Events
14 October 2020

The field of applied algebra encompasses a wide range of subjects in pure mathematics with applications in areas such as data analysis, biology, optimization and coding theory, to name a few. A hallmark of this research area is the mutual interaction between the development of abstract mathematical tools and the problems coming from real-world applications. In the talk I will illustrate this fascinating subject by means of a few examples that are especially related to my own research, including applications to computational biology, consensus methods in data analysis and neural networks. The exposition will not assume any preliminary knowledge, and everybody is welcome to attend.
Manno, Galleria 1, 2nd floor @12h00 Room 222

30 September 2020

Due to the current situation this event will be postponed to September/October 20202 Recent striking success in Artificial Intelligence have made the public believe that in a not so far future machines could be even more intelligent than human beings. The actual and possible developments of Artificial Intelligence open up a series of striking, deep and pressing questions such as: – Can a computer ever think in the way a human being does? – Can a computer have a mind and conscious experiences, such as thoughts, desires, and emotions? – What is artificial intelligence? Is it the same as human intelligence? Are they even comparable or are they something essentially different? – Can a machine be morally responsible for its actions? Can a machine be good or evil? What other moral considerations are related to AI? With the goal of enhancing their scientific and educational collaborations around those important topics, the Swiss AI Lab IDSIA USI-SUPSI and the USI Master in Philosophy Program are organising in Lugano on May 29-30 2020 an international meeting on current trends and perspectives in the Philosophy of AI.
Auditorium USI

23 September 2020

Salamander inspired platforms have been used to study robotic/biological locomotion in the Biorobotics Laboratory at EPFL for over a decade, drawing parallels between humanmade and natural systems to improve our understanding of both. Our most recent iteration, Krock 2, is targeted towards search and rescue applications within the NCCR Robotics program. Previously demonstrated in terrestrial applications, it is now being introduced into the water through the use of a dry suit and undulatory swimming gait. Multimodal locomotion carries the inherent challenge that the morphology and actuation for one mode may not be useful, or even present a hindrance, for another. On the other hand, unique synergies between modes of locomotion can present themselves. Its sprawling-posture gait and capability to swim along the water's surface manage to avoid the persistent challenge of balance within legged locomotion. However, this comes with inherent challenges in the perception of the environment, which we are addressing in collaborations.
Manno, Galleria 1, 2nd floor @11h30 Room 222

15 July 2020

The form of a human body is a wide field to think about. To think about the form of my body. How do my body look like? How does it sound? How does it feel to have a body and to be one? Our bodies are made of many different materials which makes us flexible and lively. The form of our bodies is constantly changing and also now in times of digitality. What kind of new forms can be seen nowadays and how is physicality kept in mind in a world full of technics and robots? How does my connection to the body change or develop and how are terms like intimacy, vulnerability or empathy get a new perspective or are they been forgotten? My body in interaction with the surroundings in a lab of robotics and AI. What can be found out of the human being living face to face with a technical world which we just started to discover. How can I rethink touch and inner feelings.
Manno, Galleria 1, 2nd floor @12h00 Room 222

3 June 2020 - 3 June 2020

NNAISENSE is a startup originating from IDSIA. The Swiss AI lab has a long lasting track record of ground-breaking results in artificial intelligence (AI). From perception to reinforcement learning, the company follows IDSIA’s steps in the search for super-human performance to take AI technology into manufacturing and control systems. While AI approaches control problems from an information theoretical and statistical perspective, control theory studies the closed-loop behaviour of the physical world with a strong focus on safety, hard constraints and theoretical guarantees. Control approaches can be very robust but they can seldom be conservative due to their assumptions. This is believed to be less of a problem for Neural Networks (NN) and AI, where non-conservative results can be achieved but it is generally harder to have formal results. NN performance depends mainly on the quality and the amount of data and no unified framework exist for the analysis of stability and robustness of NN-driven control and Reinforcement Learning (RL). For this reason, while deep learning is becoming the industry standard for perception, its use in control is mostly limited to simulated or non-critical tasks. Combining the fields of control and AI has the potential for retaining the best of both Worlds. This talk will introduce NNAISENSE’s most significant publications in this emerging field, with a special focus on the latest one: “Neural Lyapunov Model Predictive Control”.
https://open.meet.switch.ch/IDSIATalk

The recent advances in Deep Learning made many tasks in Computer Vision much easier to tackle. However, working with a small amount of data, and highly imbalanced real-world datasets can still be very challenging. In this talk, I will present two of my recent projects, where modelling and training occur under those circumstances. Firstly, I will introduce a novel 3D UNet-like model for fast volumetric segmentation of lung cancer nodules in Computed Tomography (CT) imagery. This model highly relied on kernel factorisation and other architectural improvements to reduce the number of parameters and computational load, allowing its successful use in production. Secondly, I will discuss the use of representation learning or similarity metric learning for few-shot classification tasks, and more specifically its use in a competition at NeurIPS 2019 and Kaggle. This competition aimed to detect the effects of over 1000 different genetic treatments to 4 types of human cells, and published a dataset composed of 6-channel fluorescent microscopy images with only a handful of samples per target class.
Manno, Galleria 1, 2nd floor @12h00

st.wwwsupsi@supsi.ch