Select Page

Programmer

 

Advances in technology are causing a shift from traditional software programming to machine learning. Traditional programming entails a human developer inputting code in order for a computer to recognize the difference between two objects. Machine learning changes that task from telling the computer what to think, to teaching the system how to think. With these advances in artificial intelligence making such strides many tech companies are making the transition.

In a recent survey by O’Reilly Media, the top tools being utilized by today’s tech firms were compiled. As many companies are still examining the usefulness of this technology, most are in the evaluation phase of these products. The survey found that the most commonly utilized tools were for model visualization or automated model search and hyperparameter tuning. Therefore, it was not surprising to find that most of the tools noted are in the form of supervised learning, whereby large amounts of data are inputted in order to teach a system how to differentiate between items. For instance, in order to teach the system to tell a STOP sign apart from a YIELD sign, hundreds of images of each type of sign would be uploaded into the system.

In February 2019, an article was released with the results itemizing the top contending tools and their features. The following is a list of those tools and the percentage in which they are being utilized by businesses, out of the over 1300 people who participated in the survey. The survey consisted mainly of persons who work in the healthcare, technology, and finance industries.

 

  1. TensorFlow – 55%
  2. scikit-learn – 48%
  3. Keras – 34%
  4. PyTorch – 29%
  5. Azure ML Studio – 17%
  6. Google Cloud ML Engine – 16%
  7. Spark NLP – 16%
  8. Amazon SageMaker – 12 %
  9. H20 – 8%
  10. spaCy/Prodigy – 8%
  11. OpenAI Gym – 5%
  12. BigDL and Analytics Zoo – 8%
  13. AylienNLP – 1%
  14. RISE Lab Ray – 1%

 

Another forty-three percent were noted as utilizing other open source tools, as well as another 22% which utilized other cloud-based services.