Episciences.org
Not a member yet
6707 research outputs found
Sort by
Ainu–Japanese Bi-directional Neural Machine Translation: A Step Towards Linguistic Preservation of Ainu, An Under-Resourced Indigenous Language in Japan
No full textThis study presents a groundbreaking approach to preserving the Ainu language, recognized as critically endangered by UNESCO, by developing a bi-directional neural machine translation (MT) system between Ainu and Japanese. Utilizing the Marian MT framework, known for its effectiveness with resource-scarce languages, the research aims to overcome the linguistic complexities inherent in Ainu's polysynthetic structure. The paper delineates a comprehensive methodology encompassing data collection from diverse Ainu text sources, meticulous preprocessing, and the deployment of neural MT models, culminating in the achievement of significant SacreBLEU scores that underscore the models' translation accuracy. The findings illustrate the potential of advanced MT technology to facilitate linguistic preservation and educational endeavors, advocating for integrating such technologies in safeguarding endangered languages. This research not only underscores the critical role of MT in bridging language divides but also sets a precedent for employing computational linguistics to preserve cultural and linguistic heritage
Normalization of Arabic Dialects into Modern Standard Arabic using BERT and GPT-2
No full textWe present an encoder-decored based model for normalization of Arabic dialects using both BERT and GPT-2 based models. Arabic is a language of many dialects that not only differ from the Modern Standard Arabic (MSA) in terms of pronunciation but also in terms of morphology, grammar and lexical choice. This diversity can be troublesome even to a native Arabic speaker let alone a computer. Several NLP tools work well for MSA and in some of the main dialects but fail to cover Arabic language as a whole. Based on our manual evaluation, our model normalizes sentences entirely correctly 46\% of the time and almost correctly 26\% of the time
Singularities and growth of higher order discrete equations
No full textWe study the link between the degree growth of integrable birational mappingsof order higher than two and their singularity structures. The higher ordermappings we use in this study are all obtained by coupling mappings that areintegrable through spectral methods, typically belonging to the QRT family, toa variety of linearisable ones. We show that by judiciously choosing theselinearisable mappings, it is possible to obtain higher order mappings thatexhibit the maximal degree growth compatible with integrability, i.e. for whichthe degree grows as a polynomial of order equal to the order of the mapping. Inall the cases we analysed, we found that maximal degree growth was associatedwith the existence of an unconfining singularity pattern. Several cases withsubmaximal growth but which still possess unconfining singularity patterns arealso presented. In many cases the exact degrees of the iterates of the mappingswere obtained by applying a method due to Halburd, based on the preimages ofspecific values that appear in the singularity patterns of the mapping, but wealso present some examples where such a calculation appears to be impossible.Comment: 16 pages, no figures; final accepted versio
Two-fold degeneracy of a class of rational Painlev\'e V solutions
No full textWe present a construction of a class of rational solutions of the Painlev\'eV equation that exhibit a two-fold degeneracy, meaning that there exist twodistinct solutions that share identical parameters. The fundamental object of our study is the orbit of translation operators of affine Weyl group acting on the underlying seed solution thatonly allows action of some symmetry operations. By linking points on this orbitto rational solutions, we establish conditions for such degeneracy to occurafter involving in the construction additional B\"acklund transformations thatare inexpressible as translation operators. This approach enables us to deriveexplicit expressions for these degenerate solutions. An advantage of thisformalism is that it easily allows generalization to higher Painlev\'e systemsassociated with dressing chains of even period .Comment: 15 page
Résolution du Pigment Sequencing Problem avec lesalgorithmes génétiques (Submission to Episciences)
No full textSubmission to EpisciencesLot sizing is important in production planning. It consists of determining a production plan that meets the orders and other constraints while minimizing the production cost. Here, we consider a Discrete Lot Sizing and Scheduling Problem (DLSP), specifically the Pigment Sequencing Problem (PSP). We have developed a solution that uses Genetic Algorithms to tackle the PSP. Our approach introduces adaptive techniques for each step of the genetic algorithm, including initialization, selection, crossover, and mutation. We conducted a series of experiments to assess the performance of our approach across some multiple trials using publicly available instances of the PSP. Our experimental results demonstrate that Genetic Algorithms are practical and effective approaches for solving DLSP.Le dimensionnement de lots tient une place importante en planification de production en industrie. Il consiste à trouver un plan de production qui à la fois satisfait les demandes et autres contraintes tout en minimisant les coûts de production. Dans cet article, nous étudions une instance de problèmes de dimension discret (DLSP), le Pigment Sequencing Problem (PSP). Nous avons développé une approche basée sur les algorithmes génétiques afin de le résoudre. Notre approche propose des méthodes adaptatives pour chacune des étapes des algorithmes génétiques que sont l'initialisation, la sélection, le croisement et la mutation. Les expériences menées nous ont permis d'évaluer la performance de cette approche sur des instances en accès publique de PSP. Les résultats obtenus montrent que les algorithmes génétiques constituent une approche intéressante et effective dans la résolution des DLSP
Analyzing Robustness of Angluin's L Algorithm in Presence of Noise
No full textAngluin's L algorithm learns the minimal deterministic finite automaton(DFA) of a regular language using membership and equivalence queries. Itsprobabilistic approximatively correct (PAC) version substitutes an equivalencequery by numerous random membership queries to get a high level confidence tothe answer. Thus it can be applied to any kind of device and may be viewed asan algorithm for synthesizing an automaton abstracting the behavior of thedevice based on observations. Here we are interested on how Angluin's PAClearning algorithm behaves for devices which are obtained from a DFA byintroducing some noise. More precisely we study whether Angluin's algorithmreduces the noise and produces a DFA closer to the original one than the noisydevice. We propose several ways to introduce the noise: (1) the noisy deviceinverts the classification of words w.r.t. the DFA with a small probability,(2) the noisy device modifies with a small probability the letters of the wordbefore asking its classification w.r.t. the DFA, (3) the noisy device combinesthe classification of a word w.r.t. the DFA and its classification w.r.t. acounter automaton, and (4) the noisy DFA is obtained by a random process fromtwo DFA such that the language of the first one is included in the second one.Then when a word is accepted (resp. rejected) by the first (resp. second) one,it is also accepted (resp. rejected) and in the remaining cases, it is acceptedwith probability 0.5. Our main experimental contributions consist in showingthat: (1) Angluin's algorithm behaves well whenever the noisy device isproduced by a random process, (2) but poorly with a structured noise, and, that(3) is able to eliminate pathological behaviours specified in a regular way.Theoretically, we show that randomness almost surely yields systems withnon-recursively enumerable languages.Comment: arXiv admin note: substantial text overlap with arXiv:2209.1031
Improved quantum data analysis
No full textWe provide more sample-efficient versions of some basic routines in quantumdata analysis, along with simpler proofs. Particularly, we give a quantum"Threshold Search" algorithm that requires only samples of a -dimensional state . That is, given observables 0 \leA_1, A_2, ..., A_m \le 1 such that for atleast one , the algorithm finds with . As a consequence, we obtain a Shadow Tomography algorithmrequiring only samples, whichsimultaneously achieves the best known dependence on each parameter , ,. This yields the same sample complexity for quantum HypothesisSelection among states; we also give an alternative Hypothesis Selectionmethod using samples
On isomorphisms to a free group and beyond
No full textThe isomorphism problem for infinite finitely presented groups is probablythe hardest among standard algorithmic problems in group theory. Classes ofgroups where it has been completely solved are nilpotent groups, hyperbolicgroups, and limit groups. In this short paper, we address the problem ofisomorphism to particular groups, including free groups. We also address thealgorithmic problem of embedding a finitely presented group in a given limitgroup.Comment: 5 pages. Published in journal of Groups, Complexity, Cryptolog
Semiring Provenance for B\"uchi Games: Strategy Analysis with Absorptive Polynomials
No full textThis paper presents a case study for the application of semiring semanticsfor fixed-point formulae to the analysis of strategies in B\"uchi games.Semiring semantics generalizes the classical Boolean semantics by permittingmultiple truth values from certain semirings. Evaluating the fixed-pointformula that defines the winning region in a given game in an appropriatesemiring of polynomials provides not only the Boolean information on who wins,but also tells us how they win and which strategies they might use. This iswell-understood for reachability games, where the winning region is definableas a least fixed point. The case of B\"uchi games is of special interest, notonly due to their practical importance, but also because it is the simplestcase where the fixed-point definition involves a genuine alternation of agreatest and a least fixed point. We show that, in a precise sense, semiringsemantics provide information about all absorption-dominant strategies --strategies that win with minimal effort, and we discuss how these relate topositional and the more general persistent strategies. This information enablesapplications such as game synthesis or determining minimal modifications to thegame needed to change its outcome. Lastly, we discuss limitations of ourapproach and present questions that cannot be immediately answered by semiringsemantics