David Marr

British neuroscientist (1945–1980) who recurs in the vault along two distinct threads. His three levels of analysis — computational, algorithmic, implementational — are a standing framework the vault reaches for when separating what a system computes from how it is wired, appearing across the substrate-laundering and rational-analysis notes. Separately, his 1969 theory of the cerebellar cortex (extended by Ito 1970 and Albus 1971) is the seed of one of the vault's cases of a computational circuit theory whose synaptic rule was confirmed experimentally only later — though the rule that was confirmed, parallel-fiber–Purkinje-cell long-term depression, was Ito's 1970 revision, not Marr's: Marr's own model assumed synaptic potentiation, and the sign flip is a defining difference between the two. The confirmation came from Ito's own lab in 1982.

References

• claim-marr-ito-cerebellar-ltd-1969-confirmed-ito-kano-1982 — the 1969/70 cerebellar LTD theory, confirmed 1982
• observation-substrate-laundering-across-marr-levels — Marr's levels as a lens on substrate independence
• claim-rational-analysis-anderson-optimal-solution-marr-level — rational analysis at Marr's computational level
• claim-anderson-traces-rational-analysis-lineage-to-brunswik-gibson-marr-shepard — Anderson's own ancestry claim, with Marr as one of two "more recent" figures giving rigorous form to an older Brunswik/Gibson adaptationist tradition
• Pattern context: claim-neural-circuit-theories-wait-decades-for-anatomical-confirmation
• observation-yuille-person-bridges-hawking-and-marr-entity-hubs — Alan Yuille's 2003 ICCV Marr Prize is the unexpected hinge linking this hub to Stephen Hawking's

Log

• 2026-08-28: Anderson's 1991 BBS précis names Marr (with Shepard) as providing "rigorous developments" of a perception-adaptationist tradition traced back through Brunswik and Gibson — a positioning of Marr the vault hadn't recorded before, distinct from the three-levels and cerebellar-theory material above (claim-anderson-traces-rational-analysis-lineage-to-brunswik-gibson-marr-shepard).
• 2026-09-13 (propagation-repair): cerebellar-theory paragraph corrected. Was: Marr's 1969 theory described as the case "whose synaptic rule — parallel-fiber–Purkinje-cell long-term depression — was confirmed experimentally only later," attributing the LTD rule to Marr. Now: the LTD rule is Ito's (1970) alone; Marr's 1969 model assumed synaptic potentiation, and the 2020 Kawato/Ohmae/Hoang/Sanger review names "potentiation versus depression of synaptic plasticity" as a defining difference between the models. Propagates the AUDIT 2026-07-19 correction to claim-marr-ito-cerebellar-ltd-1969-confirmed-ito-kano-1982. Unchanged: the 1969/1970/1971 Marr–Ito–Albus lineage, Marr's priority as the model the others extend, and the 1982 same-lab confirmation.
• 2026-11-02: the ICCV 2003 Marr Prize's own award page names Marr as the prize's namesake ("a Cambridge (UK) theoretical neuroscientist... who made profound contributions to the theory of both human and machine vision at the MIT AI Lab") and records Alan Yuille as a 2003 co-winner — the same Yuille who did his PhD under Stephen Hawking, making this hub's first recorded link to the Hawking hub (observation-yuille-person-bridges-hawking-and-marr-entity-hubs).