The "Artificial Mathematician" Objection: exploring the (im)possibility of automating mathematical understanding (PROOF VERSION (original) (raw)
Related papers
Correctness, Artificial Intelligence, and the Epistemic Value of Mathematical Proof
We argue that it is neither necessary nor sufficient for a mathematical proof to have epistemic value that it be “correct”, in the sense of formalizable in a formal proof system. We then present a view on the relationship between mathematics and logic that clarifies the role of formal correctness in mathematics. Finally, we discuss the significance of these arguments for recent discussions about automated theorem provers and applications of AI to mathematics.
Towards computer aided mathematics
Journal of Applied Logic, 2006
In his autobiography 1 Bertrand Russell characterizes mathematics as follows: "It seems to me now that mathematics is capable of an artistic excellence as great as that of any music, perhaps greater; not because the pleasure it gives (although very pure) is comparable, either in intensity or in the number of people who feel it, to that of music, but because it gives in absolute perfection that combination, characteristic of great art, of godlike freedom, with the sense of inevitable destiny; because, in fact, it constructs an ideal world where everything is perfect and yet true". Actually the perception of mathematical research as an artistic discipline has a long history and a significant number of today's mathematicians share this view. In contrast, however, Russell himself dedicated large parts of his life to defending logicism, that is, the view that mathematics is reducible to logic, and together with Alfred Whitehead he proposed in his influential Principia Mathematicae an axiomatic system to build all mathematics upon. The two viewpoints-mathematics as an art versus logicism-may appear contradictory at first. They are not though, if we separate the different aspects of mathematical practice. The invention and shaping of new mathematical structures based on mathematical knowledge as well as on aesthetic and social criteria or the discovery of the essential arguments in complex mathematical proof, for instance, are activities that typically require human ingenuity. On the other hand the verification and grounding of already pre-structured and established chunks of mathematics in foundational systems or the search for simple (sub-)proofs are examples of tasks that often require far less ingenuity. Some overoptimistic and improperly reflected predictions in the field of artificial intelligence and automated reasoning on the mechanization and automation of mathematics have unfortunately generally questioned the role of human ingenuity in mathematics without making the above distinction clear. Unlike in chess, however, where human intelligence is no longer dominating over machine intelligence, it seems to me that human ingenuity will remain dominant in many essential aspects in mathematics research and education for a long time to come. Taking our distinction above into account this does not mean, however, that there is no need for assistance systems for mathematics and Russell would presumably 1
Automated theorem provers: a practical tool for the working mathematician?
Annals of Mathematics and Artificial Intelligence, 2011
In contrast to the widespread use of computer algebra systems in mathematics automated theorem provers have largely met with indifference. There are signs that this is at last beginning to change. We argue that it is inevitable that automated provers will be adopted as a practical tool for the working mathematician. Mathematical applications of automated provers raises profound challenges for their developers. Keywords Automated theorem proving • Mathematician's assistant • Very large proofs • Proof understanding The research reported in this paper was supported by EPSRC grants EP/F033559/1 and EP/H023119/1. I am indebted to Bogdan Grechuk, Lucas Dixon, Ursula Martin and an anonymous amai referee for valuable feedback on earlier drafts, and to Michael Chan for help with the submission.
The Automation of Mathematical Reasoning
One of the main goals of automated reasoning has been the automation of mathematics. Reasoning is the ability to make inferences, and automated reasoning is concerned with the building of computing systems that automate this process. Although the overall goal is to mechanize different forms of reasoning, the term has largely been identified with valid deductive reasoning as practiced in mathematics and formal logic. Building an automated reasoning program means providing an algorithmic description to a formal calculus so that it can be implemented on a computer to prove theorems of the calculus in an efficient manner. [7] Automated reasoning has made a lot of striking successes over the last decades. It evolved into a rich scientific discipline, with many subdisciplines and with solid grounds in mathematics and computer science. Over the years, automated reasoning transformed from a research field based on mathematical logic into a field that is a driving force for mathematical logic. Nowadays, automated reasoning tools are used in everyday practice in mathematics, computer science and engineering. So, automated reasoning programs are being used by researchers to attack open questions in mathematics and logic, provide important applications in computing science, solve problems in engineering, and find novel approaches to questions in exact philosophy. Also, its role in education increases and will increase in the future.
Discovery and reasoning in mathematics
1985
We discuss the automation of mathematical reasoning, surveying the abilities displayed by human mathematicians and the computational techniques available for automating these abilities. We argue the importance of the simultaneous study of these techniques, because problems inherent in one technique can often be solved if it is able to interact with others.
6th International Week Dedicated to Maths 2014, Mathematical meetings looking for the infinity,Thessaloniki,Greece, ISBN: 978-960-89672-5-0, 2014
In this paper we give the main moments of the historical evolution of AI which as we argue is closed related with the development of mathematics in 1870-1930.We illustrate this relationship with some examples from mathematics which are widely used for programming intelligent machines.
Alan Turing and the Mathematical Objection
Minds and Machines, 2003
This paper concerns Alan Turing's ideas about machines, mathematical methods of proof, and intelligence. By the late 1930s, Kurt Gödel and other logicians, including Turing himself, had shown that no finite set of rules could be used to generate all true mathematical statements. Yet according to Turing, there was no upper bound to the number of mathematical truths provable by intelligent human beings, for they could invent new rules and methods of proof. So, the output of a human mathematician, for Turing, was not a computable sequence (i.e., one that could be generated by a Turing machine). Since computers only contained a finite number of instructions (or programs), one might argue, they could not reproduce human intelligence. Turing called this the "mathematical objection" to his view that machines can think. Logico-mathematical reasons, stemming from his own work, helped to convince Turing that it should be possible to reproduce human intelligence, and eventually compete with it, by developing the appropriate kind of digital computer. He felt it should be possible to program a computer so that it could learn or discover new rules, overcoming the limitations imposed by the incompleteness and undecidability results in the same way that human mathematicians presumably do.
Ωmega: From Proof Planning towards Mathematical Knowledge Management
The vision of computer-based mathematical assistance systems providing integrated support for all work phases of a mathematician (see Figure 1 from [12]) has fascinated researchers in artificial intelligence, particularly the deduction systems area, and in mathematics for a long time. The dream of mechanizing mathematical reasoning dates back to Gottfried Wilhelm Leibniz in the 18th century. In the beginning of the 20th century modern mathematical logic was born and an important milestones in the formalization of ...