About the DMML group
The Data Mining and Machine Learning group was established in 2011 by Prof. Alexandros Kalousis. It is a part of the Department of Information Systems of the University of Applied Sciences, Western Switzerland, Geneva. It is a also part of the VIPER group of the Computer Science Department of the University of Geneva, with a number of group members doing their PhD there, under the joint supervision of Profs. Alexandros Kalousis and Stephane Marchand-Maillet.
Within the group we do research in various areas of data mining and machine learning. Over the years group members have explored a variety of research problems, such as metric and kernel learning, feature selection and dimensionality reduction, regularisation, and meta-learning. Group members currently explore different research problems such as structured regularisation, learning of dynamical systems, life-long learning, imitation and reinforcement learning, many of them using deep learning methods.
In our research we pay particular attention to real life problems and applications. Many of our research directions have been defined on the basis of real world problems and in the context of joint projects with industrial partners. 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.
The case for lifelong learning Lifelong learning (also known as continual learning) is the problem of learning multiple consecutive tasks in a sequential manner where knowledge gained from previous tasks is retained and used for future learning . Living in a temporal world, we as humans absorb information in a sequential manner. Children in kindergarten for ...Read more
Meet us at the NIPS workshops were we will be presenting the following works: Sample-Efficient Imitation Learning via Generative Adversarial Nets (Lionel Blondé, Alexandros Kalousis), in the Deep RL and in the Imitation Learning and its Challenges in Robotics workshops. Continual Classification Learning Using Generative Models (Frantzeska Lavda, Jason Ramapuram, Magda Gregorova, Alexandros Kalousis), in the Continual Learning workshop. Variational Saccading: Efficient ...Read more
Read the blog to know about the latest endeavours of the DMML group!Read more
Generative Adversarial Imitation Learning (GAIL) . Albeit successful at generating behaviours similar to those demonstrated to the agent, GAIL suffers from a high sample complexity in the number of interactions it has to carry out ...Read more