Skip to content

Value for OpenNARS

Patrick Hammer edited this page Feb 17, 2019 · 4 revisions

Originally, ANSNA was derived from Tony's ALANN and my work on the prototype-based OpenNARS vision channel, with a move from truth evaluation based on similarities of truth value matrices to truth evaluation over bit vectors. Now it's more clear to me that SDR encodings can potentially become the foundation of a SDR-based general-purpose sensory channel in OpenNARS, that maps incoming events with SDR's attached to one of n prototypes that enter the reasoner. This would allow to test SDR properties together with OpenNARS temporal framing ideas as proposed in https://link.springer.com/chapter/10.1007%2F978-3-319-97676-1_8 that are the reason behind OpenNARS's success in temporal inference control and procedure learning ANSNA has yet failed to match.