arXiv subjects
Publications and source records attributed to Jean-Camille Birget.
This memorial article for Mark Sapir provides a brief overview of his life and career. Among his many contributions we highlight two of his most celebrated achievements: his groundbreaking solutions to Burnside-type problems for semigroups and his innovative construction of S-machines. Additionally, reflections from his colleagues and friends offer a heartfelt tribute, blending professional insights with personal memories.
We construct a wavelet-based almost sure uniform approximation of fractional Brownian motion (fBm) B_t^(H), t in [0, 1], of Hurst index H in (0, 1). Our results show that by Haar wavelets which merely have one vanishing moment, an almost sure uniform expansion of fBm of H in (0, 1) can be established. The convergence rate of our approximation is derived. We also describe a parallel algorithm that generates sample paths of an fBm efficiently.
We reprove a result of Boppana and Lagarias: If Pi_2^P is different from Sigma_2^P then there exists a partial function f that is computable by a polynomial-size family of circuits, but no inverse of f is computable by a polynomial-size family of circuits. We strengthen this result by showing that there exist length-preserving total functions that are one-way by circuit size and that are computable in uniform polynomial time. We also prove, if Pi_2^P is different from Sigma_2^P, that there exist polynomially balanced total surjective functions that are one-way by circuit size; here non-uniformity is used.
A slight modification of the definition of the Thompson-Higman groups G_k1 and F_k1 leads to inverse monoids that map onto G_k1 (respectively F_k1), and that have interesting properties: they are finitely generated, and residually finite. These inverse monoids are closely related to the suffix expansion of G_k1 (respectively F_k1).
The Bernoulli measure on strings is used to define height functions for the dense R- and L-orders of the Thompson-Higman monoids M_{k,1}. The measure can also be used to characterize the D-relation of certain submonoids of M_{k,1}. The computational complexity of computing the Bernoulli measure of certain sets, and in particular, of computing the R- and L-height of an element of M_{k,1} is investigated.
The Thompson-Higman groups G_{k,i} have a natural generalization to monoids M_{k,i}, and inverse monoids Inv_{k,i}. We study some structural features of M_{k,i} and Inv_{k,i} and investigate the computational complexity of decision problems. The main interest of these monoids is their close connection with circuits and circuit complexity. The maximal subgroups of M_{k,1} are isomorphic to the groups G_{k,j} (1 \leq j \leq k-1); so we rediscover all the Thompson-Higman groups within M_{k,1}. The Green relations \leq_J and \equiv_D of M_{k,1} can be decided in deterministic polynomial time when the inputs are words over a finite generating set of M_{k,1}. When a circuit-like generating set is used for M_{k,1} then deciding \leq_J is coDP-complete. The multiplier search problem for \leq_J is xNPsearch-complete, whereas the multiplier search problems of \leq_R and \leq_L are not in xNPsearch unless NP = coNP. Deciding \equiv_D for M_{k,1} when the inputs are words over a circuit-like generating set, is \oplus_{k-1}.NP-complete. For Inv_{k,1} over a circuit-like generating set, deciding \equiv_D is \oplus_{k-1} P-complete.
We study the monoid generalization M_{k,1} of the Thompson-Higman groups, and we characterize the R- and the L-preorder of M_{k,1}. Although M_{k,1} has only one non-zero J-class and k-1 non-zero D-classes, the R- and the L-preorder are complicated; in particular, <_R is dense (even within an L-class), and <_L is dense (even within an R-class). We study the computational complexity of the R- and the L-preorder. When inputs are given by words over a finite generating set of M_{k,1}, the R- and the L-preorder decision problems are in P. The main result of the paper is that over a "circuit-like" generating set, the R-preorder decision problem of M_{k,1} is Pi_2^P-complete, whereas the L-preorder decision problem is coNP-complete. We also prove related results about circuits: For combinational circuits, the surjectiveness problem is Pi_2^P-complete, whereas the injectiveness problem is coNP-complete.
Computational asymmetry, i.e., the discrepancy between the complexity of transformations and the complexity of their inverses, is at the core of one-way transformations. We introduce a computational asymmetry function that measures the amount of one-wayness of permutations. We also introduce the word-length asymmetry function for groups, which is an algebraic analogue of computational asymmetry. We relate boolean circuits to words in a Thompson monoid, over a fixed generating set, in such a way that circuit size is equal to word-length. Moreover, boolean circuits have a representation in terms of elements of a Thompson group, in such a way that circuit size is polynomially equivalent to word-length. We show that circuits built with gates that are not constrained to have fixed-length inputs and outputs, are at most quadratically more compact than circuits built from traditional gates (with fixed-length inputs and outputs). Finally, we show that the computational asymmetry function is closely related to certain distortion functions: The computational asymmetry function is polynomially equivalent to the distortion of the path length in Schreier graphs of certain Thompson groups, compared to the path length in Cayley graphs of certain Thompson monoids. We also show that the results of Razborov and others on monotone circuit complexity lead to exponential lower bounds on certain distortions.
We construct group codes over two letters (i.e., bases of subgroups of a two-generated free group) with special properties. Such group codes can be used for reducing algorithmic problems over large alphabets to algorithmic problems over a two-letter alphabet. Our group codes preserve aperiodicity of inverse finite automata. As an application we show that the following problems are PSpace-complete for two-letter alphabets (this was previously known for large enough finite alphabets): The intersection-emptiness problem for inverse finite automata, the aperiodicity problem for inverse finite automata, and the closure-under-radical problem for finitely generated subgroups of a free group. The membership problem for 3-generated inverse monoids is PSpace-complete.
We consider the subgroup lpG_{k,1} of length preserving elements of the Thompson-Higman group G_{k,1} and we show that all elements of G_{k,1} have a unique lpG_{k,1}.F_{k,1} factorization. This applies to the Thompson-Higman group T_{k,1} as well. We show that lpG_{k,1} is a ``diagonal'' direct limit of finite symmetric groups, and that lpT_{k,1} is a k^infinity Pr"ufer group. We find an infinite generating set of lpG_{k,1} which is related to reversible boolean circuits. We further investigate connections between the Thompson-Higman groups, circuits, and complexity. We show that elements of F_{k,1} cannot be one-way functions. We show that describing an element of G_{k,1} by a generalized bijective circuit is equivalent to describing the element by a word over a certain infinite generating set of G_{k,1}; word length over these generators is equivalent to generalized bijective circuit size. We give some coNP-completeness results for G_{k,1} (e.g., the word problem when elements are given by circuits), and #P-completeness results (e.g., finding the lpG_{k,1}.F_{k,1} factorization of an element of G_{k,1} given by a circuit).
We prove new results about the remarkable infinite simple groups introduced by Richard Thompson in the 1960s. We define the groups as partial transformation groups and we give a faithful representation in the Cuntz C*-algebra. For the finitely presented simple group T_fin (also known as V) we show that the word-length and the table size satisfy an n log n relation, just like the symmetric groups. We show that the word problem of T_fin belongs to the parallel complexity class AC^1 (a subclass of plynomial time). We show that the generalized word problem of T_fin is undecidable. We study the distortion functions of T_fin and we show that T_fin contains all finite direct products of finitely generated free groups as subgroups with linear distortion. As a consequence, up to polynomial equivalence of functions, the following three sets are the same: the set of distortions of T_fin, the set of all Dehn functions of finitely presented groups, and the set of time complexity functions of nondeterministic Turing machines.
We construct a finitely presented group with coNP-complete word problem, and a finitely generated simple group with coNP-complete word problem. These groups are represented as Thompson groups, hence as partial transformation groups of strings. The proof provides a simulation of combinational circuits by elements of the Thompson-Higman group G_{3,1}.
We extend a result of Goldreich and Ron about estimating the collision probability of a hash function. Their estimate has a polynomial tail. We prove that when the load factor is greater than a certain constant, the estimator has a gaussian tail. As an application we find an estimate of an upper bound for the average search time in hashing with chaining, for a particular user (we allow the overall key distribution to be different from the key distribution of a particular user). The estimator has a gaussian tail.
We give some connections between various functions defined on finitely presented groups (isoperimetric, isodiametric, Todd-Coxeter radius, filling length functions, etc.), and we study the relation between those functions and the computational complexity of the word problem (deterministic time, nondeterministic time, symmetric space). We show that the isoperimetric function can always be linearly decreased (unless it is the identity map). We present a new proof of the Double Exponential Inequality, based on context-free languages.
The (.)_reg construction was introduced in order to make an arbitrary semigroup S divide a regular semigroup (S)_reg which shares some important properties with S (e.g., finiteness, subgroups, torsion bounds, J-order structure). We show that (S)_reg can be described by a rather simple complete string rewrite system, as a consequence of which we obtain a new proof of the normal form theorem for (S)_reg. The new proof of the normal form theorem is conceptually simpler than the previous proofs.
We analyse the wavelet shrinkage algorithm of Donoho and Johnstone in order to assess the quality of the reconstruction of a signal obtained from noisy samples. We prove deviation bounds for the maximum of the squares of the error, and for the average of the squares of the error, under the assumption that the signal comes from a H"older class, and the noise samples are independent, of 0 mean, and bounded. Our main technique is Talgrand's isoperimetric theorem. Our bounds refine the known expectations for the average of the squares of the error.
This is the first of two papers devoted to connections between asymptotic functions of groups and computational complexity. One of the main results of this paper states that if for every $m$ the first $m$ digits of a real number $α\ge 4$ are computable in time $\le C2^{2^{Cm}}$ for some constant $C>0$ then $n^α$ is equivalent (``big O'') to the Dehn function of a finitely presented group. The smallest isodiametric function of this group is $n^{3/4α}$. On the other hand if $n^α$ is equivalent to the Dehn function of a finitely presented group then the first $m$ digits of $α$ are computable in time $\le C2^{2^{2^{Cm}}}$ for some constant $C$. This implies that, say, functions $n^{π+1}$, $n^{e^2}$ and $n^α$ for all rational numbers $α\ge 4$ are equivalent to the Dehn functions of some finitely presented group and that $n^π$ and $n^α$ for all rational numbers $α\ge 3$ are equivalent to the smallest isodiametric functions of finitely presented groups. Moreover we describe all Dehn functions of finitely presented groups $\succ n^4$ as time functions of Turing machines modulo two conjectures: \begin{enumerate} \item Every Dehn function is equivalent to a superadditive function. \item The square root of the time function of a Turing machine is equivalent to the time function of a Turing machine. \end{enumerate}