CURRICULUM VITAE
Dr. Fabio Schittler Neves
Born in Brazil in 1981, I currently live with my family in Frankfurt am Main, Germany.
Academic Education
Doctor of Philosophy
Göttingen Graduate School for Neurosciences and Molecular Biosciences — Germany (2007-2010) Doctor of philosophy (PhD). (Approved with “summa cum laude”); PhD Thesis: Universal Computation and Memory by Neural Switching
Master of Science
Federal University of Rio Grande do Sul (UFRGS) — Brazil (2003-2006) Master of Science in Physics; Master Thesis: Non-monotonic Layered Neural Networks
Bachelor of Science
Federal University of Rio Grande do Sul (UFRGS) — Brazil (2000-2003) Bachelor of Science in Physics
Experience
Researcher (2023-now)
Fraunhofer Institute for Industrial Mathematics (ITWM): working on the fields of machine learning and neural computing.
Head of Research Team “Network Dynamics for Natural Computation” (2018-2023)
Chair for Network Dynamics Center for Advancing Electronics Dresden (cfaed) and Institute of Theoretical Physics, Technische Universität Dresden: study of alternative dynamical systems’ approaches to computation, exploiting mathematical concepts of nonlinear and networked analogue systems; particularly developing artificial spiking-neural-network controllers for autonomous navigation, exploiting frameworks as Heteroclinic Computing and reconfigurable k-winners-take-all computing via purely inhibitory neural networks.
Parental leave — (2016-2018)
Postdoctoral researcher (2014-2016)
Theory of Neural Dynamics Group, Max Planck Institute for Brain Research (MPIBR) – Theory of Neural Dynamics Group: study of new forms of synaptic plasticity and facilitation production of a synaptic protein (PRG1) via changes in post- and pre-synaptic calcium levels leading to hyper-synchronization (related to epileptic events).
Postdoctoral researcher (2010-2014)
Network Dynamics Group, Max Planck Institute for Dynamics and Self-Organization (MPIDS): study in the area of natural computing by switching in networks of states, designing and identifying systems to exhibit any given computational operation.
Doctoral researcher (2007-2010)
Network Dynamics Group Max Planck Institute for Dynamics and Self-Organization (MPIDS): study of new forms of natural computation in bio-inspired systems, in particular exploiting of perturbed heteroclinic cycles to perform universal computation and gaining new insights on how complex computations may be performed in artificial and biological systems (grades: “summa cum laude” on thesis, defense and total distinction).
Federal University of Rio Grande do Sul — Institute of Physics (2003-2006)
Preparation of a Master Thesis. Advisor: Prof. Dr. Rubem Erichsen Jr. In this study, we derived analytical equations of recurrence which describe the exact macroscopic dynamics of a layered network of non-monotonic units through a signal-to-noise analysis.
Federal University of Rio Grande do Sul — Institute of Physics (2002-2003)
Development of a program, in C language, to simulate a network of spiking Fitzhugh-Nagumo neurons and the analysis of its memory storage capacity on different parameter sets
Federal University of Rio Grande do Sul — Computational Sciences Institute (2001-2002)
Development of a JAVA program to implement a “Minority Game” simulation and the analysis of the performance of different competing agents’ personalities and different memory lengths
skills
Science
Physics, dynamical systems,
bio-inspired computers and
Ai in general
Data Science
Azure cloud, data management,
ML solutions (DP-200, DP-100)
Organizational
ITIL4 framework,
Agile - Scrum,
Scientific writing
Coding
Fundamentals of: c++, c#, python, R
and Unity
Languages
Portuguese (native), English (fluent),
German (Basic)
Experience
Multi-cultural/language teams,
leading and advising team members
teaching and conferences
Teaching experience
Co-advisor: Bachelor thesis, J. Lindermeir, “Quadcopter Motion Control by Spiking Neural Networks with Proportional Inhibition” (2021)
Co-advisor: Master thesis, Arash Akrami, “Dynamic Response of Linear Reservoirs” (2020)
Coordinator: seminar series, “Coding strategies in neural networks“, MPIBR (2015)
Teaching Assistant: “Einfuehrung in die Programmierung“ (2009)
Advisor: PhD thesis, G. S. Carmantini, “Dynamical Systems Theory for Transparent Symbolic Computation in Neuronal Networks” (2017)
Co-advisor: Bachelor thesis, M. Voit, “Long- and short-term memory in heteroclinic networks” (2014)
Co-advisor: Master thesis, F. Fix, “Designing Heteroclinic Networks” (2013)
Co-advisor: Diploma thesis, G. Weber, “On networks of unstable states in symmetric systems of coupled oscillators” (2013)
Participant: Scientific Communication Workshop (2009)
Conferences
22-24 March 2021, contributed talk at “DPG-Frühjahrstagung“, virtual meeting
5-6 October 2017, invited talk at “Workshop on Dynamical Systems and Brain-inspired Information Processing” (workshop), Konstanz (Germany)
17-18 December 2015, “Heteroclinic dynamics in neuroscience” (workshop), Nice (France)
19-20 March 2015, invited talk at “SLOFADYBIO: a network for slow-fast dynamics applied to the biosciences” (meeting/conference), Paris (France)
3-7 June 2013, contributive talk at “XXXIII Dynamics Days Europe 2013“, Madrid (Spain)
8-12 July 2013, “Lakeside Research Days 2013” (workshop), Klagenfurt (Austria)
22-26 May 2011, invited talk at “SIAM Conference on Applications of Dynamical systems“, Snow Bird, Utah (USA)
26-30 July 2010, contributive talk at “Dynamics Days South America“, São José dos Campos (Brazil)
Aug. 31 – Sep. 4 2009, poster at “XXIX Dynamics Days Europe“, Göttingen (Germany)
22-27 March 2009, contributive talk at “Verhandlungen der Deutsche Physikalische Gesellschaft“, Dresden (Germany)
projects
Emergent Dynamical Intelligence (EDI)
We aim to develop code and code-free tools to generate Ai solutions for autonomous agents based on the dynamics between sensors and internal machine states
Analysis of Cycles in Asset-Network (ACAN)
We aim to stablish a set of technical procedures and models to facilitate and optimize the analysis of complex networks comprised of (financial) assets.
Heteroclinic Computing Systems
The aim of this project is to understand the requirements for a heteroclinic computer implementation.
.
Learning via controlled transitions into chaos
In this project we explore the potential of controllable phase transitions in layered neural networks to learn new patterns (solutions) during run time.