Scaia is all about meaning. Anyone can tell you what your documents say. What your documents mean is another thing altogether. AI is what breaks through the density of unstructured data to understand meaning and concepts. By training AI models to actively learn about new subject matter and developing a simple user interface to these models, we open doors to knowledge and find answers to critical problems.
Deep Learning is the future of text understanding. Legacy techniques based on older approaches will never approach the performance of today’s AI-driven models. Scaia uses many of the latest discoveries published as open source code by scientists from Google, Facebook, and other top technology companies focusing on legal, scientific, and technical documents. Our AI scientists and programmers have been able to use many of these emerging ideas and expand on them to develop our own solutions.
Models are only as good as the data they are given. Better data means better models which means better results, but finding that data can be challenging. Using data based on Wikipedia or classic literature is not going to work well on legal and technical documents, any more than a student in Shakespearean literature would be well suited to comprehending legal contracts.
Scaia has been tested and designed to run on immense volumes of data across data centers, multi-cloud environments and local networks. While Scaia’s models are great for the individual researcher, scientist or litigator, the fact that Scaia’s software can scale to large volumes makes it deployable at an enterprise level. Scaia’s SaaS (Software-as-a-Service) offerings allow end users to take advantage of Scaia’s scale with minimal requirements in our their own environments. By Summer 2020 Scaia will be deployable using Docker and Kubernetes running locally, behind corporate firewalls. The name “Scaia” is derived from “Scalable AI”.