Artificial deep neural networks are a powerful tool, able to extract information from large datasets and, using this acquired knowledge, make accurate predictions on previously unseen data. Due to the very large number of parameters required, they are also particularly difficult to understand.
Following a comparable path, robots could learn to move and walk as human being do. The goal of the MEMMO project is to develop a unified approach to motion generation for complex robots with arms and legs.
The Perception and Activity Understanding group (Jean-Marc Odobez, http://www.idiap.ch/~odobez/), along with the Robot Learning and Interaction group (Sylvain Calinon, http://calinon.ch) at Idiap (www.idiap.ch), seeks one motivated postdoctoral fellow to work on interaction strategies and deep learning for object manipulations in cluttered object scenes.
The TACT-HAND project in which Idiap participates was featured in the annual magazine Micro&Nano (p.42-43) from Micronarc, the Western Switzerland Micro-nanotech Cluster:
The thesis presents generative mixture models for learning robot manipulation skills from a few human demonstrations, and providing assistance to a human operator in performing these skills at a distance under limited bandwidth and communication delays. The models are based on invariant task-parameterized representations of hidden semi-Markov models that allow segmentation and reproduction of manipulation skills, while generalizing effectively across different environmental situations such as size, position, and orientation of objects in the environment.
Following a very strict evaluation process – comprising the nomination by Idiap's management and a formal approval by the Scientific College - , Idiap is pleased to announce the promotion of three of its researchers to Senior Researchers:
The Perception and Activity Understanding group (Jean-Marc Odobez, http://www.idiap.ch/~odobez/) seeks one PhD candidate for a Swiss NSF funded project aiming to study robot skills acquisition through active learning and social interaction strategies (ROSALIS, see below). In the project, the PhD candidate will work on the multimodal perception of persons and objects, and collaborate with two other PhD students working on skill learning and interaction modeling.
The Idiap Research Institute develops a new imaging system for research. The precision instrument has been built together with the air base of Sion and could be used by companies and the Universities of Applied Science of the Canton.
The paper of Noémie Jaquier, Mathilde Connan, Claudio Castellini and Sylvain Calinon from Idiap Research Institute and DLR-German Aerospace Center, "Combining Electromyography and Tactile Myography to Improve Hand and Wrist Activity Detection in Prostheses", has been posted on /Technologies/ journal social media platform.
The Idiap Research Institute is proud to announce that 2 new researchers have recently joined the research team. Sylvain Calinon and Michael Liebling will reinforce, complement and develop the Institute’s expertise and competencies in human-robot interaction and computational bioimaging.
Idiap pursues its development with new scientific research groups. Recently, the institute has acquired an American robot worth CHF 30’000.-. This investment opens new perspectives, in particular thanks to the interconnection between the different domains of competencies developed in Martigny.
Idiap has an open internship position for a student to work in the new Robot Learning and Interaction Group. The project involves the bimanual robot Baxter controlled by torque commands, enabling the production of natural human-like movements by modifying the stiffness at the level of the articulations.
The Robot Learning & Interaction Group at Idiap is recruiting a PhD student in the field of robot learning and control. The work will take place within a CHIST-ERA project on interactive dressing assistance (I-DRESS) starting in December 2015, in collaboration with CSIC-UPC and UWE-BRL, where Idiap contributes to the robot learning aspects.
In the summer of 2015, Rethink Robotics, the worldwide leader in collaborative robotics research and education launched the Rethink Robotics Video Challenge. The Challenge was created to highlight the amazing work being done by the research and education community with the Baxter robot. The Challenge attracted more than 90 total entries from 19 countries around the globe. The video of Dr. Sylvain Calinon’s Robot Learning & Interaction group of the Idiap Research Institute in Martigny was selected amongst the 10 finalists.
The Robot Learning & Interaction Group at Idiap is recruiting a PhD Student in the field of robot learning and control. The work will take place within the TACT-HAND project starting in April 2016, in collaboration with DLR and CITEC, where Idiap will contribute to the robot learning aspects.
DexROV is a robot of a special kind. This underwater robot is destined to take the place of human divers for tasks in very deep water (up to 1’300 meters under the sea). The Idiap Research Institute in Martigny is in charge of developing the algorithms that will help teleoperating the two arms of this robot.
Dr. Sylvain Calinon, head of Idiap’s Robot Learning & Interaction group, receives Romain Boisset, reporter for the Canap9 TV channel. They demonstrate how Idiap’s Baxter robot learns by demonstration.
Best not left alone on the back of a bus
Nineteen months after the undersea-operations project was launched, DexROV has successfully completed its first funding review as part of the European Commission’s Horizon 2020 Research and Innovation programme.
Suite à un processus d’évaluation très stricte – comprenant une nomination par la Direction et une validation par le Collège scientifique -, l’Idiap est heureux d’annoncer la promotion de trois de ses chercheurs au poste de Senior Researcher :
Thanks to Idiap researchers, the Baxter robot is able to learn from its mistakes.
Le centre de recherche Idiap développe un nouveau système de prise d’images destiné à la recherche. Un appareil de précision construit avec l’aide des apprentis de la base aérienne de Sion et qui pourrait servir à des entreprises et des Hautes Écoles du canton.
L’équipe de DexROV vient juste de compléter 2 semaines de campagne d’intégration et de tests dans le cadre du projet DexROV. Pendant ces travaux, une liaison satellite a été établie entre Bruxelles et Marseille pour permettre la surveillance à longue distance d’un robot sous-marin.
The Robot Learning & Interaction Group has an open Postdoc position within the DexROV European project, in collaboration with 6 other partners in Belgium, France, Italy, Germany and the Netherlands.
Dr Sylvain Calinon received the 2017 Best Paper Award of the Intelligent Service Robotics (ISR) journal for his article "A Tutorial on Task-Parameterized Movement Learning and Retrieval".
The DexROV team has just completed a 2 weeks long campaign of integration and tests in the DexROV project, with a satellite link established between Brussels and Marseille to enable far distance monitoring and control of an underwater robot.
The Perception and Activity Understanding group (Jean-Marc Odobez, http://www.idiap.ch/~odobez/) and the Robot Learning and Interaction group (Sylvain Calinon, http://calinon.ch) seek 3 PhD candidates for a Swiss NSF funded project aiming to study robot skills acquisition through active learning and social interaction strategies (ROSALIS, see below). The 3 PhD students will collaborate together, with 3 axes of research:
To strengthen and diversify its growth, the Idiap Research Institute invites applications for a permanent Senior Researcher position in the areas of high-performance computing for AI, visual analytics, or cyber-security and data privacy. The Robot Learning & Interaction Group has also two new open positions, a Postdoc and a phd position within the new MEMMO (Memory of Motion) ICT-H2020 European project.
"Deep neural networks remain for the most part black boxes"
