Results 1  10
of
103
COMPUTATIONALLY SOUND PROOFS
, 2000
"... This paper puts forward a new notion of a proof based on computational complexity and explores its implications for computation at large. Computationally sound proofs provide, in a novel and meaningful framework, answers to old and new questions in complexity theory. In particular, given a random o ..."
Abstract

Cited by 92 (3 self)
 Add to MetaCart
This paper puts forward a new notion of a proof based on computational complexity and explores its implications for computation at large. Computationally sound proofs provide, in a novel and meaningful framework, answers to old and new questions in complexity theory. In particular, given a random oracle or a new complexity assumption, they enable us to 1. prove that verifying is easier than deciding for all theorems; 2. provide a quite effective way to prove membership in computationally hard languages (such as CoNPcomplete ones); and 3. show that every computation possesses a short certificate vouching its correctness. Finally, if a special type of computationally sound proof exists, we show that Blum’s notion of program checking can be meaningfully broadened so as to prove that NPcomplete languages are checkable.
Optimal Ordered Problem Solver
, 2002
"... We present a novel, general, optimally fast, incremental way of searching for a universal algorithm that solves each task in a sequence of tasks. The Optimal Ordered Problem Solver (OOPS) continually organizes and exploits previously found solutions to earlier tasks, eciently searching not only the ..."
Abstract

Cited by 62 (20 self)
 Add to MetaCart
We present a novel, general, optimally fast, incremental way of searching for a universal algorithm that solves each task in a sequence of tasks. The Optimal Ordered Problem Solver (OOPS) continually organizes and exploits previously found solutions to earlier tasks, eciently searching not only the space of domainspecific algorithms, but also the space of search algorithms. Essentially we extend the principles of optimal nonincremental universal search to build an incremental universal learner that is able to improve itself through experience.
Shifting Inductive Bias with SuccessStory Algorithm, Adaptive Levin Search, and Incremental SelfImprovement
 MACHINE LEARNING
, 1997
"... We study task sequences that allow for speeding up the learner's average reward intake through appropriate shifts of inductive bias (changes of the learner's policy). To evaluate longterm effects of bias shifts setting the stage for later bias shifts we use the "successstory algorithm" (SSA). SSA ..."
Abstract

Cited by 62 (30 self)
 Add to MetaCart
We study task sequences that allow for speeding up the learner's average reward intake through appropriate shifts of inductive bias (changes of the learner's policy). To evaluate longterm effects of bias shifts setting the stage for later bias shifts we use the "successstory algorithm" (SSA). SSA is occasionally called at times that may depend on the policy itself. It uses backtracking to undo those bias shifts that have not been empirically observed to trigger longterm reward accelerations (measured up until the current SSA call). Bias shifts that survive SSA represent a lifelong success history. Until the next SSA call, they are considered useful and build the basis for additional bias shifts. SSA allows for plugging in a wide variety of learning algorithms. We plug in (1) a novel, adaptive extension of Levin search and (2) a method for embedding the learner's policy modification strategy within the policy itself (incremental selfimprovement). Our inductive transfer case studies...
The History and Status of the P versus NP Question
, 1992
"... this article, I have attempted to organize and describe this literature, including an occasional opinion about the most fruitful directions, but no technical details. In the first half of this century, work on the power of formal systems led to the formalization of the notion of algorithm and the re ..."
Abstract

Cited by 51 (0 self)
 Add to MetaCart
this article, I have attempted to organize and describe this literature, including an occasional opinion about the most fruitful directions, but no technical details. In the first half of this century, work on the power of formal systems led to the formalization of the notion of algorithm and the realization that certain problems are algorithmically unsolvable. At around this time, forerunners of the programmable computing machine were beginning to appear. As mathematicians contemplated the practical capabilities and limitations of such devices, computational complexity theory emerged from the theory of algorithmic unsolvability. Early on, a particular type of computational task became evident, where one is seeking an object which lies
Discovering Neural Nets With Low Kolmogorov Complexity And High Generalization Capability
 Neural Networks
, 1997
"... Many neural net learning algorithms aim at finding "simple" nets to explain training data. The expectation is: the "simpler" the networks, the better the generalization on test data (! Occam's razor). Previous implementations, however, use measures for "simplicity" that lack the power, universali ..."
Abstract

Cited by 49 (30 self)
 Add to MetaCart
Many neural net learning algorithms aim at finding "simple" nets to explain training data. The expectation is: the "simpler" the networks, the better the generalization on test data (! Occam's razor). Previous implementations, however, use measures for "simplicity" that lack the power, universality and elegance of those based on Kolmogorov complexity and Solomonoff's algorithmic probability. Likewise, most previous approaches (especially those of the "Bayesian" kind) suffer from the problem of choosing appropriate priors. This paper addresses both issues. It first reviews some basic concepts of algorithmic complexity theory relevant to machine learning, and how the SolomonoffLevin distribution (or universal prior) deals with the prior problem. The universal prior leads to a probabilistic method for finding "algorithmically simple" problem solutions with high generalization capability. The method is based on Levin complexity (a timebounded generalization of Kolmogorov comple...
The Complexity and Distribution of Hard Problems
 SIAM JOURNAL ON COMPUTING
, 1993
"... Measuretheoretic aspects of the P m reducibility structure of the exponential time complexity classes E=DTIME(2 linear ) and E 2 = DTIME(2 polynomial ) are investigated. Particular attention is given to the complexity (measured by the size of complexity cores) and distribution (abundance in ..."
Abstract

Cited by 46 (16 self)
 Add to MetaCart
Measuretheoretic aspects of the P m reducibility structure of the exponential time complexity classes E=DTIME(2 linear ) and E 2 = DTIME(2 polynomial ) are investigated. Particular attention is given to the complexity (measured by the size of complexity cores) and distribution (abundance in the sense of measure) of languages that are P m  hard for E and other complexity classes. Tight upper and lower bounds on the size of complexity cores of hard languages are derived. The upper bound says that the P m hard languages for E are unusually simple, in the sense that they have smaller complexity cores than most languages in E. It follows that the P m complete languages for E form a measure 0 subset of E (and similarly in E 2 ). This latter fact is seen to be a special case of a more general theorem, namely, that every P m degree (e.g., the degree of all P m complete languages for NP) has measure 0 in E and in E 2 .
Universal intelligence: A definition of machine intelligence
 Minds and Machines
, 2007
"... A fundamental problem in artificial intelligence is that nobody really knows what intelligence is. The problem is especially acute when we need to consider artificial systems which are significantly different to humans. In this paper we approach this problem in the following way: We take a number of ..."
Abstract

Cited by 42 (11 self)
 Add to MetaCart
A fundamental problem in artificial intelligence is that nobody really knows what intelligence is. The problem is especially acute when we need to consider artificial systems which are significantly different to humans. In this paper we approach this problem in the following way: We take a number of well known informal definitions of human intelligence that have been given by experts, and extract their essential features. These are then mathematically formalised to produce a general measure of intelligence for arbitrary machines. We believe that this equation formally captures the concept of machine intelligence in the broadest reasonable sense. We then show how this formal definition is related to the theory of universal optimal learning agents. Finally, we survey the many other tests and definitions of intelligence that have been proposed for machines.
Discovering Solutions with Low Kolmogorov Complexity and High Generalization Capability
, 1995
"... Many machine learning algorithms aim at finding "simple" rules to explain training data. The expectation is: the "simpler" the rules, the better the generalization on test data ( Occam's razor). Most practi cal implementations, however, use measures for "simplicity" that lack the power, univ ..."
Abstract

Cited by 37 (25 self)
 Add to MetaCart
Many machine learning algorithms aim at finding "simple" rules to explain training data. The expectation is: the "simpler" the rules, the better the generalization on test data ( Occam's razor). Most practi cal implementations, however, use measures for "simplicity" that lack the power, universality and elegance of those based on Kolmogorov complexity and Solomonoff's algorithmic probability. Likewise, most pre vious approaches (especially those of the "Bayesian" kind) suffer from the problem of choosing appropriate priors. This paper ad dresses both issues. It first reviews some ba sic concepts of algorithmic complexity theory relevant to machine learning, and how the SolomonoffLevin distribution (or universal prior) deals with the prior problem. The uni versal prior leads to a probabilistic method for finding "algorithmically simple" problem solutions with high generalization capability.
The Fastest And Shortest Algorithm For All WellDefined Problems
, 2002
"... An algorithm M is described that solves any welldefined problem p as quickly as the fastest algorithm computing a solution to p, save for a factor of 5 and loworder additive terms. M optimally distributes resources between the execution of provably correct psolving programs and an enumeration of ..."
Abstract

Cited by 35 (7 self)
 Add to MetaCart
An algorithm M is described that solves any welldefined problem p as quickly as the fastest algorithm computing a solution to p, save for a factor of 5 and loworder additive terms. M optimally distributes resources between the execution of provably correct psolving programs and an enumeration of all proofs, including relevant proofs of program correctness and of time bounds on program runtimes. M avoids Blum's speedup theorem by ignoring programs without correctness proof. M has broader applicability and can be faster than Levin's universal search, the fastest method for inverting functions save for a large multiplicative constant. An extension of Kolmogorov complexity and two novel natural measures of function complexity are used to show that the most efficient program computing some function f is also among the shortest programs provably computing f.
Reinforcement Learning With SelfModifying Policies
 IN S. THRUN , L. PRATT (EDS.), LEARNING TO LEARN
, 1997
"... A learner's modifiable components are called its policy. An algorithm that modifies the policy is a learning algorithm. If the learning algorithm has modifiable components represented as part of the policy, then we speak of a selfmodifying policy (SMP). SMPs can modify the way they modify themselve ..."
Abstract

Cited by 34 (22 self)
 Add to MetaCart
A learner's modifiable components are called its policy. An algorithm that modifies the policy is a learning algorithm. If the learning algorithm has modifiable components represented as part of the policy, then we speak of a selfmodifying policy (SMP). SMPs can modify the way they modify themselves etc. They are of interest in situations where the initial learning algorithm itself can be improved by experience  this is what we call "learning to learn". How can we force some (stochastic) SMP to trigger better and better selfmodifications? The successstory algorithm (SSA) addresses this question in a lifelong reinforcement learning context. During the learner's lifetime, SSA is occasionally called at times computed according to SMP itself. SSA uses backtracking to undo those SMPgenerated SMPmodifications that have not been empirically observed to trigger lifelong reward accelerations (measured up until the current SSA call  this evaluates the longterm effects of SMPmodifi...