Keynote Lectures

Con Espressione! AI, Machine Learning, and Musical Expressivity
Gerhard Widmer, Johannes Kepler University and LIT | AI Lab, Linz Institute of Technology, Austria, Austria

Bioinspired Soft Robotics: Nature-inspired Solutions for Artificial Intelligent Systems
Barbara Mazzolai, Istituto Italiano di Tecnologia, Italy, Italy

From Probabilistic Circuits to Probabilistic Programs and Back
Guy Van Den Broeck, UCLA, United States, United States

Explainable Machine Learning for Trustworthy AI
Fosca Giannotti, ISTI-CNR, Italy, Italy

Abstract
Much of current research in Artificial Intelligence and Music,and particularly in the field of Music Information Retrieval (MIR), focuses on algorithms that interpret musical signals and recognise musically relevant objects and patterns at various levels - from notes to beats and rhythm, to melodic and harmonic patterns and higher-level structure -, with the goal of supporting novel applications in the digital music world.
This presentation will give the audience a glimpse of what computational music perception systems can currently do with music, and what this is good for. However, we will also find that while some of these capabilities are quite impressive, they are still far from showing (or requiring) a deeper "understanding" of music. An ongoing project will be presented that aims to take AI & music research a step further, going beyond the surface and focusing on the *expressive* aspects, and how these are communicated in music.
We will look at recent work on computational models of expressive music performance and some examples of the state of the art, and will discuss possible applications of this research. In the process, the audience will be subjected to a little experiment which may, or may not, yield a surprising result.

Abstract
During the last decades, the field of robotics has been increasingly enriched by the use of biomimetics. The large variety of organisms, with smart structures and functions developed during their evolution and adaptation to different environments, represents a huge source of inspiration for the design of innovative sustainable materials and technologies.
Therefore, in the era of digitalization, virtual reality, and automation that we are living, the relationship between robotics, bioinspired design, and bio-robotic hybrid systems is getting much stronger.
Fabrication methods and addictive manufacturing are experimented as novel tools by the creativity of roboticists, while others are challenging robotically-augmented design, where robots can adapt and evolve with the world around them.
Soft, multi-functional, adaptable and growing structures, taking inspiration from the plant kingdom and from the strategies of some animal species, have started a novel trend for biomimetic robotics with relevant implications towards more sustainable applications.
Here, I will present our recent results in the field of octopus-like systems and plant-like self-creating robots. The potential impact on society of these bioinspired robots could be huge and wide, e.g., in rescue, medical applications, space, or environmental monitoring. Since the design of these robotic solutions is deeply based on a few selected biological features, a new view of robots for biology can be envisaged, with the goal to give insights on the organisms themselves and open new exciting opportunities both in science and engineering.

Abstract
Probabilistic graphical models (PGMs) are a rich staple of probabilistic AI centered around variable-level (in)dependencies. In this talk I present some recent work on probabilistic models that go beyond classical PGMs, and make a radically different choice of abstraction; one that is computational. Concretely, I will discuss two up-and-coming classes of models: probabilistic circuits and probabilistic programs. Probabilistic circuits represent distributions through the computation graph of probabilistic inference. They move beyond PGMs and deep generative models by guaranteeing tractable inference for certain classes of queries, thereby enabling new solutions to some key problems in machine learning. Probabilistic programs represent distributions through higher-level primitives of computation: iteration, branching, and procedural abstraction. They move beyond PGMs by looking "inside" of the dependencies. Finally, I will illustrate how these two computational abstractions are themselves closely related, by showing how the Dice probabilistic programming language compiles probabilistic programs into probabilistic circuits for inference.

Abstract
Black box AI systems for automated decision making, often based on machine learning over (big) data, map a user’s features into a class or a score without exposing the reasons why. This is problematic not only for the lack of transparency, but also for possible biases inherited by the algorithms from human prejudices and collection artifacts hidden in the training data, which may lead to unfair or wrong decisions. The future of AI lies in enabling people to collaborate with machines to solve complex problems. Like any efficient collaboration, this requires good communication, trust, clarity and understanding. Explainable AI addresses such challenges and for years different AI communities have studied such topic, leading to different definitions, evaluation protocols, motivations, and results. This lecture provides a reasoned introduction to the work of Explainable AI (XAI) to date, and surveys the literature with a focus on machine learning and symbolic AI related approaches. We motivate the needs of XAI in real-world and large-scale application, while presenting state-of-the-art techniques and best practices, as well as discussing the many open challenges.