Model and algorithm selection & Combining structured and unstructured data

  Machine learning

Hey Prague, we are bringing a new dose of machine learning! With speakers coming all the way from Switzerland - engineers from SQN company.

Please register here by clicking RSVP Yes! The number of seats is limited. (Entry is free)

SPECIAL NOTE: Given the circumstances, the capacity of the meetup is reduced to 50% of the standard size. Furthermore, we are going to do our best to stream the event online so that you can participate from anywhere you want. (please RSVP Yes only if you come in person)

18:00-18:15 Arrival of attendees
18:15-18:45 Model and algorithm selection
18:45-18:55 Break
18:55-19:25 Combining structured and unstructured data
19:25-21:00 Networking

Model and algorithm selection
Are you struggling with choosing the right algorithm for your problem? Do you want to automate hyper-parameter optimisation in a reproducible and computational-budget friendly way? During this talk, I make a deep dive into different approaches and discuss the most promising ones based on theory and empirical reasoning for model and algorithm selection.

Combining structured and unstructured data
The World is full of data coming in different formats - structured, text, audio, video, images. True machine learning needs to include relevant information in any format available. The presentation will cover the challenge of designing and building models taking different data types (structured and unstructured) as inputs. Multimodal machine learning will be discussed as a possible solution for this problem, along with the inevitable challenges that come with it.

Elli Tzini
Elli is a ML Engineer experienced in machine learning; specifically deep learning. After finishing her studies at ETH Zurich in Computer Science, she worked in collaboration with the Cancer Biology institute of University of Zurich on Electro-Microscopy image analysis using Deep Learning techniques. Currently, focusing on the NLP/ NLU area, building ML solutions for behavioural science.

Bertrand Buisson
Bertrand has been part of the machine-learning engineering team at SQN since November 2019 where he is responsible for assessing behaviours using a combination of structured and unstructured data. He is a mechanical engineering graduate of the Ecole Polytechnique Fédérale de Lausanne (EPFL) and the Ecole Polytechnique (France). His previous experiences lie in numerical simulations of fluid flows and engineering consulting for nuclear power plants and oil and gas companies.