WelcomeProfessor John O'Halloran, Deputy President, UCC Professor Barry O'Sullivan, Head of Department, Computer Science
Boole Lecture Series - 2015 Lecture
My Life with Boolean AlgebraProfessor Dana Scott, CMU, USA
Abstract. In 1950 the speaker entered UC Berkeley as a first-year major in Mathematics and very soon was confronted with learning about Boolean algebras. And the encounters continue to this day! The talk will first review elementary ideas and definitions, and then discuss the impact of the work of nine 20th-century pioneers the speaker knew personally. The concluding section will outline prospects for future applications, list some outstanding problems, and point out key influences on the foundations of mathematics.
|20:30||Reception (Aula Maxima)|
Opening RemarksProfessor Barry O’Sullivan
George Boole and Thinking MachinesProfessor Des MacHale, University College Cork
Abstract. We ask the question: Did George Boole foresee the invention of a thinking machine or computer? Using quotes from Boole himself and his wife Mary Everest Boole we discuss this topic. We also describe an eyewitness report by Joseph Hill of the meeting between Babbage and Boole in London in 1862 and give Boole's own reaction to this meeting.
Playing Games with Professor BooleProfessor Michael Wooldridge, University of Oxford
Abstract. Game theory is the theory of how self-interested individuals interact. Although it is usually thought of as part of economics, the main ideas developed within game theory have proved to be of value in many areas of scientific research, from biology to my own field of interest -- computer science. In this talk, I will start by introducing the basic ideas of game theory, by way of a number of examples, and will tease out some of the paradoxes and problems that arise in game theoretic analysis. I will then go on to show how game theoretic analysis is relevant to computing, by introducing Boolean games - named after the eponymous professor himself - which are increasingly used within artificial intelligence to understand how intelligent computer systems might interact. The talk assumes no knowledge of game theory, logic, or computer science.
Puzzling Out the Laws of ThoughtProfessor Eugene C. Freuder
Abstract. Puzzles can be used as a medium for thinking about "laws of thought". Reasoning and modelling methods used on puzzles can be abstracted into general principles of "constraint reasoning". Constraint reasoning principles can be programmed into computers, with broad practical applicability. One way in which University College Cork has carried on the legacy of George Boole is to have become one of the leading centers internationally for the study of this form of reasoning.
Address by Dr. Michael Murphy, President of UCC
On Boolean WingsDr. Ken Ford, Institute for Human Machine Cognition
Abstract. After decades of pundits and philosophers arguing that AI is impossible, suddenly that argument has been replaced with the assertion that not only is it possible, but that it is inevitable, perhaps imminent, and apocalyptically dangerous. In only about a decade, the conversation has shifted from you can’t do it … to we shouldn’t do it ! My purpose in this talk will not be to go into these arguments, but rather to draw your attention to an interesting historical parallel between AI and another, older, technology which was also controversial, thought to be impossible, and then deemed to be a great danger to the human race: artificial flight. From the very beginning, and until modern times, attempts at flight sought to imitate the behavior and specific implementation details of birds. But the Wright brothers were not trying to mimic bird flight, or build an ornithopter. They asked quite different questions, not about flapping or feathers, but about lift, stability and the dynamics of turning in air. The “imitation game” of the Turing Test has misdirected the ambitions of AI, just as a concern with feathers and flapping misdirected early efforts at flight. Now that we understand them, it is clear that the laws of aerodynamics apply to any wing, natural or artificial; and in the same way the laws of thought apply to reasoning done by any cognitive agent, humans, machine or — we think most interesting of all — a combination of both, working together. Boole believed, as did Leibniz and Lull before him, that human thought is mastered by laws, which could account for how people think. Perhaps computation itself, for which Boole’s ideas played a critical role, is the air that provides lift for the wings of thought.
An Overview of the Boole Library Collection & ExhibitionCronan O’Doibhin
George Boole’s Remarkable FamilyOlivia Frawley and Kevin Boole
Abstract. George Boole (1815-1864) changed the world we live in today. He was one of the Forefathers of the Information Age and the first Professor of Mathematics at Queen’s College Cork (now University College Cork). Despite the ubiquity of application of his mathematics and global familiarity with the term “Boolean Logic”, George Boole remains virtually unknown outside the world of Computer Scientists and Mathematicians. This seminar will delve into his remarkable life as well as providing a fascinating insight into his family’s accomplishments. This talk is for anyone who likes to hear a good story about a complex man and his remarkable family.
And, Or , Not - George Boole's Legacy to librariansColette McKenna, Director of Library Services, UCC
Abstract. Ever since libraries came into being, librarians have been the guardians of the wonderful world of information contained within. Helping others to find and make sense of the vast amount of information contained within physical walls and now within virtual walls has and will continue to be the raison d'être of librarians. Without the use of Boolean operators, our AND, OR and NOT, we would all drown in the vast sea of both useful and useless information. This session will illustrate how important George Boole's Boolean operators are to us.
The Real Laws of ThoughtProfessor Geoffrey Hinton, University of Toronto and Google
Abstract. For more than half a century, research in Artificial Intelligence was dominated by the idea that human reasoning consisted of manipulating discrete symbolic representations using formal rules of inference. Intuitive and analogical reasoning were regarded as curiosities to be explained later. Biologically, however, formal reasoning is a very late development that is difficult for our brains. The neural networks in our brains are far more suited to massively parallel, intuitive reasoning in which large patterns of neural activity are used as distributed representations of beliefs. These patterns directly cause other patterns of neural activity that represent intuitively obvious conclusions. Unlike formal logic in which the truth of the conclusions is guaranteed, these direct intuitive inferences are compelling because they simultaneously satisfy a large number of learned statistical regularities. I will describe a simple example that uses deep learning to achieve this type of direct intuitive inference. I will then describe some of the recent rapid progress in Artificial Intelligence that has come from learning distributed representations in neural networks.
This talk is by video link.
The Future of Artificial IntelligenceDr. Oren Etzioni, Allen Institute for Artificial Intelligence
Abstract. How should we build on the success of Machine Learning, and most recently of Deep Learning, over the coming decades? Does AI research create threats for society, or will it be a source of beneficial technology? My talk will address these issues by describing the projects and perspective at the Allen Institute for AI - www.allenai.org - in Seattle.
This talk is by video link.
WelcomeProfessor Barry O’Suillivan
Computational Logic for Artificial and Human IntelligenceProfessor Robert Kowalski, Imperial College London
Abstract. Symbolic logic has been used in artificial intelligence over the past 60 years or so, in the attempt to program computers to display human levels of intelligence. As a result, new forms of computational logic have been developed, which are arguably both more powerful and more practical than earlier forms of logic. The new computational logic that I will describe in my talk is the logic of an intelligent agent whose task in life is to make its goals true, by performing actions to change the world, in response to the changes that it observes in the world. For this purpose, the agent uses beliefs in logical form both to reason forwards, to derive consequences of its observations, and to reason backwards, to reduce goals to subgoals and actions. I will argue that this form of computational logic can be used not only for artificial intelligence, but also for more conventional computing; and because it improves upon traditional logic, it can also be used for the original purpose of logic, to help people improve their own natural intelligence.
Conversations about George BooleProfessor Michel Schellekens, University College Cork
Abstract. The project “Conversations on George Boole; the legacy in logic, mathematics, computer science and engineering”, led by Prof. Michel Schellekens (UCC, Dept. of Computer Science), traces the legacy of Boole’s work in these areas through in-depth interviews with specialists. The interviews were conducted by Michel Schellekens and filmed by UCC cameraman, Stephen Bean, at UCC, Stanford University, UC Berkeley, MIT, the Fields Institute and the Kennedy Space Centre among other locations. The project is supported by UCC’s office of the VP for teaching and learning, John O’Halloran, and is funded by UCC’s Boole Committee and a contribution from the INSPIRE national graduate education programme coordinated by Tyndall National Institute, UCC.
A series of thirteen legacy interviews with specialists have been completed, focusing on Boole’s original work, its impact on mathematics, logic, computer science and electrical engineering, as well as industrial applications. The interviews will be archived and made available through an online portal featuring full length interviews, headed by short trailers highlighting excerpts.
Our presentation will give an overview of the project, its execution and interesting insights shared by specialists on Boole’s work.
Embedding Ethical Principles in Group Decision-making SystemsProfessor Francesca Rossi, University of Padova
Abstract. The future will see autonomous agents acting in the same environment as humans, in areas as diverse as driving, assistive technology, and health care. In this scenario, collective decision making will be the norm. I will discuss the embedding of safety constraints, moral values, and ethical principles in agents, within the context of hybrid human/agents collective decision making. I will propose to do that by adapting current logic-based modelling and reasoning frameworks, such as soft constraints, CP-nets, and constraint-based scheduling under uncertainty. For ethical principles, constraints could specify the basic ethical "laws", plus sophisticated prioritised and possibly context-dependent constraints over possible actions, equipped with a conflict resolution engine. To avoid reckless behavior in the face of uncertainty, we can bound the risk of violating these ethical laws. I will also propose to replace preference aggregation with an appropriately developed constraint/value/ethics/preference fusion, an operation designed to ensure that agents' preferences are consistent with the system's safety constraints, the agents' moral values, and the ethical principles of both individual agents and the collective decision making system. I will also discuss approaches to learn ethical principles for artificial intelligent agents, as well as predict possible ethical violations.
Based on a joint research project with:
J. Greene (Harvard Univ.),
J. Tasioulas (King’s College London),
K. B. Venable (Tulane Univ. and IMHC),
B. Williams (MIT)
Human Thought as Deduction: How “The Laws of Thought” Idea Has Fared...Professor Mark Keane, UCD
Abstract. Boole’s original proposal was that his algebra captured “The Laws of Thought”; that his logic would be a good theory of what people actually do when they think and reason. In this lecture, I assess how far this idea has gone in the Cognitive Sciences by briefly reviewing theories of human deductive reasoning, with a specific emphasis on reasoning with conditionals. These theories move from natural deduction accounts (that Boole would arguably recognise) to mental model’s theories to more recent probabilistic theories. I conclude with my views on the impact of Boole’s proposal on the Cognitive Science of thinking and the extent to which these accounts have been directly or indirectly influenced by him.
Grenville Place - The Boole House of InnovationProfessor Barry O'Sullivan, Computer Science, UCC; Mark Poland, Buildings and Estates, UCC; Neil Purkiss and Sorcha Coleman, City Architect’s Department, Cork City Council.
Abstract. George Boole wrote is most famous work, The Laws of Thought, at 5 Grenville Place. In this session we will present the collaboration between UCC and Cork City Council focusing on the renovation of the house, and the Boole House of Innovation that it will be located there.
Coder Poets - Where to next for George Boole's legacy?Bill Liao, CoderDofo and SOSVentures
Abstract. Traversing between the modern global phenomena that is CoderDojo to the latest advances in bio-technology Bill Liao will take us on a grand tour of where the amazingly astute theories of George Boole will take us in the near future.
From Leibniz to Boole to Big DataProfessor John Hooker
Abstract. George Boole is primarily recognized for his work in propositional logic, but he made a strikingly original contribution to the logic of probabilities, one that is more relevant than ever in our age of big data. Even more, it represents a pivot from the rationalism of the Enlightenment to the information processing perspective we have today. Western civilization’s faith in the intelligibility of the world reached its zenith in 17th century rationalism, personified by Gottfried Wilhelm von Leibniz, who undertook to design a calculus of reasoning that could deduce all truths. Boolean algebra is a 19th century reincarnation of such a calculus, but meanwhile rationalism had given way to an empirical worldview that replaces certainty with probability. Boole sensed this transformation and addressed it with a possible-world semantics for probabilistic inference that anticipates modern optimization methods. These ideas were ignored or dismissed for a century, until his semantics reappeared in modal logics of the 1960s, and his optimality formulation was reinvented in the AI community of the 1980s. The natural world of the Enlightenment had become today’s bewildering world of endless data that must somehow be processed and assimilated. Our faith in intelligibility persists even in this new world, as we attempt to extract knowledge through data mining, crowdsourcing, and the like. Boole’s probability logic was the first to address this task, and even now its potential has not been fully exploited.
The Legacy of George Boole: from Boolean Satisfiability to Constraint ProgrammingProfessor Barry O’Sullivan
Abstract. There are many problems in computer hardware design, production scheduling, timetabling, product configuration, planning, diagnosis, etc. that can be stated in propositional logic. Determining whether there exists a design that is bug free, or finding an acceptable schedule, timetable, configuration, etc., corresponds to determining whether the statement in propositional logic is true or false. These are some of the many applications of Boolean satisfiabiliity and the software tools, the solvers, that have been developed over the past few decades. In this talk we will give examples of how Boole’s work leaves a legacy in solving extremely challenging and important problems. We will see how quickly this field has advanced, and how it relates to other fundamental problem-solving technologies such as constraint programming. It has often been said that constraint programming is the closest that computer science has come towards the ‘holy grail’ of programming.
|17:30||Wrap-up and Close|