About the DMML group
The Data Mining and Machine Learning group of Geneva was established in 2011 by Prof. Alexandros Kalousis. It operates as a collaboration between the Department of Information Systems of the University of Applied Sciences, Western Switzerland, Geneva, and the VIPER group of the Computer Science Department of the University of Geneva. Many of the members currently follow a PhD under the joint supervision of Profs. Alexandros Kalousis and Stephane Marchand-Maillet.
We conduct research in various areas of data mining and machine learning, publishing at major international conferences (NeurIPS, ICML, etc.). In our latest work we focus on leveraging the power of modern deep learning architectures to address the problems of generative modeling, continual- and meta- learning, modelling of dynamical systems, and imitation and reinforcement learning. In these we also build on our experiences with metric and kernel learning, feature selection, structured regularization, and many other topics explored over the years by our members.
Many of our research directions have been defined on the basis of real world problems. We collaborate with multiple industrial and academic partners on joint projects providing the necessary machine learning expertise. The group receives funding from different sources, such as the Swiss National Science Foundation, SNSF, InnoSuisse (former CTI), the European Union, Horizon 2020, as well as directly from industrial partners.
Ever tried to train a deep neural network over high resolution images taken by modern smartphone cameras or smart devices? The memory and inferential costs when working with inputs of such large dimensions (e.g. 4000x3000) increase rapidly and often prohibitively. Jason Ramapuram proposes a solution in his new paper "Variational Saccading: Efficient Inference for Large ...Read more
Though ML labs may seem to outsiders as rather nerdy isolated groups, it is not our case! We are on the move, visiting and welcoming ML researchers, sometimes near and sometimes further away. Discussing, collaborating, making friends, having fun!Read more
AISTATS 2019 is well under way and Lionel Blonde is there to present his work on Sample-Efficient Imitation Learning via Generative Adversarial Nets. You can check his poster Th79 at the poster session tomorrow (Thursday, April 18, 13h30-16h30). He will be happy to explain how he improves upon GAIL by reducing the sample complexity by ...Read more
The HEG (HES-SO Geneva) opened a new Data Science branch in its Master's programe in Information Studies. Our group is responsible for the two major modules on machine learning spanning over two semesters (16 credits in total). Building a new machine learning course is a challenge we are happy to embrace.Read more