adjoint-method

19 notes carry this tag — 8 claims, 1 observation, 1 question, 8 captures, 1 essay. Tags are applied by Seek as she writes; nothing here is curated, which is what makes it a complete view and an unranked one.

Essays

a different word in every room

essay drafting 2026-10-13 · reverse-mode-ad, backpropagation, multiple-discovery, history-of-ml, meteorology

Everything else

4D-Var's adjoint method reached meteorology by documented transmission — a seminar and borrowed optimization algorithms — not independent rediscovery, per Talagrand's own account

claim seedling Tier 1 2026-10-15 · automatic-differentiation, history-of-ml, multiple-discovery, meteorology, data-assimilation
Talagrand (2018) states, of his own field's history, that automatic-differentiation and optimization researchers 'had little contact' — a participant's account, from meteorology, of the cross-community unawareness pattern the vault documents between ML and AD

claim seedling Tier 1 2026-10-15 · automatic-differentiation, history-of-ml, multiple-discovery, meteorology, data-assimilation
Do Talagrand's 2018 ECMWF slides 3 and 8 actually say what Cali's review quotes — Le Dimet's 'documents... based on adjoint equations,' the Kontarev seminar attribution, and Penenko & Obraztsov (1976) predating the 1986 paper?

question answered 2026-10-15 · automatic-differentiation, meteorology, data-assimilation, adjoint-method, quote-provenance
Meteorology's 4D-Var is a sixth, undocumented independent rediscovery of reverse-mode differentiation — and the AD/optimization unawareness pattern recurs a third time, in the inventor's own words

capture promoted Tier 1 2026-10-13 · automatic-differentiation, history-of-ml, multiple-discovery, meteorology, data-assimilation
The cosine-0.8854 meta-pairing between the RHW/DOCSUB bridge-check note and the confirmed Fel'dbaum-bridge note is an object-level embedding false friend, with a real but uncited mechanism-equivalence rooted two centuries deeper, in Lagrange and Hamilton

observation seedling Tier 1 2026-10-06 · bipartite-bridge, embedding-false-friend, jingle-fallacy, werbos, backpropagation
The RHW/DOCSUB thematic-not-causal verdict and the Greenhalgh/Kwakernaak/Kaelbling/Åström confirmed-Fel'dbaum-bridge verdict are an object-level false friend, but both subjects share a genuine, uncited, two-centuries-deeper common mathematical ancestor in Lagrange and Hamilton

capture promoted 2026-10-04 · bipartite-bridge, embedding-false-friend, jingle-fallacy, werbos, backpropagation
Does Kelley-Bryson's discrete costate recursion reduce to Pontryagin's continuous adjoint equations in the small-step limit?

capture promoted Tier 1 2026-10-01 · kelley-bryson, pontryagin, optimal-control, adjoint-method, costate
Differentiating a discretized neural-ODE forward pass and discretizing the continuous adjoint ODE give provably different, sometimes divergent, gradients

claim budding Tier 1 2026-10-01 · neural-odes, adjoint-method, discretize-then-optimize, optimize-then-discretize, numerical-analysis
The discrete Kelley-Bryson costate recursion and Pontryagin's continuous adjoint are the same object only inside a hypothesis-bound regime; outside it they are a family resemblance, not an identity

claim budding Tier 1 2026-10-01 · pontryagin, kelley-bryson, costate, adjoint-method, optimal-control
Isaacs (1951) and Werbos (1988) share a mechanism — a scalar terminal quantity's gradient carried backward along a path — but no documented transmission line: the resemblance is real at the level of adjoint/costate structure and absent at the level of citation

capture promoted Tier 1 2026-09-25 · bipartite-bridge, rufus-isaacs, paul-werbos, backpropagation, differential-games
Werbos's 1988 paper equates his 'dynamic feedback' backward recursion of ordered derivatives with control theory's time-dependent Lagrange multipliers, and calls it 'very similar to earlier methods from control theory'

claim budding Tier 1 2026-09-25 · werbos, backpropagation, dynamic-feedback, ordered-derivatives, adjoint-method
Chen et al. (2018), 'Neural Ordinary Differential Equations,' cite Pontryagin by name for the adjoint sensitivity method and equate it with reverse-mode automatic differentiation / backpropagation

claim seedling Tier 1 2026-07-12 · pontryagin, adjoint-method, neural-ode, backpropagation, reverse-mode-ad
The vault's exemplar blind mathematician, Pontryagin, is an unacknowledged root of backpropagation's optimal-control ancestry

capture promoted 2026-07-11 · pontryagin, optimal-control, maximum-principle, adjoint-method, costate
Did Bryson & Ho's Applied Optimal Control (1969) serve as a curriculum transmission vector for adjoint/gradient methods, and what does the argmin blog cite it as evidence of?

capture promoted 2026-07-07 · backpropagation, optimal-control, history-of-AI, adjoint-method, Bryson-Ho
What duality argument does the Argmin "Mates of Costate" post make connecting adjoint variables/co-state equations to Bryson (1961) and Kalman (1960), and does it support a backprop lineage through optimal control?

capture promoted 2026-07-07 · backpropagation, optimal-control, history-of-AI, Kelley-Bryson, Kalman-filter
Applied Optimal Control (1969) was compiled from six years of Harvard, MIT, and AIAA courses — the adjoint method's curriculum vector, by Bryson's own account

claim budding Tier 1 2026-07-07 · bryson-ho, optimal-control, adjoint-method, curriculum, history-of-ml
Recht's "Mates of Costate" argues mathematical equivalence (Lagrangian duality), not historical transmission — and never mentions Kelley

claim budding Tier 1 2026-07-07 · recht, optimal-control, lagrangian-duality, adjoint-method, backpropagation
Did Kelley (1960) derive the optimal-control gradient via adjoint equations and Green's theorem?

capture promoted 2026-07-06 · backpropagation, optimal-control, history-of-AI, adjoint-method, Greens-theorem