Abstract Machine learning models have achieved stunning successes in the IID setting. Yet, beyond this setting, existing models still suffer from two grand challenges: brittle under covariate shift and inefficient for knowledge transfer. In this talk, I will introduce three approaches to tackle these challenges, namely self-supervised learning, causal representation learning, and test-time training. More

Abstract: What is perception? The most intuitive and influential answer to this question has long been the one given by David Marr: To see the world is “to know what is where by looking” - to transform light into representations of objects and their features, located somewhere ins pace. But is this all that perception

Abstract: The brain performs critical calculations on visual information as we swiftly, yet effortlessly, navigate around objects and obstacles in our cluttered environment. Perhaps one of the most important calculations is for the perception of depth using the apparent relative motion of objects in the environment created by our own translation known as motion parallax.

Abstract: The world is modular. So – intuitively – it seems clear that cognitive systems that deal with the world should benefit from a modular architecture. Simple or less important problems should use less cognitive resources than complex or important problems, which – intuitively – may be achieved by changing the degree of modularity that

The talks will be held by Arinze Lawrence Folarin, "My 175 days journey in Berlin"; Juliana T.C. Marcos "GSEP Internship: More than a research experience in neuromorphic vision at SCIoI"; and Kiprono Elijah Koech "Action Recognition in a Wildlife Setting - Taken a Leap". The Zoom Link will be sent the day before the lecture.

The talks will be held by Emmanuel Ousu Ahenkan and Tatiana Ngoli Moteu Marcos. The Zoom Link will be sent the day before the lecture.

Meet Philipp Semmer and Frederic du Bois-Reymond, both partners at the venture capital firm Earlybird UNI-X. They will talk about funding for university spin-offs and why they believe that scientist and researcher should be more excited about entrepreneurship. They will also share their insights on deep tech companies becoming the next generation of unicorns. The

Abstract: One of SCIoI's goals is to establish and evaluate synthetic models for biological data. Achieving this goal is clearly beneficial not only to produce more capable artificial behavior, but also to get a better idea about the processes that may underlie biological intelligence. Why can a cockatoo or mouse open a lock box? What