proofs
Provedalphamulti-device-learningrepartitioning

Alpha's repartitioning laws

map⁡(repartition⁡(range⁡(6),3))(length⁡)=[2,2,2]∧map⁡(repartition⁡(range⁡(6),5))(length⁡)=[2,1,1,1,1]\begin{aligned} & \htmlData{mx=11}{\htmlData{mx=9}{\operatorname{map}\left(\htmlData{mx=8}{\htmlData{mx=5}{\operatorname{repartition}}(\htmlData{mx=7}{\htmlData{mx=6}{\operatorname{range}}(6)}, 3)}\right)}(\htmlData{mx=10}{\operatorname{length}})} \mathrel{\htmlData{mx=13}{=}} \htmlData{mx=12}{\left[ 2, 2, 2 \right]} \\ & \htmlData{mx=0}{\land} \htmlData{mx=24}{\htmlData{mx=22}{\operatorname{map}\left(\htmlData{mx=21}{\htmlData{mx=18}{\operatorname{repartition}}(\htmlData{mx=20}{\htmlData{mx=19}{\operatorname{range}}(6)}, 5)}\right)}(\htmlData{mx=23}{\operatorname{length}})} \mathrel{\htmlData{mx=26}{=}} \htmlData{mx=25}{\left[ 2, 1, 1, 1, 1 \right]} \end{aligned}
What the notation means · or point at any symbol
e.fmap
map of repartition(range(6), 3): Lean's dot notation e.map means map applied to e. ((repartition (List.range 6) 3).map)
=Equality
map(repartition(range(6), 3))(length) and [2, 2, 2] are the same. ((repartition (List.range 6) 3).map List…)
∧And
Both hold: map(repartition(range(6), 3))(length) = [2, 2, 2], and map(repartition(range(6), 5))(length) = [2, 1, 1, 1, 1]. ((repartition (List.range 6) 3).map List…)
repartition
repartition: a definition from the library, not special notation. Its name is its meaning; open it in Lean to see how it is defined. (repartition)
range
List.range: a definition from the library, not special notation. Its name is its meaning; open it in Lean to see how it is defined. (List.range)
f(x)Function application
List.range applied to 6. Lean writes f x; papers write f(x). (List.range 6)
length
List.length: a definition from the library, not special notation. Its name is its meaning; open it in Lean to see how it is defined. (List.length)
[a, b]List
The list of 2, 2, 2, in order. ([2, 2, 2])

Six samples on three owners are shares of 2, 2, 2, and on five owners of 2, 1, 1, 1, 1 (the gates resume four-owner checkpoints on three and five owners). Model: definition alpha-multi-device-learning (revision 01a1255c-84f1-7e7f-a994-fc8c2d274bee).

How it was proved

Proved
10 October 2026
By
Ajax
Checked by
Lean's kernel (Lean 4.33.1, with Mathlib); the proof is signed and replays offline
Axioms
None
Proof
19 lines of Lean
Sign in to see the proofThe proof, its signed report and its history.

More proved statements