Explore chapters and articles related to this topic
Classification, Rule Generation and Evaluation using Modular Rough-fuzzy MLP
Published in Sankar K. Pal, Pabitra Mitra, Pattern Recognition Algorithms for Data Mining, 2004
Inductive logic programming (ILP) [179] is a machine learning technique used for construction of first-order clausal theories from examples and background knowledge. The aim is to discover, from a given set of preclassified examples, a set of classification rules with high predictive power. The PRO-GOL [180] and FOIL [233] classification algorithms, based on this method, were successfully applied in many domains. However, a limitation of these algorithms is their high computational complexity. Recently, several ILP-based scalable rule induction algorithms, e.g., TILDE [26] and GOLEM [181], are developed.
Six Human-Centered Artificial Intelligence Grand Challenges
Published in International Journal of Human–Computer Interaction, 2023
Ozlem Ozmen Garibay, Brent Winslow, Salvatore Andolina, Margherita Antona, Anja Bodenschatz, Constantinos Coursaris, Gregory Falco, Stephen M. Fiore, Ivan Garibay, Keri Grieman, John C. Havens, Marina Jirotka, Hernisa Kacorri, Waldemar Karwowski, Joe Kider, Joseph Konstan, Sean Koon, Monica Lopez-Gonzalez, Iliana Maifeld-Carucci, Sean McGregor, Gavriel Salvendy, Ben Shneiderman, Constantine Stephanidis, Christina Strobel, Carolyn Ten Holter, Wei Xu
One example of legislation currently in draft is the EU Artificial Intelligence Act (European Commission, 2021 from the European Commission. Article 3 of the AI Act defines an AI system as “software that is developed with one or more of the techniques that can, for a given set of human defined objectives, generate outputs such as content, predictions, recommendations, or decisions influencing the environments they interact with.” The techniques identified include: (a) ML approaches, including supervised, unsupervised, and reinforcement learning, using a wide variety of methods including deep learning; (b) Logic- and knowledge-based approaches, including knowledge representation, inductive (logic) programming, knowledge bases, inference, and deductive engines, (symbolic) reasoning and expert systems; and (c) Statistical approaches, Bayesian estimation, search- and optimization methods.
A Prolog application for reasoning on maths puzzles with diagrams
Published in Journal of Experimental & Theoretical Artificial Intelligence, 2022
Riccardo Buscaroli, Federico Chesani, Giulia Giuliani, Daniela Loreti, Paola Mello
Csenki (2006) focuses on a specific kind of puzzles connected to the numbers-in-diagram class, and proposes a logic-based solution employing a combination of discrete mathematics and Prolog. However, the approach was only devoted to exploit the power of first order logic to create an effective declarative solver, rather than an autonomous, multimodal one. In (Lev et al., 2004), the authors propose a solver for logical puzzles expressed in an intermediate language. An automatic translation into first-order logic enable the resolution through a theorem prover. More recently, Dries et al. (2017) adopt a two-step approach to solve probability word problems. In the first step, the text in natural language is analysed and converted into an ad-hoc declarative modelling language, which takes inspiration from the entities defined in (Hosseini et al., 2014). Then, in the second step, Probabilistic Inductive Logic Programming (PILP; De Raedt et al., 2008; De Raedt & Kimmig, 2015) is used to infer the solution. Akin to Statistical Relational Learning (SRL; Koller et al., 2007; De Raedt et al., 2016), PILP is particularly suitable for the domain of probability world problems because it allows to manage the complexity through logic, while uncertainty is addresses by probabilistic reasoning. A combination of logical inference and statistical approach is used also in (Liang et al., 2016) to transform the text of maths world problems into an intermediate form, and then into a first-order logic program.
Adjudication of coreference annotations via answer set optimisation
Published in Journal of Experimental & Theoretical Artificial Intelligence, 2018
While coreference resolution is different from coreference adjudication, methods related with Answer Set Programming have been used to perform coreference resolution. Denis and Baldridge (2009) described an approach for coreference resolution and named entity classification based on Integer Linear Programming, which includes transitivity of mention-mention links. Inoue, Ovchinnikova, Inui, and Hobbs (2012) created an approach for coreference resolution based on weighted abduction that is evaluated using an Integer Linear Programming formulation. Note that Integer Linear Programming is a formalism that is related to ASP (Liu, Janhunen, & Niemelä, 2012). Mitra and Baral (2016) describe an approach for question answering based on Inductive Logic Programming under ASP semantics, which includes coreference resolution as a subtask.