Oracle OCI AI Services

Posted on Updated on

Oracle Cloud have been introducing new AI Services over the past few months, and we see a few more appearing over the coming few months. When you look at the list you might be a little surprised that these are newly available cloud services from Oracle. You might be surprised for two main reasons. Firstly, AWS and Google have similar cloud services available for some time (many years) now, and secondly, Oracle started talking about having these cloud services many years ago. It has taken some time for these to become publicly available. Although some of these have been included in other applications and offerings from Oracle, so probably they were busy with those before making them available as stand alone services.

These can be located in your Oracle Cloud account from the hamburger menu, as shown below

As you can see most of these AI Services are listed, except for the OCI Forecasting, which is due to be added “soon”. We can also expect to have an OCI Translation services and possibly some additional ones.

  • OCI Language: This services can work with over 75 languages and allows you to detect and perform knowledge extraction from the text to include entity identification and labelling, classification of text into more than 600 categories, sentiment analysis and key phrase extraction. This can be used automate knolwedge extraction from customer feedback, product reviews, discussion forums, legal documents, etc
  • OCI Speech: Performs Speech to Text, from live streaming of speech, audio and video recordings, etc creating a transcription. It works across English, Spanish and Portuguese, with other languages to be added. A nice little feature includes Profanity filtering, allowing you to tag, remove or mask certain words
  • OCI Vision: This has two parts. The first is for processing documents, and is slightly different to OCI Language Service, in that this service looks at processing text documents in jpeg, pdf, png and tiff formats. Text information extraction is performed identifying keep terms, tables, meta-data extraction, table extraction etc. The second part of OCI Vision deals with image analysis and extracting key information from the image such as objects, people, text, image classification, scene detection, etc. You can use either the pretrained models or include your own models.
  • OCI Anomaly Detection: Although anomaly detection is available via algorithms in the Database and OCI Data Science offerings, this new services allow for someone with little programming experience to utilise an ensemble of models, including the MSET algorithm, to provide greater accuracy with identifying unusual patterns in the data.

Note: I’ve excluded some services from the above list as these have been available for some time now or have limited AI features included in them. These include OCI Data Labelling, OCI Digital Assistant.

Some of these AI Services, based on the initial release, have limited functionality and resources, but this will change over time.