In this paper we prove the intractability of learning several classes of Boolean functions in the distribution-free model (also called the Probably Approximately Correct or PAC model) of learning from examples. These results are representation independent, in that they hold regardless of the syntactic form in which the learner chooses to represent its hypotheses. Our methods reduce the problems of cracking a number of well-known public-key cryptosystems to the learning problems. We prove that a polynomial-time learning algorithm for Boolean formulae, deterministic finite automata or constant-depth threshold circuits would have dramatic consequences for cryptography and number theory: in particular, such an algorithm could be used to break the RSA cryptosystem, factor Blum integers (composite numbers equivalent to 3 modulo 4), and detect quadratic residues. The results hold even if the learning algorithm is only required to obtain a slight advantage in prediction over random guessing. T...
|
7271
|
Computers and Intractability - A Guide to the Theory of NP-Completeness
– Garey, Johnson
- 1979
|
|
2091
|
A method for obtaining digital signatures and public-key cryptosystems
– Rivest, Shamir, et al.
- 1997
|
|
2010
|
The Design and Analysis of Computer Algorithms
– Aho, Hopcroft, et al.
- 1974
|
|
1876
|
New directions in cryptogra-phy
– Diffie, Hellman
- 1976
|
|
1328
|
A theory of the learnable
– Valiant
- 1984
|
|
525
|
Learnability and the Vapnik-Chervonenkis dimension
– Blumer, Ehrenfeucht, et al.
- 1989
|
|
483
|
How to construct random functions
– GOLDREICH, GOLDWASSER, et al.
- 1986
|
|
463
|
A measure of asymptotic efficiency for tests of a hypothesis based on the sum of observations
– Chernoff
- 1952
|
|
460
|
How to Generate Cryptographically Strong Sequences of Pseudo-Random Bits
– Blum, Micali
- 1984
|
|
457
|
The strength of weak learnability
– Schapire
- 1990
|
|
409
|
Theory and Applications of Trapdoor Functions
– Yao
- 1982
|
|
365
|
Learning Regular Sets from Queries and Counterexamples
– Angluin
- 1987
|
|
224
|
Hellman: New Directions in Cryptography
– Die, Martin
- 1976
|
|
216
|
Digital signatures and Public-key Functions as Factorization
– Rabin
- 1979
|
|
213
|
Fast probabilistic algorithms for hamiltonian circuits and matchings
– Angluin, Valiant
- 1979
|
|
211
|
Constant depth circuits, Fourier transform, and learnability
– Linial, Mansour, et al.
- 1989
|
|
179
|
Learning from noisy examples
– Angluin, Laird
- 1988
|
|
175
|
Training a 3-node neural network is NPcomplete
– Blum
|
|
174
|
Complexity of automaton identification from given data
– Gold
- 1978
|
|
169
|
Computational limitations on learning from examples
– Pitt, Valiant
- 1988
|
|
140
|
On the learnability of Boolean formulae
– Kearns, Li, et al.
- 1987
|
|
111
|
RSA and Rabin Functions: Certain Parts are as Hard as the Whole
– Alexi, Chor, et al.
|
|
99
|
One-Way Function and Pseudorandom Generators
– Levin
- 1987
|
|
97
|
Log-depth circuits for division and related problem
– Beame, Cook, et al.
- 1986
|
|
97
|
Constant depth reducibility
– Chandra, Stockmeyer, et al.
- 1984
|
|
91
|
When won’t membership queries help
– Angluin, Kharitonov
- 1995
|
|
83
|
Equivalence of models for polynomial learnability
– Haussler, Kearns, et al.
- 1991
|
|
64
|
On the Markov chain simulation method for uniform combinatorial distributions and simulated
– Aldous
- 1987
|
|
63
|
The minimum consistent DFA problem cannot be approximated within any polynomial
– Pitt, Warmuth
- 1989
|
|
46
|
On threshold circuits and polynomial computations
– Reif, Tate
- 1992
|
|
25
|
On taking roots in finite fields
– Adleman, Manders, et al.
- 1977
|
|
23
|
A polynomial-time algorithm for learning kvariable pattern languages from examples
– Kearns, Pitt
- 1989
|
|
23
|
Reductions among prediction problems: On the difficulty of predicting automata
– Pitt, Warmuth
- 1988
|
|
19
|
Ameasure of asymptotic e ciency for tests of a hypothesis based on the sum of observations
– Cherno
- 1952
|
|
16
|
Complexity of automaton identi cation from given sets. Information and Control 37:302{320
– Gold
- 1978
|
|
16
|
Primality and Cryptography
– Kranakis
- 1986
|
|
11
|
Learning in neural networks
– Judd
- 1988
|
|
9
|
Lecture notes on the complexity of some problems in number theory
– Angluin
- 1982
|
|
7
|
On the learnability of finite automata
– Li, Vazirani
- 1988
|
|
6
|
Reductions among prediction problems: on the di culty of predicting automata
– Pitt, Warmuth
- 1988
|
|
5
|
An ~ O(n 0:4 )-approximation algorithm for 3-coloring
– Blum
- 1989
|
|
5
|
A general lower bound on the number of examples needed for learning
– Valiant
- 1989
|
|
5
|
A new approximate graph coloring algorithm
– Wigderson
- 1982
|
|
5
|
On the difficulty of finding small consistent decision trees
– Hancock
- 1989
|
|
3
|
On taking roots in nite elds
– Adleman, Manders, et al.
- 1977
|
|
3
|
On the di culty of nding small consistent decision trees
– Hancock
- 1989
|
|
2
|
A polynomial-time algorithm for learningk-variable pattern languages from examples
– Kearns, Pitt
- 1989
|
|
1
|
On the learnability of nite automata
– Li, Vazirani
- 1988
|
|
