Artificial intelligence (AI) investments have grown threefold since 2013, but adoption outside of the tech sector remaining at an early stage. As reported by a recent study from McKinsey, key factors in driving this imbalance are the ability of businesses to overcome technical, expertise, and commercial challenges.
When launching SAP Leonardo Machine Learning Foundation, SAP started on a mission to overcome these challenges and help all customers transition to the intelligent enterprise, no matter their level of digital maturity and AI expertise. Now, SAP expands the capabilities of its machine learning platform, including the training of image classification services with training data belonging to the customer, the opportunity for customers to deploy their own models on SAP Leonardo Machine Learning Foundation, and new ready-to-use services.
Customization 2.0 to Build the Intelligent Enterprise
To help customers and partners address a larger number of use cases, SAP has opened the enterprise-class model training capability of SAP Leonardo Machine Learning Foundation. It is now possible for customer to tailor services to their business needs by training services on their unique data. This new functionality is enabled via a secure extraction of data via SAP Cloud Platform and predefined training routines. On completion, customers can validate the training results and then apply the retrained model in productive scenarios. In practice, this new functionality now empowers a business with little AI expertise to roll out a mobile application that allows customers to submit a picture of any texture and automatically be presented with the closest pattern in the company’s catalog.
In addition, SAP now enables customers to upload their own machine learning models to SAP Cloud Platform. This way, businesses are able to utilize the enterprise-grade service provisioning capabilities of SAP Machine Learning Foundation also for the models created by their machine learning experts and partners. For example, it enables a global system integrator with deep investments in data science to publish sophisticated pipeline monitoring model in the cloud, so that utilities can leverage it to trigger predictive maintenance.
Democratizing Machine Learning
With SAP Leonardo Machine Learning Foundation, application developers with no AI expertise have been able to use ready-to-use services to jumpstart the development of intelligent applications. These services are pre-trained to work on real data, readily available on SAP Cloud Platform, and accessible via standard (RESTfull) APIs for easy consumption.
But with the breadth of applicability of machine learning, even when it comes to generalized problems often no one size fits all. Let’s take image recognition as an example. SAP pre-trained model for image recognition can classify a variety of different objects. It can tell apart anything from cars to people, to trees. However, if a car manufacturer wants to facilitate visual shopping by accurately matching a given picture to one of its products, it needs to teach the foundation’s Image Recognition service to recognize different car models. In the same way, the wallpaper retailer mentioned before had to teach the image recognition model how to recognize different kinds of surface textures.
In the world of machine learning, this is done by training the model on specific data, and to do so companies can now use the enterprise-class model training capability of SAP Leonardo Machine Learning Foundation.
The process of training a model is intuitive – simple APIs calls – and does not require data science expertise, complicated equipment provisioning, or capital expenditure, allowing any company to benefit from tailored machine learning intelligence.
Helping Companies Capitalize on AI Investments with Custom-build Services
SAP Leonardo Machine Learning Foundation also provides value to customers and partners that have deep data science expertise in-house. These businesses have accumulated precious data and defined sophisticated models to improve business efficiencies and grow competitive advantage.
For these businesses, SAP Leonardo Machine Learning Foundation now offers the possibility to deploy their Google TensorFlow-based models. By doing so, provisioning a model for productive use becomes easy. Via SAP Leonardo Machine Learning Foundation, authorized users can easily manage their models status, monitor models and make them accessible to third-parties for consumption. SAP Machine Learning Foundation is delivered on SAP Cloud platform, which ensures that customers can leverage these capabilities in a managed, secure, and scalable cloud environment.
Expanding Use Case Coverage and Performance
Innovation never stops. To help companies jump-start innovation in new fields SAP is continuing to add new ready-to-use services to SAP Leonardo Machine Learning Foundation. Two new services have just been announced and are now accessible in SAP API Business Hub on SAP Cloud Platform:
- Multi-Dimensional Time Series Forecast: Allows time-series forecasting to take into account multiple dimensions — e.g., to predict the optimal price for a product based on different factors such as past sales, competition pricing, advertisement expenses
- Object Character Recognition: Outputs the text and numbers shown in a given image
Equally important to the success of machine learning implementations is the ability of SAP to leverage new ecosystem innovations. For example, SAP just upgraded the hardware infrastructure on SAP Cloud platform to NVIDIA Volta GPUs to ensure resource-intensive tasks are carried on with state-of-the art speed and reliability.
In essence, SAP’s approach to machine learning allows companies with different expertise and budgets to utilize artificial intelligence in the most appropriate way as they pursue their preferred path toward becoming an intelligent enterprise. And all of this, while they are gaining business value each step of the way.
For more information on SAP Leonardo Machine Learning Foundation, visit sap.com/ML. If you attending SAP TechEd Barcelona, stop by booth SE02 at the show floor area to meet our team of experts, or attend one of the many machine learning sessions.