Oracle bolsters cloud infrastructure with suite of AI tools
Oracle has updated its cloud infrastructure service (OCI) to include a new suite of artificial intelligence services with six new tools that aim to make it easier and faster for developers and data scientists to apply AI, including machine learning techniques, to different enterprise scenarios.
The new suite of AI services on OCI, Oracle’s public cloud platform used to develop and run big data applications, is available now. It will compete with Amazon Web Services’ (AWS) SageMaker platform and Microsoft’s Azure Machine Learning Studio, which is designed for use by enterprise developers who may not have in-depth expertise or knowledge of data science.
Despite playing catchup with the new OCI services, Oracle’s strategy seems logical, according to Constellation Research VP and Principal Analyst Holger Mueller. “Enterprises that are already on Oracle databases stand to benefit from the new services. It also means that Oracle has managed to keep Oracle’s database load in house on Oracle and has shown that OCI runs the Oracle database best,” Mueller said.
The new services enable developers to avoid worrying about installing, updating and managing AI platforms, allowing them to focus on programming, Mueller noted.
AI to accelerate deployment
The new set of AI services are expected to reduce the time taken to manage data and the time taken to build and deploy applications, said Greg Pavlik, chief technology officer at Oracle Cloud. The time taken to respond to varied scenarios by enterprises could be the difference for their survival in “volatile and uncertain times,” Pavlik said.
The new tools include AI-based language, speech, vision, anomaly detection, data labelling and forecasting services. The new OCI Language service is designed to allow developers to perform text analysis at large scale. The service can understand unstructured text in documents, customer feedback interactions, support tickets and social media, Pavlik said. The service also comes with pretrained models that allow developers to deploy them out-of-the-box and gain insights in the form of sentiment analytics, phrase detection and entity recognition, among other capabilities.
Oracle’s competitors offer similar capabilities. AWS has intelligent language services such as Comprehend, Lex, and Polly, while Microsoft offers Text Analytics API for advanced analysis.
The Speech service, according to Oracle, comes with prebuilt models that can understand speech in several languages in real time. Pavlik said that developers can apply the service to convert file-based audio containing human speech into text transcriptions. The service can be used to provide in-workflow closed captions, index content, and enhance analytics on audio and video content.
AWS’ Transcribe and Translate service could be considered as an equivalent to this service. Azure, too, offers a similar service.
The OCI Vision service is aimed at making it easy for developers to train visual models. It comes with pretrained models for image recognition and document analysis tasks, Pavlik said.
“It also enables users to extend the models to other industry and customer-specific use cases such as scene monitoring, defect detection, and document processing with their own data,” Pavlik said, adding that the service can be used to detect visual anomalies in manufacturing, extract text from forms to automate business workflows, and tag items in images to count products or shipments.
AWS’ Rekognition service and the Azure Computer Vision offers similar capabilities.
Weeding out anomalies and cleaning data
Enterprises spend a lot of time detecting problems with their data and AI models. In order to curb the time needed to detect such anomalies, the new AI services suite includes the OCI Anomaly Detection service, which can flag critical irregularities early, leading to faster resolution times.
“OCI Anomaly Detection, built on the MSET2 algorithm, provides REST APIs and SDKs for several programming languages, which developers can use to easily integrate anomaly detection models into business applications,” Pavlik said. He added that the tool can be used to detect fraud, predict machinery breakdown and record data from multiple sources.
Anomaly detection can be considered as a key aspect of AI services and all vendors should offer it, said Constellation’s Mueller. “It is even more relevant for Oracle, given the massive amounts of transactional data stored in its databases. And being able to detect an anomaly — then flag it in analytics — or even resolve it with actions through AI is huge for enterprises to move faster and become more agile,” Mueller said.
As part of the new suite, Oracle also released the OCI Forecasting service, which automatically provides time-series forecasts based on prebuilt machine learning models without the need to code, Pavlik said. It allows developers to create forecasts for critical business metrics such as product demand and revenue.
Oracle also announced OCI Data Labelling, which helps users create labelled datasets to easily train AI models. According to Pavlik, the new service will allow developers to assemble data, create and browse datasets and apply labels to them through user interfaces or public APIs.
“The labelled data sets can be exported and used for model development across many of Oracle’s AI and data science services, including OCI Vision and OCI Data Science, for a consistent model-building experience,” Pavlik said.