Normal numbers and automatic complexity alexander.shen@lirmm.fr, - - PowerPoint PPT Presentation

normal numbers and automatic complexity
SMART_READER_LITE
LIVE PREVIEW

Normal numbers and automatic complexity alexander.shen@lirmm.fr, - - PowerPoint PPT Presentation

Normal numbers and automatic complexity alexander.shen@lirmm.fr, www.lirmm.fr/~ashen LIRMM CNRS & University of Montpellier September 2017, FCT alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity Individual


slide-1
SLIDE 1

Normal numbers and automatic complexity

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen

LIRMM CNRS & University of Montpellier

September 2017, FCT

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-2
SLIDE 2

Individual random sequence

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-3
SLIDE 3

Individual random sequence

If a coin tossing gives 00000 . . . or 01010101 . . ., we become suspicious

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-4
SLIDE 4

Individual random sequence

If a coin tossing gives 00000 . . . or 01010101 . . ., we become suspicious individual sequence of 0 and 1: can it be “random”/“nonrandom”?

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-5
SLIDE 5

Individual random sequence

If a coin tossing gives 00000 . . . or 01010101 . . ., we become suspicious individual sequence of 0 and 1: can it be “random”/“nonrandom”? von Mises (1919): Kollektiv: a basic notion of probability theory; frequency stability

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-6
SLIDE 6

Individual random sequence

If a coin tossing gives 00000 . . . or 01010101 . . ., we become suspicious individual sequence of 0 and 1: can it be “random”/“nonrandom”? von Mises (1919): Kollektiv: a basic notion of probability theory; frequency stability Borel: normal numbers

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-7
SLIDE 7

Normal numbers

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-8
SLIDE 8

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-9
SLIDE 9

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits simply normal: #0(n)/n → 1/2, #1(n) → 1/2

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-10
SLIDE 10

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits simply normal: #0(n)/n → 1/2, #1(n) → 1/2 #00(n) = number of occurences of 00 in the first n positions

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-11
SLIDE 11

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits simply normal: #0(n)/n → 1/2, #1(n) → 1/2 #00(n) = number of occurences of 00 in the first n positions #00(n) + #01(n) + #10(n) + #11(n) = n

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-12
SLIDE 12

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits simply normal: #0(n)/n → 1/2, #1(n) → 1/2 #00(n) = number of occurences of 00 in the first n positions #00(n) + #01(n) + #10(n) + #11(n) = n normal: #00(n)/n → 1/4 and the same for all other blocks (any length)

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-13
SLIDE 13

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits simply normal: #0(n)/n → 1/2, #1(n) → 1/2 #00(n) = number of occurences of 00 in the first n positions #00(n) + #01(n) + #10(n) + #11(n) = n normal: #00(n)/n → 1/4 and the same for all other blocks (any length) Another approach: cut the sequence into k-bit blocks and count the number of blocks of each type (aligned

  • ccurrences); these two definitions are equivalent

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-14
SLIDE 14

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits simply normal: #0(n)/n → 1/2, #1(n) → 1/2 #00(n) = number of occurences of 00 in the first n positions #00(n) + #01(n) + #10(n) + #11(n) = n normal: #00(n)/n → 1/4 and the same for all other blocks (any length) Another approach: cut the sequence into k-bit blocks and count the number of blocks of each type (aligned

  • ccurrences); these two definitions are equivalent

almost all numbers are normal

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-15
SLIDE 15

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits simply normal: #0(n)/n → 1/2, #1(n) → 1/2 #00(n) = number of occurences of 00 in the first n positions #00(n) + #01(n) + #10(n) + #11(n) = n normal: #00(n)/n → 1/4 and the same for all other blocks (any length) Another approach: cut the sequence into k-bit blocks and count the number of blocks of each type (aligned

  • ccurrences); these two definitions are equivalent

almost all numbers are normal e, π, √ 2??? Champernowne: 0 1 10 11 100 101 110 . . .

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-16
SLIDE 16

Normal numbers

00100111010111 . . . #0(n) = number of 0 among the first n bits simply normal: #0(n)/n → 1/2, #1(n) → 1/2 #00(n) = number of occurences of 00 in the first n positions #00(n) + #01(n) + #10(n) + #11(n) = n normal: #00(n)/n → 1/4 and the same for all other blocks (any length) Another approach: cut the sequence into k-bit blocks and count the number of blocks of each type (aligned

  • ccurrences); these two definitions are equivalent

almost all numbers are normal e, π, √ 2??? Champernowne: 0 1 10 11 100 101 110 . . . Wall: α is normal, n integer ⇒ nα, α/n normal

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-17
SLIDE 17

Randomness as incompressibility

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-18
SLIDE 18

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-19
SLIDE 19

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-20
SLIDE 20

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient Martin-L¨

  • f: random ⇔ obeys all “effective laws” of

probability theory

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-21
SLIDE 21

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient Martin-L¨

  • f: random ⇔ obeys all “effective laws” of

probability theory Kolmogorov, Levin, Chaitin,. . . : randomness = incompressibility of prefixes

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-22
SLIDE 22

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient Martin-L¨

  • f: random ⇔ obeys all “effective laws” of

probability theory Kolmogorov, Levin, Chaitin,. . . : randomness = incompressibility of prefixes 000 . . . 000 not random: short description: “million zeros”

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-23
SLIDE 23

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient Martin-L¨

  • f: random ⇔ obeys all “effective laws” of

probability theory Kolmogorov, Levin, Chaitin,. . . : randomness = incompressibility of prefixes 000 . . . 000 not random: short description: “million zeros” What is “description”? Different answers possible

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-24
SLIDE 24

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient Martin-L¨

  • f: random ⇔ obeys all “effective laws” of

probability theory Kolmogorov, Levin, Chaitin,. . . : randomness = incompressibility of prefixes 000 . . . 000 not random: short description: “million zeros” What is “description”? Different answers possible Normality = weak randomness

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-25
SLIDE 25

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient Martin-L¨

  • f: random ⇔ obeys all “effective laws” of

probability theory Kolmogorov, Levin, Chaitin,. . . : randomness = incompressibility of prefixes 000 . . . 000 not random: short description: “million zeros” What is “description”? Different answers possible Normality = weak randomness Limited class of descriptions: finite memory

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-26
SLIDE 26

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient Martin-L¨

  • f: random ⇔ obeys all “effective laws” of

probability theory Kolmogorov, Levin, Chaitin,. . . : randomness = incompressibility of prefixes 000 . . . 000 not random: short description: “million zeros” What is “description”? Different answers possible Normality = weak randomness Limited class of descriptions: finite memory Well-known since 1960s (Agafonov, Schnorr, Stimm, Dai, Lathrop, Lutz, Mayordomo, Becher, Heiber,. . . )

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-27
SLIDE 27

Randomness as incompressibility

Individual random sequences: plausible as outcomes of coin tossing experiment Normality is necessary but hardly sufficient Martin-L¨

  • f: random ⇔ obeys all “effective laws” of

probability theory Kolmogorov, Levin, Chaitin,. . . : randomness = incompressibility of prefixes 000 . . . 000 not random: short description: “million zeros” What is “description”? Different answers possible Normality = weak randomness Limited class of descriptions: finite memory Well-known since 1960s (Agafonov, Schnorr, Stimm, Dai, Lathrop, Lutz, Mayordomo, Becher, Heiber,. . . )

  • ur (small) contribution: clean definitions and proofs

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-28
SLIDE 28

Kolmogorov complexity: framework

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-29
SLIDE 29

Kolmogorov complexity: framework

Relation D(p, x) on strings: “p is a description of x”

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-30
SLIDE 30

Kolmogorov complexity: framework

Relation D(p, x) on strings: “p is a description of x” CD(x) = min{|p|: D(p, x)}

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-31
SLIDE 31

Kolmogorov complexity: framework

Relation D(p, x) on strings: “p is a description of x” CD(x) = min{|p|: D(p, x)} trivial D: Λ is a description of everything, CD(x) = 0

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-32
SLIDE 32

Kolmogorov complexity: framework

Relation D(p, x) on strings: “p is a description of x” CD(x) = min{|p|: D(p, x)} trivial D: Λ is a description of everything, CD(x) = 0 restrictions for D needed

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-33
SLIDE 33

Kolmogorov complexity: framework

Relation D(p, x) on strings: “p is a description of x” CD(x) = min{|p|: D(p, x)} trivial D: Λ is a description of everything, CD(x) = 0 restrictions for D needed plain Kolmogorov complexity: D is a c.e. functional relation (only one x for each p)

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-34
SLIDE 34

Kolmogorov complexity: framework

Relation D(p, x) on strings: “p is a description of x” CD(x) = min{|p|: D(p, x)} trivial D: Λ is a description of everything, CD(x) = 0 restrictions for D needed plain Kolmogorov complexity: D is a c.e. functional relation (only one x for each p)

  • ur requirement: the relation D is an O(1)-valued

function (each description describes O(1) objects) that “can be checked with finite memory”

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-35
SLIDE 35

Kolmogorov complexity: framework

Relation D(p, x) on strings: “p is a description of x” CD(x) = min{|p|: D(p, x)} trivial D: Λ is a description of everything, CD(x) = 0 restrictions for D needed plain Kolmogorov complexity: D is a c.e. functional relation (only one x for each p)

  • ur requirement: the relation D is an O(1)-valued

function (each description describes O(1) objects) that “can be checked with finite memory” corresponding class of complexity functions CD allows us to characterize normal sequences as incompressible

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-36
SLIDE 36

Technical details

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-37
SLIDE 37

Technical details

Idea: D(p, x) is automatic if it can be checked reading p and x bit by bit, with finite memory

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-38
SLIDE 38

Technical details

Idea: D(p, x) is automatic if it can be checked reading p and x bit by bit, with finite memory similar to rational relations but no initial/final state

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-39
SLIDE 39

Technical details

Idea: D(p, x) is automatic if it can be checked reading p and x bit by bit, with finite memory similar to rational relations but no initial/final state Formal definition: graph; edges labeled by (u, v), (u, ε), (ε, u), (ε, ε)

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-40
SLIDE 40

Technical details

Idea: D(p, x) is automatic if it can be checked reading p and x bit by bit, with finite memory similar to rational relations but no initial/final state Formal definition: graph; edges labeled by (u, v), (u, ε), (ε, u), (ε, ε) path ⇒ pair of strings

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-41
SLIDE 41

Technical details

Idea: D(p, x) is automatic if it can be checked reading p and x bit by bit, with finite memory similar to rational relations but no initial/final state Formal definition: graph; edges labeled by (u, v), (u, ε), (ε, u), (ε, ε) path ⇒ pair of strings D = the set of all pairs that can be read along paths

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-42
SLIDE 42

Technical details

Idea: D(p, x) is automatic if it can be checked reading p and x bit by bit, with finite memory similar to rational relations but no initial/final state Formal definition: graph; edges labeled by (u, v), (u, ε), (ε, u), (ε, ε) path ⇒ pair of strings D = the set of all pairs that can be read along paths “automatic relations”

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-43
SLIDE 43

Technical details

Idea: D(p, x) is automatic if it can be checked reading p and x bit by bit, with finite memory similar to rational relations but no initial/final state Formal definition: graph; edges labeled by (u, v), (u, ε), (ε, u), (ε, ε) path ⇒ pair of strings D = the set of all pairs that can be read along paths “automatic relations” multiplication and division by an integer constant are automatic relations

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-44
SLIDE 44

Technical details

Idea: D(p, x) is automatic if it can be checked reading p and x bit by bit, with finite memory similar to rational relations but no initial/final state Formal definition: graph; edges labeled by (u, v), (u, ε), (ε, u), (ε, ε) path ⇒ pair of strings D = the set of all pairs that can be read along paths “automatic relations” multiplication and division by an integer constant are automatic relations union/composition of two automatic relations is automatic

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-45
SLIDE 45

Optimal decompressors

Theorem (Becher, Heiber) A sequence x1x2x3 . . . is normal ⇔ lim inf CD(x1 . . . xn)/n ≥ 1 for every automatic O(1)-valued relation D(p, x)

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-46
SLIDE 46

Part 1: non-normal sequences are compressible

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-47
SLIDE 47

Part 1: non-normal sequences are compressible

assume that different k-bit blocks have different frequencies

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-48
SLIDE 48

Part 1: non-normal sequences are compressible

assume that different k-bit blocks have different frequencies use standard block coding (Shannon, Fano, Huffman)

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-49
SLIDE 49

Part 1: non-normal sequences are compressible

assume that different k-bit blocks have different frequencies use standard block coding (Shannon, Fano, Huffman) [frequent blocks have shorter codes]

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-50
SLIDE 50

Part 1: non-normal sequences are compressible

assume that different k-bit blocks have different frequencies use standard block coding (Shannon, Fano, Huffman) [frequent blocks have shorter codes] block coding uses finite memory

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-51
SLIDE 51

Part 1: non-normal sequences are compressible

assume that different k-bit blocks have different frequencies use standard block coding (Shannon, Fano, Huffman) [frequent blocks have shorter codes] block coding uses finite memory Technical: select a subsequence that has limit frequencies; use these frequencies for block coding, use convexity of entropy function

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-52
SLIDE 52

Part 2: normal sequences are not compressible

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-53
SLIDE 53

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . .

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-54
SLIDE 54

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . . Some automatic O(1) relation D

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-55
SLIDE 55

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . . Some automatic O(1) relation D Why x1x2 . . . xN is not compressible?

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-56
SLIDE 56

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . . Some automatic O(1) relation D Why x1x2 . . . xN is not compressible? Split it into k-bit blocks X1X2 . . . XM

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-57
SLIDE 57

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . . Some automatic O(1) relation D Why x1x2 . . . xN is not compressible? Split it into k-bit blocks X1X2 . . . XM description p can be also split into corresponding blocks

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-58
SLIDE 58

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . . Some automatic O(1) relation D Why x1x2 . . . xN is not compressible? Split it into k-bit blocks X1X2 . . . XM description p can be also split into corresponding blocks trivial crucial lemma: CD(xy) ≥ CD(x) + CD(y)

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-59
SLIDE 59

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . . Some automatic O(1) relation D Why x1x2 . . . xN is not compressible? Split it into k-bit blocks X1X2 . . . XM description p can be also split into corresponding blocks trivial crucial lemma: CD(xy) ≥ CD(x) + CD(y) all k-bit strings appear equally often among X1, X2, . . . , XM

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-60
SLIDE 60

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . . Some automatic O(1) relation D Why x1x2 . . . xN is not compressible? Split it into k-bit blocks X1X2 . . . XM description p can be also split into corresponding blocks trivial crucial lemma: CD(xy) ≥ CD(x) + CD(y) all k-bit strings appear equally often among X1, X2, . . . , XM most of k-bit strings are incompressible (even in Kolmogorov’s sense)

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-61
SLIDE 61

Part 2: normal sequences are not compressible

Normal sequence x1x2 . . . Some automatic O(1) relation D Why x1x2 . . . xN is not compressible? Split it into k-bit blocks X1X2 . . . XM description p can be also split into corresponding blocks trivial crucial lemma: CD(xy) ≥ CD(x) + CD(y) all k-bit strings appear equally often among X1, X2, . . . , XM most of k-bit strings are incompressible (even in Kolmogorov’s sense) so the economy is negligible compared to length

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-62
SLIDE 62

What do we get as byproducts?

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-63
SLIDE 63

What do we get as byproducts?

Hall: α is normal, n integer ⇒ nα and α/n are normal

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-64
SLIDE 64

What do we get as byproducts?

Hall: α is normal, n integer ⇒ nα and α/n are normal Proof: multiplication and division by a constant are O(1)-valued automatic relations and composition of automatic relations is automatic

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-65
SLIDE 65

What do we get as byproducts?

Hall: α is normal, n integer ⇒ nα and α/n are normal Proof: multiplication and division by a constant are O(1)-valued automatic relations and composition of automatic relations is automatic aligned definition ⇔ non-aligned definition

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-66
SLIDE 66

What do we get as byproducts?

Hall: α is normal, n integer ⇒ nα and α/n are normal Proof: multiplication and division by a constant are O(1)-valued automatic relations and composition of automatic relations is automatic aligned definition ⇔ non-aligned definition Proof: the criterion can be proven for non-aligned definition in a similar way

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-67
SLIDE 67

What do we get as byproducts?

Hall: α is normal, n integer ⇒ nα and α/n are normal Proof: multiplication and division by a constant are O(1)-valued automatic relations and composition of automatic relations is automatic aligned definition ⇔ non-aligned definition Proof: the criterion can be proven for non-aligned definition in a similar way Agafonov: automatic selection rule preserves normality

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-68
SLIDE 68

What do we get as byproducts?

Hall: α is normal, n integer ⇒ nα and α/n are normal Proof: multiplication and division by a constant are O(1)-valued automatic relations and composition of automatic relations is automatic aligned definition ⇔ non-aligned definition Proof: the criterion can be proven for non-aligned definition in a similar way Agafonov: automatic selection rule preserves normality Proof: if a selected subsequence is compressible, this compression can be used together with uncompressed description of the remaining terms (some care needed)

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-69
SLIDE 69

What do we get as byproducts?

Hall: α is normal, n integer ⇒ nα and α/n are normal Proof: multiplication and division by a constant are O(1)-valued automatic relations and composition of automatic relations is automatic aligned definition ⇔ non-aligned definition Proof: the criterion can be proven for non-aligned definition in a similar way Agafonov: automatic selection rule preserves normality Proof: if a selected subsequence is compressible, this compression can be used together with uncompressed description of the remaining terms (some care needed) Piatetski-Shapiro theorem: if no block appear c times more

  • ften then they should, the sequence is normal

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity

slide-70
SLIDE 70

What do we get as byproducts?

Hall: α is normal, n integer ⇒ nα and α/n are normal Proof: multiplication and division by a constant are O(1)-valued automatic relations and composition of automatic relations is automatic aligned definition ⇔ non-aligned definition Proof: the criterion can be proven for non-aligned definition in a similar way Agafonov: automatic selection rule preserves normality Proof: if a selected subsequence is compressible, this compression can be used together with uncompressed description of the remaining terms (some care needed) Piatetski-Shapiro theorem: if no block appear c times more

  • ften then they should, the sequence is normal

THANKS! [https://arxiv.org/pdf/1701.09060.pdf]

alexander.shen@lirmm.fr, www.lirmm.fr/~ashen Normal numbers and automatic complexity