Datameer has made a major announcement by revealing their newest solution for democratizing enterprise data access. This morning they unveiled SmartAI, a tool that operationalizes deep learning models within Data Lake environments. According to the company, the release will allow enterprises to democratize data science by enabling business users to do the work of data scientists for production-ready scenarios in ways that meet the traditional standards of IT. In this way, analysts can execute deep learning models against enterprise data from the Data Lake to drive analytics.
SmartAI is made up of two parts: Datameer’s Hadoop-native Data Management and data preparation capabilities, and Google TensorFlow‘s deep learning analytic cycle. TensorFlow is an open source software library for machine learning that was originally developed by Google for training neural networks in pattern recognition.
The marriage between Datameer’s existing portfolio of Big Data capabilities and TensorFlow’s deep learning toolset means that users can create their own analytic pipelines that feed deep learning models. In conjunction, analysts gain access to deep learning models inside the Datameer function library for application of the model during analysis. The connection between the model and the Data Lake allows users for runtime execution of TensorFlow models with Datameer jobs, which integrates deep learning patterns.
In a statement to Solutions Review, Datameer’s Vice President of Product Management told us: “Convolutional neural networks are emerging as the deep learning technology of choice, and Google’s TensorFlow is seeing rapid adoption. We have seen significant interest in being able to deploy these models against their enterprise data in a scalable and governed way. With SmartAI, we are seeking to address the convergence of data engineering and deep learning by integrating these models directly into enterprise data pipelines.”
Datameer SmartAI will initially be made available as a technical preview.