IBM has presented the new AI OpenScale technology platform, which aims to address the main challenges to the adoption of Artificial Intelligence such as the way in which these applications respond to the questions, the lack of skills in this field and the complexities linked to the use of AI tools produced by different companies.
IBM’s new technology platform will enable companies to manage AI transparently, regardless of what the application development and application environments are as soon as the applications of Artificial Intelligence go into execution, AI OpenScale is able to detect and correct any errors present.
Next to AI OpenScale, the company has also introduced NeuNetS that allows the creation of complex neural networks even starting from scratch. In this way, it will be possible to fill in the gaps in AI’s competencies and accelerate their growth.
Davide Kenny, Senior Vice President of IBM Cognitive Solutions, said: ” To accelerate the adoption of AI, companies need a single environment in which to run their applications of artificial intelligence. We are at the forefront of promoting the use of the AI that our customers prefer, regardless of the manufacturer, and the use of IBM’s AI OpenScale makes the AI transparent and easy to manage. This market will grow only if companies trust the AI and can scale without limitation IT platform used “.
These are the features of AI OpenScale, as indicated by IBM:
- Understanding how AI applications answer questions : AI OpenScale explains how AI recommendations are formulated with the terms used daily in the company.
- Automatically address biases in AI applications : AI OpenScale continuously monitors AI applications and prevents bias through unique, powerful, and automated ” anti-injury ” technology.
- Allows AI applications to be verifiable – AI OpenScale records every forecast, every version of the model and all the training data used, along with all the metrics to help companies comply with regulations such as GDPR.
- Use AI to build AI – AI OpenScale addresses the shortage of AI-trained engineers using AI. The Neural Network Synthesis Engine (NeuNetS) from IBM will allow companies to build neural networks that are essential to run AI applications quickly and automatically, even from scratch. NeuNetS will initially be available in AI OpenScale in beta.
- Manage and scale AI in an interoperable way – AI OpenScale can work with any model of machine learning or deep learning open source, such as Tensorflow, Scikitlearn, Keras and SparkML. It can also manage applications and models trained and hosted in any common environment, including IBM Watson, IBM PowerAI, Seldon, AWS SageMaker, AzureML, and other non-IBM engines.
AI OpenScale is an answer that IBM tries to give to all those realities that still watch, even with interest, with a certain mistrust towards artificial intelligence. The same IBM has recently announced the results of a survey that has interviewed 5 thousand managers of the company, with 82% who said to evaluate the use of AI and 63% who admitted not to own internally the ability to safely manage this technology. 60% finally expresses perplexity regarding trust and compliance.
AI OpenScale will be available via IBM Cloud and IBM Cloud Private within the year.