The struggle is real, as they say, when it comes to getting machine learning into production. That was one of the big messages of 2019 as enterprises completed successful machine learning pilots but found it much more difficult to put their efforts into production let alone scale them across the whole organization.
Even though everyone seems to be working on it, machine learning deployed in production grew at a slower rate between 2018 and 2019, according to Gartner's annual CIO survey. (In 2017, 4% of organizations had deployed AI in production; in 2018, 14% had deployed to production; and in 2019, 19% had deployed to production.)
Gartner VP analyst and fellow Rita Sallam is forecasting that enterprises that may have experimented with open source technologies in their pilot efforts will likely turn to commercial artificial intelligence and machine learning platforms to pull together those open source efforts into their enterprise deployment efforts.
What's more, enterprises are likely to turn to the AI and ML platforms offered by public cloud providers such as Amazon AWS, Google, and Microsoft Azure.
Recognizing the need, at its re:Invent 2019 event, AWS rolled out a number of machine learning and artificial intelligence offerings that are particularly geared to help organizations get up and running with their deployment efforts.
Among them is one that Amazon said is based on its own experience scaling machine learning within its own operations, and it applies the lessons learned from that experience and from other customer experiences. Think of this as Amazon experts holding the hands of enterprise leaders to guide them through the process for the first time as they learn how ML can be implemented for a particular production project at scale.
Called AWS Machine Learning Embark, the program starts with a discovery day workshop that includes the customer business and tech staff paired with AWS machine learning experts to "work backwards from a problem and align on where machine learning can have a meaningful impact," Michelle Lee, VP of the Machine Learning Solutions Lab at AWS wrote in a blog post about the offering.
The Embark program also includes instructor-led, on-site training pulled from Amazon's Machine Learning University with an option to continue education online and take the AWS Certified Machine Learning - Specialty certification exam.
The program calls for Amazon ML Solutions Lab experts to mentor participants through the entire process of getting their machine learning use case from the initial idea into production.
"Through the process, the team will gain insight into best practices, ways to avoid costly mistakes, and knowledge based on the overall experience of working with the experts who have completed hundreds of machine learning implementations," Lee wrote.
Amazon also announced several other AI services during its re:Invent event.
Amazon Kendra uses machine learning for enterprise search, helping organizations access unstructured text data across different formats and data sources with natural languages queries.
Amazon CodeGuru automates code reviews for developers, evaluating code using pre-trained models created from Amazon's own code reviews and the top 10,000 open-source projects on GitHub. The service will identify any issues and add human-readable comments to pull requests. The service will also help customers find the most expensive lines of code in terms of performance.
Amazon Fraud Detector is a managed service that uses machine learning to offer automated fraud detection including online identity and payment fraud in real time.
Amazon Transcribe Medical uses machine learning to transcribe medical speech, allowing physicians to dictate notes rather than manually enter them into an electronic health record.
Amazon Augmented Artificial Intelligence (A2I) allows developers to validate machine learning predictions with human reviewers. The service provides pre-built human review workflows for common machine learning tasks such as object detection in images, transcriptions of speech, and common machine learning tasks, according to Amazon. Developers can choose confidence thresholds for their specific applications and send all predictions with a confidence score below the threshold to a human reviewer. Where is Amazon getting these human reviewers? The company said developers can choose to have their reviews performed by Amazon Mechanical Turk's 500,000 global workers, third-party organizations with pre-authorized workers (including Startek, iVision, CapeStart Inc., Cogito, and iMerit) or their own reviewers.
"Companies across various industry segments tell us that they want to leverage Amazon's extensive experience with machine learning to address some of the common challenges they face as enterprises on an on-going basis. These challenges include internal search, helping software developers write better code, identifying fraudulent transactions, and improving the overall quality of all machine learning systems," said Swami Sivasubramanian, VP of Amazon Machine Learning, AWS, in a statement.