the conference Monday through Wednesday and is thrilled to be attending his first KDD meeting. If you are also attending KDD and would like to have a discussion, please reach out because Mike would love to chat!
The slides for his talk, Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimization, are publicly available. The abstract is below:
Many real world applications (ML models, simulators, etc.) have multiple competing metrics that define performance; these require practitioners to carefully consider potential tradeoffs. However, assessing and ranking this tradeoff is nontrivial, especially when the number of metrics is more than two. Often times, practitioners scalarize the metrics into a single objective, e.g., using a weighted sum. In this talk, we pose this problem as a constrained multi-objective optimization problem. By setting and updating the constraints, we can efficiently explore only the region of the Pareto efficient frontier of the model/system of most interest. We motivate this problem with the application of an experimental design setting, where we are trying to fabricate high performance glass substrate for solar cell panels.