Frequently asked questions
Can’t find the answer you’re looking for? Reach out to our support team.
- What is Natural Language Processing?
- Natural Language Processing (NLP) is a branch of Artificial Intelligence that focusses on teaching machines to understand human-written text. NLP exploits numerous aspects of texts such as sentence structure, length and words used and context.
- How much data is needed to train an NLP model?
- The more the merrier, but typically you would need somewhere between 200 to 400 documents to train an initial NLP model.
- How long does it take to set up a project in Auto-Pilot?
- Setting up a project takes 1 day. First analysis and training of data takes about a week. After that many improvements are gradually put in place.
- Who is responsible for training the NLP models?
- In the beginning it would usually be Auto-Pilot or one of its partners. Once the solution is maturing the client can take control of most aspects.
- Who owns the data on Auto-Pilot?
- The user is and always will be the owner of its own data. The data is not used for any other purpose than to aid the client. Data is separated between clients and securely protected from the outside world.
- What kind of accuracy levels can I expect from the NLP models?
- The accuracy levels keep rising when more data is being processed on the platform. It is not uncommon that the accuracy level will surpass human levels within a couple of months.
- How is the predicted data exported?
- Auto-Pilot has a flexible export component that can be completely customized. This means that data can be exported to all endpoints which allow a connection.
- How is Auto-Pilot hosted?
- The entire architecture behind the platform is cloud-native and runs mostly on AWS. Other implementations such as private cloud or on-premise are negotiable.