Machine Learning Coffee Minisymposium on Agile Probabilistic AI

2018-03-26 09:00:00 2018-03-26 11:00:00 Europe/Helsinki Machine Learning Coffee Minisymposium on Agile Probabilistic AI Series of longer machine learning coffee sessions on themes of the research programmes of Finnish Center for Artificial Intelligence (FCAI) http://old.cs.aalto.fi/en/midcom-permalink-1e82b5fb96cd79c2b5f11e8801785583693ddf5ddf5 Konemiehentie 2, 02150, Espoo

Series of longer machine learning coffee sessions on themes of the research programmes of Finnish Center for Artificial Intelligence (FCAI)

26.03.2018 / 09:00 - 11:00
seminar room T6, Konemiehentie 2, 02150, Espoo, FI

Map © OpenStreetMap. Some rights reserved.

Finnish Center for Artificial Intelligence (FCAI) is proud to present a series of longer machine learning coffee sessions on themes of the research programmes of FCAI. These minisymposia will last for ~2 hours and consist of some longer talks and a set of 5-min flash talks. The purpose is both for the researchers already working on the theme to get to know what the others are working on, and for those interested in the theme to get to know who to talk to. Suggestions of flash talk topics are welcome - contact the organizers of the specific symposium.

The first minisymposium is on Agile probabilistic AI, on March 26, 9:00-11:00, in T6, CS building, Otaniemi.

Speakers include (feel free to suggest more flash talks to organizers: aki.vehtari@aalto.fi / arto.klami@helsinki.fi):

Aki Vehtari: Probabilistic programming and Stan

  • Probabilistic programming (PP) makes it easy to write new probabilistic models and PP frameworks then allow automated inference for those models. I describe the generic idea of PP, give some examples of software frameworks designed for different PP purposes, and focus more on recent development in Stan.

Arto Klami: Automated variational inference

  • Automated inference for generic models which can be programmed with probabilistic programming languages is challenging. I describe methods based on modern variational inference which are used to speed-up inference for big data.

Henri Vuollekoski: ELFI - Engine for Likelihood Free Inference

  • ELFI - Engine for Likelihood Free Inference provides a probabilistic programming framework for combining probabilistic models with stochastic simulators and performs automated inference using efficient likelihood free inference algorithms.

The second minisymposium will be on Simulator-based inference on April 9 in Kumpula. Speakers will include Jukka Corander, Samuel Kaski and Jaakko Lehtinen.

During the Spring, we will have minisymposia also on Privacy-preserving and secure AI on May 7 and Interactive AI on May 14, interleaved with normal Machine learning coffee talks.

See the full programme of the Coffee series.