Yes, and the definition works reasonably well for the core arguments they are making in Section 5. I suspect they'll follow up with a full paper with more deta

Yes, and the definition works reasonably well for the core arguments they are making in Section 5. I suspect they'll follow up with a full paper with more details (and artifacts) of their proposed approach.