FLEX scores 0.99 as robots learn broader human preferences

MIT and NYU researchers say FLEX turns human corrections into broader robot constraints, scoring 0.99 in a navigation test while leaving physical, closed-loop use unproven.