Until now we've been differentiating content description by:
Simple Content Description: give me a single object, get a description back
Complex Content Description: give me a scene of inter-related items, get a description back
For the future of Content Description, to get some really nuanced content, we will want to consider things like abstract entities which are only tenuously linked to the resources in the scene
Example: I'm in a square in a country that is suffering a severe economic crisis
The scene will need to include important context like this #36
A really amazing function would be able to find quite personal things, for example identifying that one of the characters has a particular memory triggered by some music that is playing
Until now we've been differentiating content description by:
For the future of Content Description, to get some really nuanced content, we will want to consider things like abstract entities which are only tenuously linked to the resources in the scene
Example: I'm in a square in a country that is suffering a severe economic crisis
The scene will need to include important context like this #36
A really amazing function would be able to find quite personal things, for example identifying that one of the characters has a particular memory triggered by some music that is playing