Deep Learning Solutions For
Long-term Value Investing

Find Out More

Deep Learning and Long-Term Value Investing

Quantenstein is an integrated software platform for automated long-term value investing that builds on the latest developments in Deep Learning technology. For a given investment universe (e.g. regions, industries, market cap categories) and set of constraints (e.g. portfolio size, dividend yield, holding period, transaction costs, ESG criteria), Quantenstein optimizes client-specific financial performance metrics based on large quantities of fundamental accounting data to assemble tailored investment portfolios. More than a thousand candidate configurations are tested simultaneously to arrive at the best solution for your needs.


Stock selection

Quantenstein is trained to identify the best stocks for medium to long term investment horizons.

Portfolio construction

Quantenstein integrates stock selection and portfolio construction into a single, end-to-end architecture.

Advanced Machine Learning

Quantenstein leverages state-of-the-art decision tree methods and extensive expertise in Convolutional and LSTM neural networks.

Custom built portfolios

Portfolios are optimized to your specific requirements: stock universe, holding period, yield, performance characteristics, etc. Contact us!

Recent papers

  1. D. Boscaini, J. Masci, E. Rodolà, M. Bronstein. Learning shape correspondence with anisotropic convolutional neural networks. Advances in Neural Information Processing Systems (NIPS), 2016
  2. M.F. Stollenga, J. Masci, F. Gomez, J. Schmidhuber. Deep networks with internal selective attention through feedback connections. Advances in Neural Information Processing Systems (NIPS), 2014
  3. J. Masci, M. Bronstein, A. Bronstein, J. Schmidhuber. Multimodal similarity-preserving hashing. IEEE transactions on pattern analysis and machine intelligence 36 (4), 824-830, 2014
  4. R.K. Srivastava, J. Masci, S. Kazerounian, F. Gomez, J. Schmidhuber. Compete to compute Advances in Neural Information Processing Systems (NIPS), 2013
  5. J. Bayer and C. Osendorfer. Learning stochastic recurrent networks arXiv preprint arXiv:1411.7610, 2014.
  6. J. Bayer, C. Osendorfer, P. van der Smagt. Learning sequence neighbourhood metrics Artificial Neural Networks and Machine Learning–ICANN 2012. Springer, 2012, pp. 531–538.
  7. J. Bayer, C. Osendorfer, D. Korhammer, N. Chen, S. Urban, P. van der Smagt. On fast dropout and its applicability to recurrent networks arXiv preprint arXiv:1311.0701, 2013.
  8. F. Sehnke, C. Osendorfer, T. Rückstieß, A. Graves, J. Peters, J. Schmidhuber. Parameter-exploring policy gradients Neural Networks 23 (4), 551-559, 2010
  9. J. Koutnik, F. Gomez, J. Schmidhuber Evolving neural networks in compressed weight space Proceedings of the Genetic and Evolutionary Computation Conference (GECCO), 2010.
  10. S. Hochreiter, J. Schmidhuber Long Short-Term Memory Neural Computation 9(8): 1735-1780, 1997.
  11. J. Koutnik, K. Greff, F Gomez, J. Schmidhuber A Clockwork RNN International Conference on Machine Learning, 2014.
  12. F. Trifterer Probabilistic Prediction of Returns in High-Frequency Financial Time Series using an Adaptive Bayesian Mixture of Recurrent Neural Networks Ba.S. Thesis


click the pictures for bio

Jonathan Masci
Quantenstein GM
Kevin Endler
Quantenstein GM
Portfolio Manager
Acatis Investment GmbH
Hendrik Leber
Managing Partner
Acatis Investment GmbH
Christian Osendorfer
Research Scientist
Florian Trifterer
Research Scientist


Quantenstein GmbH
c/o ACATIS Investment GmbH
Taunusanlage 18
60325 Frankfurt am Main

Telephone : +49 - 69 - 97 58 37 79
Fax : +49 - 69 - 97 58 37 99

Legal notice:
Managing directors are: Jonathan Masci, Kevin Endler.
Value-added tax identification number according to §27a of the German Value-Added Tax Law (UStG): DE 306 566 800.
Registration Court: Local Court of Frankfurt am Main
Commercial Register Number: HRB 105063

Quantenstein GmbH is responsible for the content
Copyright © Quantenstein 2016

Landing image Copyright: sborisov / 123RF Stock Photo