There is a growing awareness of the strategic value of AI in business. However, most organizations still grapple with fundamental information architecture challenges. The chores of finding, collecting and organizing fragmented and siloed data, and then preparing that data for analysis and ML comprises is often slowing AI development.
In a recent Forrester report, 60 percent of respondents said managing data quality was among the top challenges of AI. Another 44 percent attributed it to data prep. For organizations with no data scientists, AI projects are even more challenging.
Watson Studio’s new AutoAI capabilities work in conjunction with Watson Machine Learning to begin to remedy these challenges by automating and speeding a variety of the steps in the AI lifecycle.
The new AutoAI capabilities are available now in Watson Studio on the IBM Cloud. They automate the time-consuming processes of data prep and preprocessing, including model development and feature engineering. They furthermore aim to enable users to leverage hyperparameter optimization capabilities to build data science and AI models with greater ease.
In addition, AutoAI contains a suite of the most powerful model types for enterprise data science; and aims to let users quickly scale ML experimentations and deployment processes.
AutoAI for data prep
“IBM has been working closely with clients as they chart their paths to AI, and one of the first challenges many face is data prep – a foundational step in AI,” said Rob Thomas, General Manager, IBM Data and AI. “We have seen that complexity of data infrastructures can be daunting to the most sophisticated companies. However, it can be overwhelming for those with little to no technical resources. The automation capabilities we’re putting on Watson Studio smooth the process and help clients start building ML models and experiments faster.”
Also included in the AutoAI family is IBM Neural Networks Synthesis (NeuNetS). It was first previewed last fall and is currently in open beta within Watson Studio projects.
The technology aims to fast-track the development of deep-learning models by using AI to automatically synthesize customized neural networks. NeuNetS enables users to choose whether to optimize for speed or accuracy. Users can furthermore watch the model build and train itself in real-time.
Watson Studio AutoAI leverages key technologies developed in IBM Research. It furthermore builds on automation capabilities IBM has been developing and offering across its portfolio for years.