Next Meetup

Machine Learning on the Edge Between Research and Industry
Ever wondered how to apply GANs on categorical data, and what this is good for at all? Attend to our meetup on June 19th. The second talk of this session will introduce you to the impressive machinery behind the ads, you see in your browsers and mobile apps. The event is streamed on Youtube: _________________________________________________________ Generating Multi-Categorical Samples with Generative Adversarial Networks In this talk, we will propose a method to train generative adversarial networks (GANs) on multivariate feature vectors representing multiple categorical values. In contrast to the continuous domain, where methods based on GANs have delivered considerable results, these approaches struggle to perform equally well on discrete data. We will describe and compare several architectures based on multiple (Gumbel) softmax output layers taking into account the size of each variable. We will evaluate the performance of our architecture on datasets with different sparsity, number of features, ranges of categorical values, and dependencies among the features. Furthermore, we will discuss possible practical applications for data science practitioners. Ramiro CAMINO is a PhD Student at the University of Luxembourg. He received his Master's Degree in Computer Science from the University of Buenos Aires, and has several years of experience in the industry of web development, social media games and mobile games. Afterwards, he moved to the research and development area, applying machine learning and natural language processing to different problems like resume parsing, document classification and entity resolution. He is currently expanding his studies to the domain of deep learning and generative models with practical applications related to data augmentation and data imputation. ________________________________________________________ Real Time Bidding Based Online Advertising in a Nutshell Online advertising is increasingly switching to real-time bidding (RTB) mechanisms, in which the ad slots are sold through real-time auctions upon users visits. To accurately target the potential users and compete with unknown bidders in a highly stochastic environment brings a lot of challenges and research opportunities in the machine learning fields. In this talk, we will provide an introduction of the eco-system of RTB, the auction mechanisms, and an overview of the algorithms applied on user behaviour prediction, bidding market modelling, and bidding strategies optimization. Manxing Du is a PhD student at the University of Luxembourg. She is working on machine learning topics particularly in Real Time Bidding (RTB) based online advertisement. Before joining SnT, she was a research engineer at the Research Institutes of Sweden (RISE) Acreo and she received her M.Sc degree in communication systems from the Royal institute of Technology (KTH) in Sweden. Her research interests include data mining, reinforcement learning, deep learning and their applications in computational advertising. ________________________________________________________ Sponsor's Corner: We would like to thank our partners for their support: Amazon and SnT. Please see their messages underneith. SnT Distinguished Lecture - Convex Optimisation Friday, 22 June 2018, 14:00 - 15:30 Room E004, JFK Building Speaker: Stephen Boyd (Stanford University) Register: Convex optimisation has emerged as useful tool for applications that include data analysis and model fitting, resource allocation, engineering design, network design and optimisation, finance, and control and signal processing. After an overview of the mathematics, algorithms, and software frameworks for convex optimisation, we turn to common themes that arise across applications, such as sparsity and relaxation. We describe recent work on real-time embedded convex optimisation.

SnT - Luxembourg University

29 Avenue John F. Kennedy · Luxembourg

    Past Meetups (51)