Skip to main content

Overview Machine Learning and Deep Learning Frameworks

Abstract

This chapter provides an overview of the different machine learning (ML) and deep learning (DL) frameworks, aiming to show the variety ranging from different open-source initiatives through to standard software vendors and specialized start-ups contributing to the enormous amount of tools to analyze, condense and predict data.

Keywords

  • Machine learning frameworks
  • Deep learning framework

This is a preview of subscription content, access via your institution.

Buying options

Chapter
USD   29.95
Price excludes VAT (USA)
  • DOI: 10.1007/978-3-030-78821-6_12
  • Chapter length: 38 pages
  • Instant PDF download
  • Readable on all devices
  • Own it forever
  • Exclusive offer for individuals only
  • Tax calculation will be finalised during checkout
eBook
USD   69.99
Price excludes VAT (USA)
  • ISBN: 978-3-030-78821-6
  • Instant PDF download
  • Readable on all devices
  • Own it forever
  • Exclusive offer for individuals only
  • Tax calculation will be finalised during checkout
Hardcover Book
USD   89.99
Price excludes VAT (USA)
Fig. 1

(© ifb SE)

Fig. 2

(© ifb SE)

Fig. 3

(© ifb SE)

Notes

  1. 1.

    See Kopic et al. (2019).

  2. 2.

    See Akhgarnush et al. (2019).

  3. 3.

    See Steurer (2021).

  4. 4.

    While it has become more and more common to buy successful companies (start-up or established companies) to lower the development risk (or, to be more precise, to transform the development risk into an integration risk).

  5. 5.

    The initiator of Kafka, see Steurer (2021).

  6. 6.

    Google did not initiate Hadoop but contributed to the concept (BigFiles and later Google File System), which had an impact on Hadoop.

  7. 7.

    See The Apache Software Foundation (2020).

  8. 8.

    See Bettio et al. (2019) and The Linux Foundation, Hyperledger (2020).

  9. 9.

    See The Linux Foundation, LF AI & Data (2020).

  10. 10.

    Customer relationship management.

  11. 11.

    Next best offer, Next best action (see May 2019).

  12. 12.

    Microsoft is located somewhere between the standard software vendors and Big Tech (GAFA or GAFAM) and has a special role in the ML/DL framework area.

  13. 13.

    The Microsoft Cognitive Toolkit.

  14. 14.

    Referring to the continent, or New Zealand, to be precise.

  15. 15.

    E.g. KNIME (Konstanz Information Miner).

  16. 16.

    Caffe was replaced by Caffe2 in 2017 and then integrated into PyTorch by Facebook.

  17. 17.

    Especially the American Big Techs (GAFA), and we have already discussed the localization of Microsoft (standard software vendor or Big Tech) in this chapter.

  18. 18.

    Due to their “deep pockets” filled with revenues from advertising (particularly Google and Facebook) or other differentiating activities like Amazon’s cloud services.

  19. 19.

    Like GNU General Public License (GPL), GNU Lesser General Public License (LGPL), Mozilla Public License, BSD license (Berkeley Software Distribution) or Apache Software License (for details see Sect. 2.3 in Liermann, Open-Source Software 2021).

  20. 20.

    Application-specific integrated circuit.

  21. 21.

    Graphics processing unit.

  22. 22.

    Convolutional neural networks or Deep convolutional networks (DCN), see Sect. 3.6 in Liermann et al., Deep learning—an introduction (2019).

  23. 23.

    Recurrent neural network, see Sect. 3.2 in Liermann et al., Deep learning—an introduction (2019).

  24. 24.

    Baidu is a multinational technology company focused on services and products in an internet-related context and artificial intelligence (AI). Baidu is one of the Chinese Big Tech companies (BATX).

  25. 25.

    Compute Unified Device Architecture by Nvidia is an application programming interface (API) model and a parallel computing platform.

  26. 26.

    Caffe/Caffe2, Chainer, Microsoft Cognitive Toolkit, MATLAB, Mxnet, PaddlePaddle, PyTorch, Tensorflow, Torch and Mathematica’s Wolfram Language.

  27. 27.

    Intel, Nvidia, Microsoft and IBM.

  28. 28.

    Also referred to as artificial general intelligence (AGI) or full AI.

  29. 29.

    See Baum (2017).

  30. 30.

    Also referred to as narrowed AI.

  31. 31.

    It is to be expected that things will have already moved forward by the time the book is published.

  32. 32.

    See Dlib (2020).

  33. 33.

    See H2O.ai, h2o.ai Overview (2020).

  34. 34.

    Formerly known as Data Science Experience or DSX.

  35. 35.

    See also Microsoft Corporation (2020).

  36. 36.

    GBT—Gradient boosted trees.

  37. 37.

    GBDT—Gradient boosted decision trees.

  38. 38.

    GBRT—Gradient boosted regression trees.

  39. 39.

    GBM—Gradient boosting machine.

  40. 40.

    MART—Multiple additive regression trees.

  41. 41.

    Intelligent optimization is automated creation and selection of improving solutions.

  42. 42.

    Python module that provides Python bindings for ML.NET.

  43. 43.

    Armadillo is a library for linear algebra (see Conrad Sanderson and Ryan Curtin 2020).

  44. 44.

    RCASE offers methods aiming to identify the root causes of faults or problems. It has applications—among others—in incomplete data and dirty data cleansing.

  45. 45.

    Density-Based Spatial Clustering of Applications with Noise is a popular data mining algorithm for cluster analysis (see Ester et al. 1996).

  46. 46.

    See XGBoost Contributors (2020).

  47. 47.

    GBM—Gradient boosting machine.

  48. 48.

    GBRT—Gradient boosted regression trees.

  49. 49.

    GBDT—Gradient boosted decision trees.

  50. 50.

    Sometimes referred as The Microsoft Cognitive Toolkit.

  51. 51.

    YOLO is a one-step, real-time object detection system. A one-step detection strategy means that the images to be analyzed need to be read only once.

  52. 52.

    See fast.ai (2020).

  53. 53.

    See flux (2020).

  54. 54.

    Julia is a high-level programming language targeting numerical and scientific computing. Julia can also be used as a general-purpose language with high execution speed. Julia was created with the goal of being as easy on statistics as R is.

  55. 55.

    DeepMind (formerly Google DeepMind) is a company specializing in artificial intelligence (AI) programming. DeepMind was founded in September 2010 and acquired by Google Inc. in 2014.

  56. 56.

    The following libraries are supported: Caffe, TensorFlow, Torch, Darknet, models in ONNX format.

  57. 57.

    See plaidML (2020).

  58. 58.

    Application-specific chips, to support and accelerate machine learning optimized for tensor operations.

  59. 59.

    NERSeeAlsoSeeAlsoNamed Entity Recognition (Named Entity Recognition), Word embeddings and Seq2Seq models (see Sect. 4 in Liermann et al., Deep learning—an introduction 2019).

  60. 60.

    Including to some extent Natural Language Understanding and Natural Language Generation.

  61. 61.

    For inside and outside digitalization (see Liermann and Stegmann, Introduction 2021).

  62. 62.

    clinical Text Analysis and Knowledge Extraction System.

  63. 63.

    Unstructured Information Management Architecture framework. UIMA was initially developed by IBM.

  64. 64.

    Bidirectional Encoder Representations from Transformers.

  65. 65.

    Systematic Role Labeling.

  66. 66.

    SYSTRAN (SYStem TRANslation) is a software system for machine translation founded in the 1960s.

  67. 67.

    Ubisqus is a language service provider.

  68. 68.

    For the NPL frameworks BERT, DistilBert, GPT-2, RoBERTa, XLM and XLNet.

  69. 69.

    Natural Language Understanding.

  70. 70.

    Natural Language Generation.

  71. 71.

    Extraction, transformation, loading (ETLSeeAlsoSeeAlsoExtraction, transformation, loading).

  72. 72.

    Analysis, visualization and reporting.

  73. 73.

    See QuantumBlack (2020).

  74. 74.

    Pip—a Python package-management system—targets installing and managing Python software packages.

  75. 75.

    See ROOT/TMVA team (2020).

  76. 76.

    Driven by the algorithmic demand (for a simple example see subsection 2.1.4 in Liermann et al., Deep learning—an introduction 2019) for training the deep learning networks, which is to some extent similar to the requirements for 3D graphics.

  77. 77.

    The programming language is a superset of the Python programming language and aims to give C-like performance to most code parts in Python with additional C-oriented syntax.

  78. 78.

    R is a free software environment and programming language for statistical computing and graphic representation. The development started in 1993.

  79. 79.

    S is a statistical programming language developed in 1970 at Bell Laboratories.

  80. 80.

    The R package reticulate offers an integration of python code to R see Ushey (2020).

  81. 81.

    The SAS language is a statistical analysis programming language designed by Anthony James Barr at North Carolina State University and is the standard language for the SAS software modules.

  82. 82.

    App programming language of iOS and other Apple operating systems.

  83. 83.

    Kotlin designed for interoperability with Java. Kotlin’s JVM version is based on the Java Class Library.

  84. 84.

    See The Linux Foundation, Open Neural Network Exchange (2020).

  85. 85.

    Allow developers to more easily move between frameworks, some of which may be more desirable for specific phases of the development process, such as fast training, network architecture flexibility or inferencing on mobile devices.

  86. 86.

    Allow hardware vendors and others to improve the performance of artificial neural networks of multiple frameworks at once by targeting the ONNX representation.

Literature

  • Akhgarnush, Eljar, Lars Broeckers, and Thorsten Jakoby. 2019. “Hadoop—A Standard Framework for Computer Clusters.” In The Impact of Digital Transformation and Fintech on the Finance Professional, edited by Volker Liermann and Claus Stegmann. New York: Palgrave Macmillan.

    Google Scholar 

  • Baum, Seth. 2017. A Survey of Artificial General Intelligence Projects for Ethics, Risk, and Policy. Working Paper 17-1, Global Catastrophic Risk Institute.

    Google Scholar 

  • Bettio, Martina, Fabian Bruse, Achim Franke, Thorsten Jakoby, and Daniel Schärf. 2019. “Hyperledger Fabric as a Blockchain Framework in the Financial Industry.” In The Impact of Digital Transformation and Fintech on the Finance Professional, edited by Volker Liermann and Claus Stegmann. New York: Palgrave Macmillan.

    Google Scholar 

  • Conrad Sanderson, and Ryan Curtin. 2020. Armadillo. Accessed December 15, 2020. http://arma.sourceforge.net/.

  • Devlin, Jacob, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. Pre-training of Deep Bidirectional Transformers for Language Understanding. Google AI Language.

    Google Scholar 

  • Dlib. 2020. Dlib. Accessed December 15, 2020. http://dlib.net/.

  • Ester, Martin, Hans-Peter Kriegel, Jörg Sander, and Xiaowei Xu. 1996. “A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise.” Proceedings of 2nd International Conference on Knowledge Discovery and Data Mining (KDD-96).

    Google Scholar 

  • Facebook Open Source. 2017. “Caffe2 Open Source Brings Cross Platform Machine Learning Tools to Developers.” Caffe2. April 18. Accessed December 15, 2020. https://caffe2.ai/blog/2017/04/18/caffe2-open-source-announcement.html.

  • fast.ai. 2020. fastai. Accessed December 15, 2020. https://docs.fast.ai/.

  • flux. 2020. flux. Accessed December 15, 2020. https://fluxml.ai/.

  • Google Inc. 2020. Google APIs Explorer. Accessed December 15, 2020. https://developers.google.com/apis-explorer.

  • H2O.ai. 2020. h2o.ai Overview. December 15. Accessed December 15, 2020. http://docs.h2o.ai/h2o/latest-stable/h2o-docs/index.html.

  • Kopic, Eva, Bezu Teschome, Thomas Schneider, Ralph Steurer, and Sascha Florin. 2019. “In-Memory Databases and Their Impact on Our (Future) Organizations.” In The Impact of Digital Transformation and Fintech on the Finance Professiona, edited by Volker Liermann and Claus Stegmann. New York: Palgrave Macmillan.

    Google Scholar 

  • Liermann, Volker. 2021. “Open-Source Software.” In The Digital Journey of Banking and Insurance, Volume II—Digitalization and Machine Learning, edited by Volker Liermann and Claus Stegmann. New York: Palgrave Macmillan.

    Google Scholar 

  • Liermann, Volker, and Claus Stegmann. 2021. “Introduction.” By Volker Liermann and Claus Stegmann, The Digital Journey of Banking and Insurance, Volume I–Disruption and DNA. New York: Palgrave Macmillan.

    Google Scholar 

  • Liermann, Volker, Sangmeng Li, and Norbert Schaudinnus. 2019. “Deep Learning—An Introduction.” In The Impact of Digital Transformation and Fintech on the Finance Professional, edited by Volker Liermann and Claus Stegmann. New York: Palgrave Macmillan.

    Google Scholar 

  • May, Uwe. 2019. “The Concept of the Next best Action/Offer in the age of Customer Experience.” In The Impact of Digital Transformation and Fintech on the Finance Professional, edited by Volker Liermann and Claus Stegmann. New York: Palgrave Macmillan.

    Google Scholar 

  • Microsoft Corporation. 2020. LightGBM. Accessed December 15, 2020. https://lightgbm.readthedocs.io/.

  • plaidML. 2020. plaidML. Accessed December 15, 2020. https://github.com/plaidml/plaidml.

  • QuantumBlack. 2020. Kedro’s Documentation. Accessed December 15, 2020. https://kedro.readthedocs.io/en/stable/.

  • ROOT/TMVA team. 2020. Toolkit for Multivariate Data Analysis with ROOT. Accessed December 15, 2020. http://tmva.sourceforge.net/.

  • SageMath. 2020. Sage. Accessed December 15, 2020. https://www.sagemath.org/.

  • Steurer, Ralph. 2021. “Kafka—Real Time Streaming for the Finance Industry.” In The Digital Journey of Banking and Insurance, Volume III—Data Storage, Processing, and Analysis, edited by Volker Liermann and Claus Stegmann. New York: Palgrave Macmillan.

    Google Scholar 

  • The Apache Software Foundation. 2020. The Apache Software Foundation. Accessed December 15, 2020. https://www.apache.org/.

  • The Linux Foundation. 2020. LF AI & Data. Accessed December 15, 2020. https://lfaidata.foundation/.

  • ———. 2020. Hyperledger. Accessed December 15, 2020. https://www.hyperledger.org/.

  • ———. 2020. Open Neural Network Exchange. Accessed December 15, 2020. https://onnx.ai/.

  • Ushey, Kevin. 2020. The Comprehensive R Archive Network. Accessed December 15, 2020. https://cran.r-project.org/web/packages/reticulate/index.html.

  • Witten, Ian H., Eibe Frank, and Mark A. Hall. 2011. Data Mining: Practical Machine Learning Tools and Techniques. San Francisco, CA: Morgan Kaufmann.

    Google Scholar 

  • XGBoost Contributors. 2020. XGBoost. Accessed December 15, 2020. https://xgboost.ai/.

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Volker Liermann .

Editor information

Editors and Affiliations

Rights and permissions

Reprints and Permissions

Copyright information

© 2021 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this chapter

Verify currency and authenticity via CrossMark

Cite this chapter

Liermann, V. (2021). Overview Machine Learning and Deep Learning Frameworks. In: Liermann, V., Stegmann, C. (eds) The Digital Journey of Banking and Insurance, Volume III. Palgrave Macmillan, Cham. https://doi.org/10.1007/978-3-030-78821-6_12

Download citation

  • DOI: https://doi.org/10.1007/978-3-030-78821-6_12

  • Published:

  • Publisher Name: Palgrave Macmillan, Cham

  • Print ISBN: 978-3-030-78820-9

  • Online ISBN: 978-3-030-78821-6

  • eBook Packages: Economics and FinanceEconomics and Finance (R0)