IBM Paves Way for Closing Skills Gap in New AI Tool Release

“Deep Learning-as-a-Service” (DLaaS) can now be added to the “as-a-service” trend thanks to IBM’s recent announcement that it launched a new AI tool within Watson Studio that “allows data scientists to visually design their neural networks and scale out their training runs while auto-allocation means paying only for the resources utilized.” According to the Wall Street Journal, the goal of this new component of the Watson Studio is to “bridge a skills gap for creating custom artificial-intelligence systems to draw value from data.”

IBM’s DLaaS strives to allow AI and deep learning platforms, algorithms and models to become more accessible to a wider audience. Such an objective holds particularly true as these fields are still young and it stands very difficult to replicate and create such systems from the ground-up due to large computing power required and complex algorithm construction.

The new AI tool’s target audience is for developers, data scientists, business analysts and those in similar roles to upload data to the platform and from there create and implement deep learning algorithms such as neural network models to garner new levels of insights via a “plug-and-play” interface. Such an interface enables individuals and organizations to leverage the power of deep learning and artificial intelligence models without the heavy lifting involved in coding and developing the core framework. IBM’s DLaaS touts features such as an experiment assistant, hyperparameter optimization and a beta version of its Neural Network Modeler.

[Related Article: Understanding AI Toolkits posted by Edd Wilder-James]

This latest addition to Watson Studio illustrates the exponential growth that the AI and deep learning spaces face. According to Statistica, revenues from AI totaled at $2.42 billion in 2017 and are predicted to scale to $10.53 billion in 2020 and $59.49 billion in 2025. IBM’s tool release symbolizes only the beginning of the DLaaS segment.

Read more articles like this at OpenDataScience.com