Compositional neurosymbolic representations enable efficient active exploration
P. Michael Furlong, Nicole S.-Y. Dumont, Rika Antonova, Jeff Orchard, Chris Eliasmith · 2026
Read Compositional neurosymbolic representations enable efficient active exploration by P. Michael Furlong, Nicole S.-Y. Dumont, Rika Antonova, Jeff Orchard, Chris Eliasmith (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13594323.txt.
Description
Furlong and co-authors show in 2026 how a brain like vector code can make self driven search far cheaper. Long running robots must learn as they roam, yet standard search tools grow in cost with each new sight. On small gear, that growth ends the run early.
Their code packs parts into one fixed size vector, so time and memory stay flat as data grow. A search method built on the code ran sixty to two hundred times faster than top rivals at the same score. On brain style chips, energy per sample fell thirty to one hundred eighty eight fold.
Catalog record
| CREATOR | P. Michael Furlong, Nicole S.-Y. Dumont, Rika Antonova, Jeff Orchard, Chris Eliasmith |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 153 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |