We consider here the problem of building a never-ending language learner; that is, an intelligent computer agent that runs forever and that each day must (1) extract, or read, information from the web to populate a growing structured knowledge base, and (2) learn to perform this task better than on the previous day. In particular, we propose an approach and a set of design principles for such an agent, describe a partial implementation of such a system that has already learned to extract a knowledge base containing over 242,000 beliefs with an estimated precision of 74 % after running for 67 days, and discuss lessons learned from this preliminary attempt to build a never-ending learning agent.

We address the problem of how to reinforcement learn in ultra-complex environments, with huge state spaces, where one must learn to exploit compact structure of the problem domain. The approach we propose is to simulate the evolution of an artificial economy of computer programs. The economy is constructed based on two simple principles so as to assign credit to the individual programs for collaborating on problem solutions. We find empirically that, starting from programs that are random computer code, we are able to evolve systems that solve hard problems. In particular our economy as learned to solve almost all random Blocks World problems with goal stacks 200 blocks high. Competing methods solve such problems only up to goal stacks of at most 8 blocks. Our economy has also learned to unscramble about half a randomly scrambled Rubik's cube, and to solve several among a collection of commercially sold puzzles.

The Letter Spirit project is an attempt to model central aspects of human high-level perception and creativity on a computer, focusing on the creative act of artistic letter-design. The aim is to model the process of rendering the 26 lowercase letters of the roman alphabet in many different, internally coherent styles. Two important and orthogonal aspects of letterforms are basic to the project: the categorical sameness possessed by instances of a single letter in various styles (e.g., the letter `a' in Baskerville, Palatino, and Helvetica) and the stylistic sameness possessed by instances of various letters in a single style (e.g., the letters `a', `b', and `c' in Baskerville). Starting with one or more seed letters representing the beginnings of a style, the program will attempt to create the rest of the alphabet in such a way that all 26 letters share the same style, or spirit. Letters in the domain are formed exclusively from straight segments on a grid in order to make decisions ...

Creative products are generally recognised as satisfying two requirements: firstly they are useful, and secondly they are novel. Much effort in AI and design computing has been put into developing systems that can recognise the usefulness of the products that they generate. In contrast, the work presented in this thesis has concentrated on developing computational systems that are able to recognise the novelty of their work. The research has shown that when computational systems are given the ability to recognise both the novelty and the usefulness of their products they gain a level of autonomy that opens up new possibilities for the study of creative behaviour in single agents and the emergence of social creativity in multi-agent systems. The work

In a hostile environment, an autonomous cognitive system requires a reflective capability to detect problems in its own operation and recover from them without external intervention. We present an architecture in which reflection is distributed so that components mutually observe and protect each other, and where the system has a distributed model of all its components, including those concerned with the reflection itself. Some reflective (or “meta-level”) components enable the system to monitor its execution traces and detect anomalies by comparing them with a model of normal activity. Other components monitor “quality ” of performance in the application domain. Implementation in a simple virtual world shows that the system can recover from certain kinds of hostile code attacks that cause it to make wrong decisions in its application domain, even if some of its self-monitoring components are also disabled.

The science of the mind, and of consciousness in particular, needs to carefully consider people's common-sense views of the mind, not just what the mind really is. Such views are themselves an aspect of the nature of (conscious) mind, and therefore part of the object of study for a science of mind. Also, since the common-sense views allow broadly successful social interaction, it is reasonable to look to the common-sense views for some rough guidance as to the real nature of the mind. On the other hand, to the extent that common-sense views are inaccurate, and perhaps even in gross conflict with the true nature of the mind, one interesting scientific question is: why do we hold such views, given our access to our own minds? Why should introspection be limited in a way that allows inaccurate views to hold sway? Now, common-sense views of the mind are revealed in natural language discourse that describes mental states, and such descriptions are largely metaphorical. The metaphors are use...

as a testbed for designing intelligent agents. The system consists of an overall agent architecture and five components within the architecture. The five components are: 1. Goal-Directed Learning (GDL), a decision-theoretic method for selecting learning goals. 2. Probabilistic Bias Evaluation (PBE), a technique for using probabilistic background knowledge to select learning biases for the learning goals. 3. Uniquely Predictive Theories (UPTs) and Probability Computation using Independence (PCI), a probabilistic representation and Bayesian inference method for the agent's theories. 4. A probabilistic learning component, consisting of a heuristic search algorithm and a Bayesian method for evaluating proposed theories. 5. A decision-theoretic probabilistic planner, which searches through the probability space defined by the agent's current theory to select the best action. PAGODA is given as input an initial planning goal (its ove

Induction and deduction are two opposite operations in data mining applications. Induction extracts knowledge in the form of, say, rules or decision trees from existing data, and deduction applies induction results to interpret new data. An intelligent learning database (ILDB) system integrates machine learning techniques with database and knowledge base technology. It starts with existing database technology and performs both induction and deduction. The integration of database technology, induction (from machine learning), and deduction (from knowledgebased systems) plays a key role in the construction of ILDB systems, as does the design of efficient induction and deduction algorithms. This paper presents a system structure for ILDB systems, and discusses practical issues for ILDB applications, such as instance selection and structured induction. 1 Introduction Over the past thirty years database research has evolved technologies that are now widely used in almost every co...

Research efforts for automating the design process of mechanical artifacts have been intensified in recent years. A variety of approaches have been proposed. However, as is usual in a new field, no unified theory or terminology has yet emerged. This situation is complicated further by lack of general agreement on what constitutes "design". Evidently, some common framework or vocabulary is necessary for researchers to be able to communicate and compare their efforts. This paper attempts to outline such a vocabulary and examines, by example, how it could be used to review current published research in mechanical design automation.

This paper reports a computational model of Boole's discovery of Logic as a part of Mathematics. George Boole (1815-1864) found that the symbols of Logic behaved as algebraic symbols, and he then rebuilt the whole contemporary theory of Logic by the use of methods such as the solution of algebraic equations. Study of the different historical factors that influenced this achievement has served as background for our two main contributions: a computational representation of Boole's Logic before it was mathematized; and a production system, BOOLE2, that rediscovers Logic as a science that behaves exactly as a branch of Mathematics, and that thus validates to some extent the historical explanation. The system's discovery methods are found to be general enough to handle three other cases: two versions of a Geometry due to a contemporary of Boole, and a small subset of the Differential Calculus. 1 Introduction In 1847, George Boole found that, by adequately representing Logic, it became a bra...