Quantum Queries on Permutations

K. Iwama and R. Freivalds considered query algorithms where the black box contains a permutation. Since then several authors have compared quantum and deterministic query algorithms for permutations. It turns out that the case of \(n\)-permutations where \(n\) is an odd number is difficult. There was no example of a permutation problem where quantization can save half of the queries for \((2m+1)\)-permutations if \(m\ge 2\). Even for \((2m)\)-permutations with \(m\ge 2\), the best proved advantage of quantum query algorithms is the result by Iwama/Freivalds where the quantum query complexity is \(m\) but the deterministic query complexity is \((2m-1)\). We present a group of \(5\)-permutati…

Languages Recognizable by Quantum Finite Automata

There are several nonequivalent definitions of quantum finite automata. Nearly all of them recognize only regular languages but not all regular languages. On the other hand, for all these definitions there is a result showing that there is a language l such that the size of the quantum automaton recognizing L is essentially smaller than the size of the minimal deterministic automaton recognizing L. For most of the definitions of quantum finite automata the problem to describe the class of the languages recognizable by the quantum automata is still open. The partial results are surveyed in this paper. Moreover, for the most popular definition of the QFA, the class of languages recognizable b…

Boolean Functions with a Low Polynomial Degree and Quantum Query Algorithms

The complexity of quantum query algorithms computing Boolean functions is strongly related to the degree of the algebraic polynomial representing this Boolean function. There are two related difficult open problems. First, Boolean functions are sought for which the complexity of exact quantum query algorithms is essentially less than the complexity of deterministic query algorithms for the same function. Second, Boolean functions are sought for which the degree of the representing polynomial is essentially less than the complexity of deterministic query algorithms. We present in this paper new techniques to solve the second problem.

On the impact of forgetting on learning machines

People tend not to have perfect memories when it comes to learning, or to anything else for that matter. Most formal studies of learning, however, assume a perfect memory. Some approaches have restricted the number of items that could be retained. We introduce a complexity theoretic accounting of memory utilization by learning machines. In our new model, memory is measured in bits as a function of the size of the input. There is a hierarchy of learnability based on increasing memory allotment. The lower bound results are proved using an unusual combination of pumping and mutual recursion theorem arguments. For technical reasons, it was necessary to consider two types of memory : long and sh…

On the Size Complexity of Deterministic Frequency Automata

Austinat, Diekert, Hertrampf, and Petersen [2] proved that every language L that is (m,n)-recognizable by a deterministic frequency automaton such that m > n/2 can be recognized by a deterministic finite automaton as well. First, the size of deterministic frequency automata and of deterministic finite automata recognizing the same language is compared. Then approximations of a language are considered, where a language L′ is called an approximation of a language L if L′ differs from L in only a finite number of strings. We prove that if a deterministic frequency automaton has k states and (m,n)-recognizes a language L, where m > n/2, then there is a language L′ approximating L such that L′ c…

NON-CONSTRUCTIVE METHODS FOR FINITE PROBABILISTIC AUTOMATA

Size (the number of states) of finite probabilistic automata with an isolated cut-point can be exponentially smaller than the size of any equivalent finite deterministic automaton. However, the proof is non-constructive. The result is presented in two versions. The first version depends on Artin's Conjecture (1927) in Number Theory. The second version does not depend on conjectures not proved but the numerical estimates are worse. In both versions the method of the proof does not allow an explicit description of the languages used. Since our finite probabilistic automata are reversible, these results imply a similar result for quantum finite automata.

Group Input Machine

We introduce a new type of internal memory for finite automata and real-time automata. Instead of using tapes with a prescribed Euclidean structure (one-dimensional or two-dimensional tapes) we allow arbitrary group structure of the internal memory of the automata.

Ultrametric Vs. Quantum Query Algorithms

Ultrametric algorithms are similar to probabilistic algorithms but they describe the degree of indeterminism by p-adic numbers instead of real numbers. This paper introduces the notion of ultrametric query algorithms and shows an example of advantages of ultrametric query algorithms over deterministic, probabilistic and quantum query algorithms.

Postselection Finite Quantum Automata

Postselection for quantum computing devices was introduced by S. Aaronson[2] as an excitingly efficient tool to solve long standing problems of computational complexity related to classical computing devices only. This was a surprising usage of notions of quantum computation. We introduce Aaronson's type postselection in quantum finite automata. There are several nonequivalent definitions of quantumfinite automata. Nearly all of them recognize only regular languages but not all regular languages. We prove that PALINDROMES can be recognized by MM-quantum finite automata with postselection. At first we prove by a direct construction that the complement of this language can be recognized this …

Algorithmic Information Theory and Computational Complexity

We present examples where theorems on complexity of computation are proved using methods in algorithmic information theory. The first example is a non-effective construction of a language for which the size of any deterministic finite automaton exceeds the size of a probabilistic finite automaton with a bounded error exponentially. The second example refers to frequency computation. Frequency computation was introduced by Rose and McNaughton in early sixties and developed by Trakhtenbrot, Kinber, Degtev, Wechsung, Hinrichs and others. A transducer is a finite-state automaton with an input and an output. We consider the possibilities of probabilistic and frequency transducers and prove sever…

Amount of Nonconstructivity in Finite Automata

When D. Hilbert used nonconstructive methods in his famous paper on invariants (1888), P.Gordan tried to prevent the publication of this paper considering these methods as non-mathematical. L. E. J. Brouwer in the early twentieth century initiated intuitionist movement in mathematics. His slogan was "nonconstructive arguments have no value for mathematics". However, P. Erdos got many exciting results in discrete mathematics by nonconstructive methods. It is widely believed that these results either cannot be proved by constructive methods or the proofs would have been prohibitively complicated. R.Freivalds [7] showed that nonconstructive methods in coding theory are related to the notion of…

Structured Frequency Algorithms

B.A. Trakhtenbrot proved that in frequency computability (introduced by G. Rose) it is crucially important whether the frequency exceeds \(\frac{1}{2}\). If it does then only recursive sets are frequency-computable. If the frequency does not exceed \(\frac{1}{2}\) then a continuum of sets is frequency-computable. Similar results for finite automata were proved by E.B. Kinber and H. Austinat et al. We generalize the notion of frequency computability demanding a specific structure for the correct answers. We show that if this structure is described in terms of finite projective planes then even a frequency \(O(\frac{\sqrt{n}}{n})\) ensures recursivity of the computable set. We also show that …

Multiple Usage of Random Bits in Finite Automata

Finite automata with random bits written on a separate 2-way readable tape can recognize languages not recognizable by probabilistic finite automata. This shows that repeated reading of random bits by finite automata can have big advantages over one-time reading of random bits.

Learning with belief levels

AbstractWe study learning of predicate logics formulas from “elementary facts,” i.e. from the values of the predicates in the given model. Several models of learning are considered, but most of our attention is paid to learning with belief levels. We propose an axiom system which describes what we consider to be a human scientist's natural behavior when trying to explore these elementary facts. It is proved that no such system can be complete. However we believe that our axiom system is “practically” complete. Theorems presented in the paper in some sense confirm our hypothesis.

On the inductive inference of recursive real-valued functions

AbstractWe combine traditional studies of inductive inference and classical continuous mathematics to produce a study of learning real-valued functions. We consider two possible ways to model the learning by example of functions with domain and range the real numbers. The first approach considers functions as represented by computable analytic functions. The second considers arbitrary computable functions of recursive real numbers. In each case we find natural examples of learnable classes of functions and unlearnable classes of functions.

Improved constructions of mixed state quantum automata

Quantum finite automata with mixed states are proved to be super-exponentially more concise rather than quantum finite automata with pure states. It was proved earlier by A. Ambainis and R. Freivalds that quantum finite automata with pure states can have an exponentially smaller number of states than deterministic finite automata recognizing the same language. There was an unpublished ''folk theorem'' proving that quantum finite automata with mixed states are no more super-exponentially more concise than deterministic finite automata. It was not known whether the super-exponential advantage of quantum automata is really achievable. We prove that there is an infinite sequence of distinct int…

Size of Quantum Finite State Transducers

Sizes of quantum and deterministic finite state transducers are compared in the case when both quantum and deterministic finite state transducers exist. The difference in size may be exponential.

On the Influence of Technology on Learning Processes

Probabilistic computations and frequency computations were invented for the same purpose, namely, to study possible advantages of technology involving random choices. Recently several authors have discovered close relationships of these generalizations of deterministic computations to computations taking advice. Various forms of computation taking advice were studied by Karp and Lipton [1], Damm and Holzer [2], and Freivalds [3]. In the present paper, we apply the nonconstructive, probabilistic, and frequency methods to an inductive inference paradigm originally due to Gold [4] and investigate their impact on the resulting learning models. Several trade-offs with respect to the resulting l…

Choosing a learning team

Super-Exponential Size Advantage of Quantum Finite Automata with Mixed States

Quantum finite automata with mixed states are proved to be super-exponentially more concise rather than quantum finite automata with pure states. It was proved earlier by A.Ambainis and R.Freivalds that quantum finite automata with pure states can have exponentially smaller number of states than deterministic finite automata recognizing the same language. There was a never published "folk theorem" proving that quantum finite automata with mixed states are no more than super-exponentially more concise than deterministic finite automata. It was not known whether the super-exponential advantage of quantum automata is really achievable. We use a novel proof technique based on Kolmogorov complex…

On the role of procrastination for machine learning

Quantum Queries on Permutations with a Promise

This paper studies quantum query complexities for deciding (exactly or with probability 1.0) the parity of permutations of n numbers, 0 through n *** 1. Our results show quantum mechanism is quite strong for this non-Boolean problem as it is for several Boolean problems: (i) For n = 3, we need a single query in the quantum case whereas we obviously need two queries deterministically. (ii) For even n , n /2 quantum queries are sufficient whereas we need n *** 1 queries deterministically. (iii) Our third result is for the problem deciding whether the given permutation is the identical one. For this problem, we show that there is a nontrivial promise such that if we impose that promise to the …

Ultrametric Algorithms and Automata

We introduce a notion of ultrametric automata and Turing machines using p-adic numbers to describe random branching of the process of computation. These automata have properties similar to the properties of probabilistic automata but complexity of probabilistic automata and complexity of ultrametric automata can differ very much.

Computer graphics for quantum computation

Finite State Transducers with Intuition

Finite automata that take advice have been studied from the point of view of what is the amount of advice needed to recognize nonregular languages. It turns out that there can be at least two different types of advice. In this paper we concentrate on cases when the given advice contains zero information about the input word and the language to be recognized. Nonetheless some nonregular languages can be recognized in this way. The help-word is merely a sufficiently long word with nearly maximum Kolmogorov complexity. Moreover, any sufficiently long word with nearly maximum Kolmogorov complexity can serve as a help-word. Finite automata with such help can recognize languages not recognizable …

Co-learning of total recursive functions

Artin’s Conjecture and Size of Finite Probabilistic Automata

Size (the number of states) of finite probabilistic automata with an isolated cut-point can be exponentially smaller than the size of any equivalent finite deterministic automaton. The result is presented in two versions. The first version depends on Artin's Conjecture (1927) in Number Theory. The second version does not depend on conjectures but the numerical estimates are worse. In both versions the method of the proof does not allow an explicit description of the languages used. Since our finite probabilistic automata are reversible, these results imply a similar result for quantum finite automata.

On the relative sizes of learnable sets

Abstract Measure and category (or rather, their recursion-theoretical counterparts) have been used in theoretical computer science to make precise the intuitive notion “for most of the recursive sets”. We use the notions of effective measure and category to discuss the relative sizes of inferrible sets, and their complements. We find that inferable sets become large rather quickly in the standard hierarchies of learnability. On the other hand, the complements of the learnable sets are all large.

A new family of nonstochastic languages

Öz bulunamadı.

Non-constructive Methods for Finite Probabilistic Automata

Size (the number of states) of finite probabilistic automata with an isolated cut-point can be exponentially smaller than the size of any equivalent finite deterministic automaton. The result is presented in two versions. The first version depends on Artin's Conjecture (1927) in Number Theory. The second version does not depend on conjectures but the numerical estimates are worse. In both versions the method of the proof does not allow an explicit description of the languages used. Since our finite probabilistic automata are reversible, these results imply a similar result for quantum finite automata.

Ultrametric Finite Automata and Turing Machines

We introduce a notion of ultrametric automata and Turing machines using p-adic numbers to describe random branching of the process of computation. These automata have properties similar to the properties of probabilistic automata but complexity of probabilistic automata and complexity of ultrametric automata can differ very much.

Active Learning of Recursive Functions by Ultrametric Algorithms

We study active learning of classes of recursive functions by asking value queries about the target function f, where f is from the target class. That is, the query is a natural number x, and the answer to the query is f(x). The complexity measure in this paper is the worst-case number of queries asked. We prove that for some classes of recursive functions ultrametric active learning algorithms can achieve the learning goal by asking significantly fewer queries than deterministic, probabilistic, and even nondeterministic active learning algorithms. This is the first ever example of a problem where ultrametric algorithms have advantages over nondeterministic algorithms.