The concept of brain-inspired or neuro-inspired information processing dates back to the very beginning of computing, as described by Alan Turing in its report “Intelligent Machinery” of 1948. In 1990, Carver Mead showed that biological information processing can be efficiently coupled to newly designed electronic devices, defining for the first time the concept of Neuromorphic System. Taking inspiration from the human brain means trying to mimic its basic elements, namely neurons and synapses. This endeavor to emulate the brain entails the development of spiking neural networks (SNNs), which transmit information through temporal and spatial patterns similar to those found in the brain. This brain-inspired approach to computation, which relies on binary events referred to as spikes, is known as neuromorphic computing. Its multifaceted nature can be observed and understood from two perspectives: first, through computation, it can lead to a better understanding and wider adoption of biological primitives; second, through the physical realization of brain-inspired devices, it can offer a viable path towards low-energy solutions for efficient intelligent systems. Neuromorphic brain-inspired computing is thus a truly multidisciplinary field that embeds different possible ways of being alternative to traditional approaches: ● From the software standpoint, it introduces the concept of sparse and discrete computation. ● From the hardware perspective, it aims at achieving elemental units with bio-plausible behaviors. ● From the computing architecture point of view, it targets the overcome of the von Neumann architecture in favor of the in-memory paradigm naturally chosen by the human brain as the best solution. Finally, getting inspired by the famous quote “What I cannot create, I don’t understand” chalked by Nobel prizewinning Richard Feynman on his blackboard, neuromorphic approaches are envisioned to help in the future the understanding of human brains and the treating of neurological disorders.

The student will gain the following knowledge about: - fundamentals of artificial synapses and neurons - fundamentals of neuromorphic computing (algorithms and devices) - techniques for integrating SNN-based neuroAI models in IoT applications - neuromorphic SW/HW platforms, sensors, encoding techniques, The student will gain the following skills: - ability to model neurons and synapses - ability to compute with brain-inspired algorithms and devices - ability to integrate neuromorphic technologies in next-generation solution design

Basic knowledge of programming, materials science, electronics and circuit theory. Elements of machine learning and nanotechnology are helpful, even if not strictly needed.

Given the intrinsic multidisciplinarity of brain-inspired approaches, lecturers and relative topics of information processing and nanodevices will be mixed all along the course, following a bottom-up approach from single unit to networks. Three main blocks can be identified: I) Emulating neurons and synapses ○ Introduction to neuromorphic computing as “beyond von Neumann” architecture ○ Introduction to biological neurons and synapses ○ Hardware emulation of neurons and synapses ○ Software emulation of neurons and synapses ○ LAB 1 - Modeling of single neuron II) Brain-inspired computing ○ Neuromorphic Engineering or “Why get inspired by the brain?” ○ Computing with Spiking Neurons: general concepts ○ Neuromorphic vs von Neumann computing ■ Algorithms: networks and learning methods ■ Architecture: digital, analog, mixed ■ Encoding: rate-coding, temporal-coding ■ Devices: physical nanodevices able to learn ■ Sensing: dynamic vision sensors (DVS), silicon cochlea, silicon nose, tactile ■ NeuroBench: evaluation and performance measurements of SNN models ○ LAB 2: in materia computing with nanodevices coupled to neuromorphic classifiers ○ LAB 3: SNN in IoT application with neuromorphic sensors and SW/HW platforms. III) State of the art of current available neuromorphic approaches ○ Successful Neuromorphic applications and limits ○ Future trends and challenges

Theoretical lectures with exercises and computational labs will be alternated during the course (2/3 and 1/3, respectively). For computational labs, students will be divided in small groups and work under the supervision of the teachers.

Slides; Dispense; Esercizi; Esercitazioni di laboratorio; Strumenti di simulazione;

Modalità di esame: Prova orale obbligatoria; Elaborato progettuale in gruppo;

Exam: Compulsory oral exam; Group project;

... Students will be asked to select and read a technical paper (provided during the course) related to the topics covered in the course and to carry out a design project in a group (2-3 people), developed on a simulation platform and finally on an "evaluation board". The exam consists of an oral exam and a group presentation of the developed project. Each group will be asked to submit the developed code as well as a report (6-10 pages, double-spaced) with a research paper-like description of their work, including challenges and motivations, state of the art, proposed solution and discussion of the results obtained. The project will then be presented and discussed by all members of the group in an oral session. After the ppt group presentation (approximately 15 minutes), each group member will be asked questions about the project and their individual contribution to the group activity, as well as general questions about the topics covered in the course. Each student's final mark will consist of - Evaluation of the group project (same for all group members, 70%). This evaluation will take into account: the complexity of the problem addressed; the originality and richness of the proposed solution; the methodological and technical correctness of the solution; the completeness and quality of the report; the completeness and quality of the oral presentation. - Individual oral evaluation (30%). This evaluation takes into account: the quality of the individual presentation; the individual effort and contribution to the group activity; the correctness of the answers to the theoretical and technical questions; the individual communication skills.

Gli studenti e le studentesse con disabilità o con Disturbi Specifici di Apprendimento (DSA), oltre alla segnalazione tramite procedura informatizzata, sono invitati a comunicare anche direttamente al/la docente titolare dell'insegnamento, con un preavviso non inferiore ad una settimana dall'avvio della sessione d'esame, gli strumenti compensativi concordati con l'Unità Special Needs, al fine di permettere al/la docente la declinazione più idonea in riferimento alla specifica tipologia di esame.