Were Lucky's (Bell Labs) and Widrow's (Stanford) adaptive-filter groups aware of each other in the mid-1960s, or did they converge on steepest-descent transversal filters independently?
Raised while promoting claim-lucky-1965-adaptive-equalizer-steepest-descent-transversal-filter. Two groups arrive at essentially the same object — an adjustable transversal (tapped delay line) filter whose weights are learned by iterative steepest descent to minimize an error — at almost the same moment and for different ends:
- Widrow & Hoff (Stanford, ~1960) — the LMS rule training the Adaline, later carried into adaptive antennas, noise cancelling, and seismic processing (claim-widrow-abandoned-multilayer-training-until-1985-backprop).
- Lucky (Bell Labs, 1965) — the automatic equalizer for voiceband telephone channels, breaking the ~2400 bps distortion barrier (claim-voiceband-distortion-barrier-capped-data-at-2400-bps).
Widrow & Lehr's 1990 retrospective (the source behind claim-madaline-rule-i-first-layer-trainable-second-fixed) credits "outstanding work by Lucky and others at Bell Laboratories," which shows Widrow was aware of Lucky by 1990 — but says nothing about mid-1960s cross-pollination in either direction.
What would answer it: check the reference lists of Lucky 1965 (does he cite Widrow & Hoff 1960 or Widrow's adaptive-filter work?) and of Widrow's 1960s papers (do they cite Lucky?); look for a shared citation ancestor (e.g. Kallmann's 1940s transversal-filter work); and see whether either man's later oral history addresses the parallel. This is a history-of-ideas thread, not a load-bearing verification — but a clean answer would be the intellectual payoff of this hop.
Progress — 2026-07-21
Half of the first test is now done. Lucky's 1965 bibliography was read in full (4 entries) and cites no Widrow-Hoff work (claim-lucky-1965-bibliography-omits-widrow-hoff-citation). Three Widrow-authored primaries — a 2013 CHM oral history, his 2005 IEEE SPM retrospective, and the 1990 Widrow & Lehr survey — consistently describe Lucky's algorithm as independently developed and different from LMS, with LMS entering Bell Labs' adaptive-modem work only afterward (claim-widrow-frames-lucky-equalizer-as-independent-invention-lms-later-graft). Synthesized at claim-lucky-widrow-independent-origination-favored-lucky-side-unverified: the weight of evidence favors independent origination, better than even odds, but every piece of it is Widrow's side of the story. Leaving this open, not answered — the second test (Lucky's own reference lists are covered; his own oral history / writing is not) and the shared-Kallmann- ancestor test are both still unresolved, routed together at question-lucky-own-account-of-widrow-hoff-awareness. Widrow's 1960s papers' own reference lists (do they cite Lucky, or pre-Lucky Bell Labs transversal-filter work?) also remain unchecked.