My Profile Photo

Sebastian Blaes

Ph.D. candidate in the Autonomous Learning Group at the Max Planck Institute for Intelligent Systems, Tuebingen.

My Profile Photo

Sebastian Blaes

Education

PhD in Computer Science

MPI for Intelligent Systems, Tübingen, GER

2017 - Present

Coursework:

  • Seminar: AI, Science, Society, Responsibility
  • Machine Learning: Algorithms and Theory
  • Probabilistic Inference

MSc in Computer Science

Goethe University, Frankfurt, GER

2013 - 2017

With Distinction

Coursework:

  • Approximation Algorithms
  • Semantics and Analysis of Functional Programming Languages
  • Operating Systems
  • Theoretical Computer Science

MSc in Physics

Goethe University, Frankfurt, GER

2012 - 2015

With Distinction

Coursework:

  • Theoretical Neuroscience
  • Computational Neuroscience
  • Quantum Computing and Information Theory
  • Machine Learning
  • Electronics and Sensors
  • Digital Electronics

BSc in Physics

Goethe University, Frankfurt, GER

2008 - 2012

Coursework:

  • Brain Dynamics: From Neuron to Cortex
  • Plasma Physics
  • Visual Systems: Principles of Attention

Research

Autonomous State Representation Learning
for Efficient Reinforcement Learning and
Intrinsically Motivated Behaviour

PhD Research

2017 – Present

Max-Planck-Institute for Intelligent Systems (Autonomous Learning Group)

Deep Convolutional Networks for Visual Object Recognition:
Few-Shot Learning of New Categories

Graduate Research

2017

Frankfurt Institute for Advanced Studies (Burwick Lab)

Attentional Selection and Suppression Mechanism
in an Oscillatory Neural Network Model

Graduate Research

2015

Frankfurt Institute for Advanced Studies (Burwick Lab)

Plasma Confinement of the Weibel Type

Undergraduate Research

2012

Goethe University (Institute for Applied Physics)

Honors, Awards, Scholarships

  • Scholar of the International Max Planck Research School (IMPRS) for Intelligent Systems (IS)

Experience

Teaching Assistant

  • Reinforcement Learning (University Tuebingen, 2018, 2020/2021)
  • Introduction to Programming I (C++) (Goethe University, 2014 - 2016)
  • Introduction to Programming II (Functional Programming, Databases) (Goethe University, 2014 - 2016)
  • Theoretical Computer Science I (Algorithm Engineering and Analysis) (Goethe University, 2014 - 2016)

Workshop Instructor

MPI-IS Tuebingen

  • Tübingen Robust Learning Symposium (2020)
  • One day Workshop on Machine Learning at Leipzig University, GER (2018)

Lab Instructor

Goethe University

2012 - 2014

  • Lab Experiments: Electricity and Magnetism

IT Assistant

Wachendorf Elektronik GmbH & Co. KG

2004 - 2006

  • PHP Programming
  • IT Support

Training

  • Transylvanian Machine Learning Summer School (TMLSS) (2018)
  • Deep Learning and Reinforcement Learning Summer School by Vector Institute (2018)

Publications

Journals

  • S.Blaes and T.Burwick, Few-Shot Learning in Deep Networks through Global Prototyping, Neural Networks, 94 (2017) 159-172
  • S.Blaes and T.Burwick, Attentional Bias through Oscillatory Coherence between Excitatory Activity and Inhibitory Minima, Neural Computation, 27 (2015) 1405--1437
  • Teske, C., Y. Liu, S. Blaes and J. Jacoby, Electron Density and Plasma Dynamics of a Spherical Theta Pinch, Physics of Plamsas, (1994-present) 19, no. 3 (2012): 033505

Conferences

  • Cristina Pinneri, Shambhuraj Sawant, Sebastian Blaes, Jan Achterhold, Jörg Stückler, Michal Rolinek, Georg Martius, Sample-efficient Cross-Entropy Method for Real-time Planning, Conference on Robot Learning 2020,
  • Blaes, S., Vlastelica, M., Zhu, J., Martius, G., Control What You Can: Intrinsically Motivated Task-Planning Agent, Advances in Neural Information Processing (NeurIPS’19), pages 12520 -- 12531
  • M.Mundt, S. Blaes and T. Burwick, Feature Binding in Deep Convolution Networks with Recurrences, Oscillations and Top-Down Modulated Dynamics, ESANN'2016, Bruges, Belgium, pages 423 -- 428

Preprints

Attended Conferences

2019

  • Control What You Can. Intrinsically Motivated Task-Planning Agent, Advances in Neural Information Processing (NeurIPS’19) (O)
  • Control What You Can. Intrinsically Motivated Task-Planning Agent, The Fourth International Workshop on Intrinsically Motivated Open-ended Learning (P)
  • Control what you can: Intrinsically motivated reinforcement learner with task planning structure, Workshop on Structure & Priors in Reinforcement Learning (ICLR 2019) (O)
  • Control What You Can: Intrinsically Motivated Reinforcement Learner with Task Planning Structure, Workshop on Task-Agnostic Reinforcement Learning (ICLR 2019) (O)

2018

  • Control What You Can: Intrinsically Motivated Hierarchical Reinforcement Learner, Deep Reinforcement Learning Workshop at NeurIPS2018 (O)
  • Embodied Exploration Strategy for Off-Policy Reinforcement Learning, Third Machine Learning in Planning and Control of Robot Motion Workshop (ICRA2018) (O)
  • Using Embodied Exploration for Reinforcement Learning, Second Max Planck ETH Workshop on Learning Control (P)

A: attended, P: presentation, O: poster

© 2020 Sebastian Blaes. All rights reserved.