Data-Powered Algorithms - I - PowerPoint PPT Presentation

About This Presentation
Title:

Data-Powered Algorithms - I

Description:

Data-Powered Algorithms - I Bernard Chazelle Princeton University – PowerPoint PPT presentation

Number of Views:38
Avg rating:3.0/5.0
Slides: 75
Provided by: Berna103
Category:

less

Transcript and Presenter's Notes

Title: Data-Powered Algorithms - I


1
Data-PoweredAlgorithms - I
  • Bernard Chazelle
  • Princeton University

2
  • Amazing
  • predictions

3
Lord Kelvin ( 1824-1907 )
X-rays will prove to be a hoax"
Radio has no future. "
4
Albert Einstein ( 1932 )
There is not the slightest indication that
nuclear energy will ever be obtainable. "
5
Thomas Watson IBM Chairman (1943 )
"I think there is a world market for maybe five
computers."
6
Gordon Moore Intel Co-founder (1965 )
Computing power doubles every two years."
Moores Law
7
(No Transcript)
8
  • Moore's Law

9
In a few decades
Moores Law repealed
  • then what?

10
Lord Kelvin ( 1824-1907 )
Theres nothing to be discovered in physics
today. "
11
Lord Chazelle (2006)
Theres nothing to be discovered in computer
science today. "
12
Lord Chazelle (2006)
Theres nothing to be discovered in computer
science today. "
13
Lord Chazelle (2006)
Computing will be the most disruptive
scientific paradigm since quantum mechanics."
14
Lord Chazelle (2006)
and the end of Moores Law will make this
even more obvious."
15
What is computing ?
  • 4 big ideas
  • Universality
  • Duality
  • Self-reference
  • Tractability

16
  • Universality

17
control
18
Let em eat cake
19
Before Turing
data
20
Before Turing
data
21
Fishing
22
Fishing
Fishing manual
program
data
23
Fishing
Confucius
Fishing manual
program
data
24
needs to know how to fish
25
needs to know nothing about fishing
Fishing manual
program
data
26
knows nothing
control
27
turn bits into sounds
001010100010100010011111010001010
28
display/organize email
001010100010100010011111010001010
29
Earth simulator
algebra
001010100010100010011111010001010
30
  • All the same !

31
  • Duality

32
Saussure (1857-1913)
signifier
signified
data
program
Let em eat cake
33
This is not a pipe
Magritte
34
WHOS ON FIRST ?
Abbott and Costello
35
  • Self-reference

36
Print this
37
  • Virus !

Self-replication
Print this twice
Print this twice
Print this twice Print this twice
38
James Watson Francis Crick, 1953
39
signified
signifier
Duality gene/protein
Self-reference base pairs
40
  • Tractability

41
Protein folding
Scheduling
all seem intractable
Map coloring
Andrew
Wiles
Traveling salesman
Theorem proving
42
Protein folding
Scheduling
equivalent
Map coloring
Andrew
Wiles
Traveling salesman
Theorem proving
43
Protein folding
E-commerce security
intractable ?
Map coloring
Andrew
Wiles
Traveling salesman
Theorem proving
44
Two Amazing Consequences of Intractability
Zero Knowledge
Probabilistically Checkable Proofs
45
  • Zero Knowledge

46
dunno, but I wont tell you how much Im worth
Are you richer than me ?
Bill
Bob
47
So, whos richer ?
I wont tell you either
Bill
Bob
48
There exists a dialogue
Bill
Bob
49
blah blah blah blah blah blah blah blah blah blah
blah blah
blah blah blah blah blah blah blah blah blah blah
blah blah
Bill
Bob
50
at the end of which
Bill
Bob
51
1. They will know who is richer
2. They will have learned nothing else
( with probability 0.99999999999 )
Bill
Bob
52
Zero Knowledge
Prove it!
I have no nukes !
53
1. No UN inspections
2. Both parties try to cheat
54
  • P C P

55
Who will believe me?
Step 1 write proof in special
format
Step 2 verifier picks 5 random
words
My Proof of Riemanns Hypothesis
56
compiler
57
Theres something wrong. I REJECT !
Verifier
58
Everything looks fine. I ACCEPT !
Verifier is correct with probability 0.9999999
Verifier
59
Theres something wrong. I REJECT !
If my 2000-page proof is wrong in only one step,
how can verifier spot an error in 5 random words?
Verifier
60
Everything looks fine. I ACCEPT !
How does verifier know I proved Riemanns
hypothesis and not 224 ?
Verifier
61
  • The Algorithm

62
Very little does a lot
63
32 x 17
224 32
544
grade school
64
FFT
signal processing
65
RSA
encryption
e-commerce
66
PageRank
web search
67
Sloan Digital Sky Survey
10 petabytes (1MG)
10 petabytes/yr
Biomedical imaging
10,000 times the Library of Congress
150 petabytes/yr
68
Understanding Biological Function
  • 100s of sequenced genomes
  • Function of many genes unknown
  • 30 for yeast
  • ?

Genome Proteome
69
Interaction Networks
  • High-throughput experiments (Yeast two-hybrid,
    etc)
  • Form networks of interactions

70
protein-protein interaction networks
71
Analysis of Interaction Networks
Barabasi, AL. et al. (2004) Nat. Rev. Genet.
Spirin, V. et al. (2003) PNAS
Yeger-Lotem, E. et al. (2004) PNAS
72
Sciences of The Formula
math, physics, chemistry
73
Sciences of The Algorithm
Abu Abdullah Muhammad bin Musa al-Khwarizm
(780-850)
74
If Google is a religion, what is its God?
It would have to be The Algorithm.
Write a Comment
User Comments (0)
About PowerShow.com