1. Proceedings
  2. I3M
  3. 2019
  4. EMSS
  5. 10.46354/i3m.2019.emss.016

Software frameworks for artificial intelligence: comparsion of low-level and high-level approaches

  • a Michael Bogner  ,
  • b Florian Weindl  ,
  • c Franz Wiesinger  
  • a,b,c University of Applied Sciences Upper Austria – Department of Embedded Systems Engineering, Softwarepark 11, A-4232 Hagenberg, Austria
Cite as
Bogner M., Weindl F., Wiesinger F.  (2019). Software frameworks for artificial intelligence: comparsion of low-level and high-level approaches. Proceedings of the 31st European Modeling & Simulation Symposium (EMSS 2019), pp. 96-102. DOI: https://doi.org/10.46354/i3m.2019.emss.016.

Abstract

As nearly every artificial intelligence application is based on a framework, using the best fitting one for the task is key in developing an efficient solution quickly. Since there are two main types of frameworks, based on low and high abstraction level approaches, these two types will get compared and evaluated throughout this paper using Tensorflow and Keras as representatives. Key features of artificial intelligence frameworks for industrial applications are performance, expandability, abstraction level and therefore ease of use for rapid prototyping. All those features are major factors to keep development time and costs as low as possible, while maximizing product quality. To evaluate both approaches by these criteria a neural network classifying handwritten digits is implemented.

References

  1. Gershenson C.G., 2003. Artificial Neural Networks for Beginners. Eprint: cs/0308031.
  2. Schmidhuber J, 2015. Deep learning in neural networks: An overview. Available from: http://www.sciencedirect.com/science/article/pii/S0893608014002135 [accessed 10.07.2018].
  3. Lunze J, 2016. Künstliche Intelligenz für Ingenieure: Methoden zur Lösung ingenieurtechnischer Probleme mit Hilfe von Regeln, logischen Formeln und Bayesnetzen. Available from: https://books.google.at/books?id=NqBlCwAAQBAJ [accessed 11.08.2018].
  4. LeCun Y.L., Cortes C.C., Burges C.J.C.B., 1998. THE MNIST DATABASE of handwritten digits. Available from: http://yann.lecun.com/exdb/mnist/ [accessed 13.07.2019].
  5. Abadi M.A., 2016.Tensorflow: A system for large-scale machine learning. Available from: http://arxiv.org/abs/1605.08695 [accessed 06.07.2018].
  6. Keras Home, 2019. Keras Home. Available from: https://keras.io/#keras-the-python-deep-learninglibrary [accessed 02.08.2018].
  7. Keras, 2019. Keras Why use Keras. Available from: https://keras.io/#why-this-name-keras [accessed 02.08.2018].
  8. Google Inc., 2019. Tensorflow Homepage. Available from: https://www.tensorflow.org/ [accessed 11.07.2019].
  9. LeCun Y.L., Bengio Y., 1995. The Handbook of Brain Theory and Neural Networks.
  10. Theano, 2019. Theano Homepage. Available from: http://deeplearning.net/software/theano/ [accessed 11.07.2019].
  11. Berkeley AI Research, 2019. Caffe webpage. Available from: https://caffe.berkeleyvision.org/ [accessed 11.07.2019]
  12. NumPy, 2019. NumPy Homepage. Available from: https://www.numpy.org/ [accessed 11.07.2019]
  13. Gazi Yalcin O., 2018. Image Classification in 10 Minutes with MNIST Dataset. Available from: https://towardsdatascience.com/imageclassification-in-10-minutes-with-mnist-dataset-54c35b77a38d [accessed 10.07.2019].