kurg.org is the eponymous online home of Marcus Edel—developer, maintainer, computer science student, and open source enthusiast.


BIOGRAPHY

Helping people answer questions using data they can't see—AI, stats, data science, computation.

I was born and raised in the heart of Germany and grew up a very outgoing and active child. Always staying proactive. I studied Computer Science and Physics at the Free University of Berlin.

In my research, one important thing I do is develop the C++ machine learning library mlpack, which contains a growing collection of high-quality implementations of cutting-edge machine learning algorithms and is in use by both researchers and industry groups worldwide. I also work on ensmallen—a flexible library for efficient mathematical optimization, which provides a simple set of abstractions for writing an objective function to optimize. It also provides a large set of standard and cutting-edge optimizers that can be used for virtually any mathematical optimization task.

In short, my research interests include (modular/hierarchical) reinforcement learning, (stochastic/black-box) optimization and (deep/recurrent) neural networks. I have used these techniques in application domains ranging from computer vision and neural decoding to localization and robotics.


Publications

Here is a list of publications, listed in reverse order of publication. I also have a Google Scholar profile although the information there may not always be up-to-date.
Github
Résumé
  • M. Edel, E. Köppe."Deep Neural Networks for Multimodal Wearable Activity Recognition", in preparation.
  • S. Bhardwaj, R. R. Curtin, M. Edel, Y. Mentekidis, C. Sanderson. "ensmallen: a flexible C++ library for efficient function optimization", in NeurIPS 2018 Workshop on Systems for ML, 2018. [bib] [pdf] [code]
  • M. Edel. "mlpack open-source machine learning library and community", in NeurIPS 2018 Workshop on Machine Learning Open Source Software, 2018. [bib] [pdf] [code]
  • R.R. Curtin, M. Edel, M. Lozhnikov, Y. Mentekidis, S. Ghaisas, S. Zhang. "mlpack: a fast, flexible machine learning library", The Journal of Open Source Software, vol. 3, issue 26, pp. [bib] [pdf] [code]
  • R. R. Curtin, S. Bhardwaj, M. Edel, Y. Mentekidis. "A generic and fast C++ optimization framework", arXiv preprint arXiv:1711.06581, 2017. [bib] [pdf]
  • R.R. Curtin, M. Edel. "Designing and building the mlpack open-source machine learning library", submitted to The Fourth International Conference of PUST (ICOPUST), 2017. [pdf] [code]
  • M. Edel, J. Lausch. "Capacity Visual Attention Networks", in Second Global Conference on Artificial Intelligence (GCAI), Proceeedings (Christoph Benzmüller Raul Rojas, Geoff Sutcliffe, eds.), EasyChair, EPiC Series in Computing, 2016. [bib] [pdf]
  • M. Edel, E. Köppe. "Binarized-BLSTM-RNN based Human Activity Recognition", in Proceedings of the 2016 IEEE International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2016. [bib] [pdf]
  • M. Edel, E. Köppe. "An Advanced Method for Pedestrian Dead Reckoning using BLSTM-RNNs", in Proceedings of the 2015 IEEE International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2015. [bib] [pdf]
  • M. Edel, A. Soni, R.R. Curtin. "An automatic benchmarking system", in NIPS 2014 Workshop on Software Engineering for Machine Learning, 2014. [bib] [pdf] [code]
  • Demos

    In order to make my research as reproducible as possible, most of the algorithms and benchmarks I use are available open-source through the mlpack machine learning library. Here is a list of demos, that entirely run in your browser. No software requirements, no compilers, no installations, no GPUs.
  • Deep Learning Optimizer Visualization [demo] [code]
  • Classify MNIST digits with a Convolutional Neural Network [demo] [code]
  • Classify MNIST digits with the Recurrent Model of Visual Attention [demo] [code]
  • An automatic benchmarking system [interface] [code]