C-LARA

We have first versions of two large items from the priority list, parallelism and better support for non-AI languages. The Melbourne students are starting to set up their projects.

Priority list

Parallelism. A first version of parallelism for annotation is now installed on the server. Processing is much faster, particularly if you use MWE. I have been testing, and so far everything appears to work correctly. I think there are several easy things we can do to get further efficiency gains.

Support for non-AI languages. I have installed new functionality in the Edit Images screen, so that it now also shows the different versions of the text for each page in editable form. This should make it much easier to see when things don’t line up, a very common problem when you’re working with non-AI languages, and correct mistakes.

This functionality hasn’t yet been tested much. I am meeting Sophie Rendina online tomorrow (Thursday), and we’ll look at one of her texts together to see how it works in practice.

Encyclopaedia article

The Palgrave Encyclopaedia are still thinking about whether to accept the AI author for our article.

Melbourne students

I have been talking more with the students at Melbourne Uni about their projects. The Voice Recorder project has now developed a clear spec for what they will do, and will soon start on implementation work.

Looking around in the OpenAI doc, I find that there’s a way to use the Image Annotator that I hadn’t previously thought of: once you’ve got an outline for an object, you can use it to resubmit an image generation request, asking for only the part within the outline to be changed. It’s hard to say how well this will work in practice, but it opens up interesting possibilities.

Next Zoom call

The next call will be at:

Thu Aug 22 2024, 18:00 Adelaide (= 08.30 Iceland = 09.30 Ireland/Faroe Islands = 10.30 Europe = 11.30 Israel = 12.00 Iran = 16.30 China = 18:30 Melbourne = 19.30 New Caledonia)