arXiv subjects

Luca Manzoni

Publications and source records attributed to Luca Manzoni.

32 records · Page 2Linked to original sources

arXiv2020

Tip the Balance: Improving Exploration of Balanced Crossover Operators by Adaptive Bias

The use of balanced crossover operators in Genetic Algorithms (GA) ensures that the binary strings generated as offsprings have the same Hamming weight of the parents, a constraint which is sought in certain discrete optimization problems. Although this method reduces the size of the search space, the resulting fitness landscape often becomes more difficult for the GA to explore and to discover optimal solutions. This issue has been studied in this paper by applying an adaptive bias strategy to a counter-based crossover operator that introduces unbalancedness in the offspring with a certain probability, which is decreased throughout the evolutionary process. Experiments show that improving the exploration of the search space with this adaptive bias strategy is beneficial for the GA performances in terms of the number of optimal solutions found for the balanced nonlinear Boolean functions problem.

cs.NE↗

arXiv2020

Towards an evolutionary-based approach for natural language processing

Tasks related to Natural Language Processing (NLP) have recently been the focus of a large research endeavor by the machine learning community. The increased interest in this area is mainly due to the success of deep learning methods. Genetic Programming (GP), however, was not under the spotlight with respect to NLP tasks. Here, we propose a first proof-of-concept that combines GP with the well established NLP tool word2vec for the next word prediction task. The main idea is that, once words have been moved into a vector space, traditional GP operators can successfully work on vectors, thus producing meaningful words as the output. To assess the suitability of this approach, we perform an experimental evaluation on a set of existing newspaper headlines. Individuals resulting from this (pre-)training phase can be employed as the initial population in other NLP tasks, like sentence generation, which will be the focus of future investigations, possibly employing adversarial co-evolutionary approaches.

cs.CL↗

arXiv2019

Balanced Crossover Operators in Genetic Algorithms

In several combinatorial optimization problems arising in cryptography and design theory, the admissible solutions must often satisfy a balancedness constraint, such as being represented by bitstrings with a fixed number of ones. For this reason, several works in the literature tackling these optimization problems with Genetic Algorithms (GA) introduced new balanced crossover operators which ensure that the offspring has the same balancedness characteristics of the parents. However, the use of such operators has never been thoroughly motivated, except for some generic considerations about search space reduction. In this paper, we undertake a rigorous statistical investigation on the effect of balanced and unbalanced crossover operators against three optimization problems from the area of cryptography and coding theory: nonlinear balanced Boolean functions, binary Orthogonal Arrays (OA) and bent functions. In particular, we consider three different balanced crossover operators (each with two variants: "left-to-right" and "shuffled"), two of which have never been published before, and compare their performances with classic one-point crossover. We are able to confirm that the balanced crossover operators performs better than all three balanced crossover operators. Furthermore, in two out of three crossovers, the "left-to-right" version performs better than the "shuffled" version.

cs.NE↗

arXiv2019

Search Space Reduction of Asynchrony Immune Cellular Automata by Center Permutivity

We continue the study of asynchrony immunity in cellular automata (CA), which can be considered as a weaker version of correlation immunity in the context of vectorial Boolean functions. The property could have applications as a countermeasure for side-channel attacks in CA-based cryptographic primitives, such as S-boxes and pseudorandom number generators. We first give some theoretical results on the necessary conditions that a CA rule must satisfy in order to meet asynchrony immunity, the most important one being center permutivity. Next, we perform an exhaustive search of all asynchrony immune CA rules of neighborhood size up to $5$, leveraging on the discovered theoretical properties to greatly reduce the size of the search space.

nlin.CG↗

arXiv2019

The many roads to the simulation of reaction systems

Reaction systems are a computational model inspired by the bio-chemical reactions that happen inside biological cells. They have been and currently are studied for their many nice theoretical properties. They are also a useful modeling tool for biochemical systems, but in order to be able to employ them effectively in the field the presence of efficient and widely available simulators is essential. Here we explore three different algorithms and implementations of the simulation, comparing them to the current state of the art. We also show that we can obtain performances comparable to GPU-based simulations on real-world systems by using a carefully tuned CPU-based simulator.

cs.FL↗

arXiv2019

Complexity of the dynamics of reaction systems

Reaction systems are discrete dynamical systems inspired by bio-chemical processes, whose dynamical behaviour is expressed by set-theoretic operations on finite sets. Reaction systems thus provide a description of bio-chemical phenomena that complements the more traditional approaches, for instance those based on differential equations. A comprehensive list of decision problems about the dynamical behavior of reaction systems (such as cycles and fixed/periodic points, attractors, and reachability) is provided along with the corresponding computational complexity, which ranges from tractable problems to PSPACE-complete problems.

cs.FL↗

arXiv2019

Characterizing PSPACE with shallow non-confluent P systems

In P systems with active membranes, the question of understanding the power of non-confluence within a polynomial time bound is still an open problem. It is known that, for shallow P systems, that is, with only one level of nesting, non-confluence allows them to solve conjecturally harder problems than confluent P systems, thus reaching PSPACE. Here we show that PSPACE is not only a bound, but actually an exact characterization. Therefore, the power endowed by non-confluence to shallow P systems is equal to the power gained by confluent P systems when non-elementary membrane division and polynomial depth are allowed, thus suggesting a connection between the roles of non-confluence and nesting depth.

cs.CC↗

arXiv2019

On the dynamical behaviour of linear higher-order cellular automata and its decidability

Higher-order cellular automata (HOCA) are a variant of cellular automata (CA) used in many applications (ranging, for instance, from the design of secret sharing schemes to data compression and image processing), and in which the global state of the system at time $t$ depends not only on the state at time $t-1$, as in the original model, but also on the states at time $t-2, \ldots, t-n$, where $n$ is the memory size of the HOCA. We provide decidable characterizations of two important dynamical properties, namely, sensitivity to the initial conditions and equicontinuity, for linear HOCA over the alphabet $\mathbb{Z}_m$. Such characterizations extend the ones shown in [23] for linear CA (LCA) over the alphabet $\mathbb{Z}^{n}_m$ in the case $n=1$. We also prove that linear HOCA of size memory $n$ over $\mathbb{Z}_m$ form a class that is indistinguishable from a specific subclass of LCA over $\mathbb{Z}_m^n$. This enables to decide injectivity and surjectivity for linear HOCA of size memory $n$ over $\mathbb{Z}_m$ using the decidable characterization provided in [2] and [19] for injectivity and surjectivity of LCA over $\mathbb{Z}^n_m$. Finally, we prove an equivalence between LCA over $\mathbb{Z}_m^n$ and an important class of non-uniform CA, another variant of CA used in many applications.

cs.FL↗

arXiv2019

Solving QSAT in sublinear depth

Among $\mathbf{PSPACE}$-complete problems, QSAT, or quantified SAT, is one of the most used to show that the class of problems solvable in polynomial time by families of a given variant of P systems includes the whole $\mathbf{PSPACE}$. However, most solutions require a membrane nesting depth that is linear with respect to the number of variables of the QSAT instance under consideration. While a system of a certain depth is needed, since depth 1 systems only allows to solve problems in $\mathbf{P^{\#P}}$, it was until now unclear if a linear depth was, in fact, necessary. Here we use P systems with active membranes with charges, and we provide a construction that proves that QSAT can be solved with a sublinear nesting depth of order $\frac{n}{\log n}$, where $n$ is the number of variables in the quantified formula given as input.

cs.CC↗

arXiv2019

A Turing machine simulation by P systems without charges

It is well known that the kind of P systems involved in the definition of the P conjecture is able to solve problems in the complexity class $\mathbf{P}$ by leveraging the uniformity condition. Here we show that these systems are indeed able to simulate deterministic Turing machines working in polynomial time with a weaker uniformity condition and using only one level of membrane nesting. This allows us to embed this construction into more complex membrane structures, possibly showing that constructions similar to the one performed in [1] for P systems with charges can be carried out also in this case.

cs.CC↗

arXiv2018

Pruning Techniques for Mixed Ensembles of Genetic Programming Models

The objective of this paper is to define an effective strategy for building an ensemble of Genetic Programming (GP) models. Ensemble methods are widely used in machine learning due to their features: they average out biases, they reduce the variance and they usually generalize better than single models. Despite these advantages, building ensemble of GP models is not a well-developed topic in the evolutionary computation community. To fill this gap, we propose a strategy that blends individuals produced by standard syntax-based GP and individuals produced by geometric semantic genetic programming, one of the newest semantics-based method developed in GP. In fact, recent literature showed that combining syntax and semantics could improve the generalization ability of a GP model. Additionally, to improve the diversity of the GP models used to build up the ensemble, we propose different pruning criteria that are based on correlation and entropy, a commonly used measure in information theory. Experimental results,obtained over different complex problems, suggest that the pruning criteria based on correlation and entropy could be effective in improving the generalization ability of the ensemble model and in reducing the computational burden required to build it.

cs.NE↗

arXiv2017

A Distance Between Populations for n-Points Crossover in Genetic Algorithms

Genetic algorithms (GAs) are an optimization technique that has been successfully used on many real-world problems. There exist different approaches to their theoretical study. In this paper we complete a recently presented approach to model one-point crossover using pretopologies (or Cech topologies) in two ways. First, we extend it to the case of n-points crossover. Then, we experimentally study how the distance distribution changes when the number of crossover points increases.

cs.NE↗

arXiv2012

An Efficient Genetic Programming System with Geometric Semantic Operators and its Application to Human Oral Bioavailability Prediction

Very recently new genetic operators, called geometric semantic operators, have been defined for genetic programming. Contrarily to standard genetic operators, which are uniquely based on the syntax of the individuals, these new operators are based on their semantics, meaning with it the set of input-output pairs on training data. Furthermore, these operators present the interesting property of inducing a unimodal fitness landscape for every problem that consists in finding a match between given input and output data (for instance regression and classification). Nevertheless, the current definition of these operators has a serious limitation: they impose an exponential growth in the size of the individuals in the population, so their use is impossible in practice. This paper is intended to overcome this limitation, presenting a new genetic programming system that implements geometric semantic operators in an extremely efficient way. To demonstrate the power of the proposed system, we use it to solve a complex real-life application in the field of pharmacokinetic: the prediction of the human oral bioavailability of potential new drugs. Besides the excellent performances on training data, which were expected because the fitness landscape is unimodal, we also report an excellent generalization ability of the proposed system, at least for the studied application. In fact, it outperforms standard genetic programming and a wide set of other well-known machine learning methods.

cs.NE↗

arXiv2011

Computational Aspects of Asynchronous CA

This work studies some aspects of the computational power of fully asynchronous cellular automata (ACA). We deal with some notions of simulation between ACA and Turing Machines. In particular, we characterize the updating sequences specifying which are "universal", i.e., allowing a (specific family of) ACA to simulate any TM on any input. We also consider the computational cost of such simulations.

cs.FL↗