Free text parsing in OpenLabyrinth QUestions

Today we are pleased to announce that OpenLabyrinth has a powerful new feature. Case authors can now ask users to type in free-text answers, rather than always relying on clicking on a predefined list of choices.

For an example of all the new features provided by OpenLabyrinth QUestions, check out this example case:

Now what do we mean by this? First of all, we want to stress that this is not full blown natural language processing (NLP). NLP is often asked-for in virtual patient cases but is really hard to implement. The Maryland Virtual Patient, presented at the Medicine Meets Virtual Reality conference in 2009 by Nirenburg et al, was an impressive attempt at this. But it took millions of dollars and years to develop – rather out of reach for most of us.

No, this is text-matching against a set of predefined words or phrases, a much cruder solution. Our apologies to case authors who have to work a lot with such input because it is a nuisance to program in any system. We hope we have made things easier for our authors by providing a more flexible set of rules.

The educational advantage of this approach is that it allows the case author to pick the brains of the learner, asking for suggestions, without prompting or giving away the possible correct answer.

This feature is only currently available on development code. If you want early access to this, contact us through the web site.