Topological quantum reservoir computing: protected edge modes hold memory while bulk dynamics compute
Topological quantum reservoir computing moves quantum reservoir computing onto a one-dimensional Kitaev chain: topologically protected Majorana edge modes retain information over long times while faster bulk dynamics process the input, and only a linear readout is trained. Without interactions, edge retention grows exponentially with chain length while local write and read amplitudes are exponentially suppressed; a matched ancillary fermion makes the edge a programmable memory, with stronger writing shortening retention. Interactions let information load autonomously from bulk to edge at the cost of reduced retention, and this channel improves long-timescale temporal integration.