No file available
To read the file of this research,
you can request a copy directly from the author.
Toward human-level goal reasoning with a natural language of thought
Abstract
What is the nature of goal reasoning needed for human-level artificial intelligence? This presentation contends that to achieve human-level AI, a system architecture for human-level goal reasoning would benefit from a neuro-symbolic approach combining deep neural networks with a ‘natural language of thought’ and would be greatly handicapped if relying only on formal logic systems.
ResearchGate has not been able to resolve any citations for this publication.
We present a metacognitive, integrated, dual-cycle architecture whose function is to provide agents with a greater capacity for acting robustly in a dynamic environment and managing unexpected events. We present MIDCA 1.3, an implementation of this architecture which explores a novel approach to goal generation, planning and execution given surprising situations. We formally define the mechanism and report empirical results from this goal generation algorithm. Finally, we describe the similarity between its choices at the cognitive level with those at the metacognitive.
Is synthetic phenomenology a valid concept? In approaching consciousness from a computational point of view, the question of phenomenology is not often explicitly addressed. In this paper we re- view the use of phenomenology as a philosophical and a cognitive construct in order to have a meaningful transfer of the concept into the computational domain. Two architectures are discussed with respect to these definitions: our 'kernel, axiomatic' structure and the widely quoted 'Global Workspace' scheme. The conclusion suggests that architectures with phenomenal properties genu- inely address the issue of modelling consciousness and indicate and the way that a machine with synthetic phenomenology may benefit from the property
Little is clearly understood about the similarities, differences, and comparative computational and representational advantages of the many proposals extant for organizing minds into collections of "mental subagents." Using a new mathematical framework for exactly specifying the structure of mental organizations, we formulate separately the ideas of multiple perspectives, reasoned assumptions, and virtual copies. When combined, these notions form a common backbone for systems as diverse as CONLAN, NETL, and FOL, and show many particular characteristics of those systems results of the "language of thought" adopted for representing the contents of mental subagents. The framework also suggests connections between the "strengths" of mental attitudes, the ambiguity of "self," and the possibilities for self-omniscience.
Throughout the history of AI, logic has been praised by its admirers, maligned by its detractors, and discussed in confusing and misleading terms by almost everybody. Among the pioneers of AI, John McCarthy has always been a strong promoter of logic, but Marvin Minsky has been a skeptic who experimented with a wide range of alternatives. Roger Schank had no doubts about logic, which he denounced at every opportunity. He introduced the distinction between the neats who used logic for everything vs. the scruffies like himself who developed notations that were specifically designed for the problem at hand. Even advocates of logic have disagreed among themselves about the role of logic, the subset appropriate to any particular problem, and the tradeoffs of ease of use, expressive power, and computational complexity. The debates introduced many valuable ideas, but the hype and polemics confused the issues and often led to unfortunate design decisions. Controversies arise in several areas: relationships between language and logic; the range of notations for logic; procedural vs. declarative representations; distinctions between object level and metalevel statements; expressive power and complexity; and the role of logic in readable, usable, efficient interfaces. This is a preprint of an article that appeared in IEEE Intelligent Systems, March 2007, 22:2, pp. 84-87. 1. Language and Logic
In this paper we present Uniform Meaning Representation (UMR), a meaning representation designed to annotate the semantic content of a text. UMR is primarily based on Abstract Meaning Representation (AMR), an annotation framework initially designed for English, but also draws from other meaning representations. UMR extends AMR to other languages, particularly morphologically complex, low-resource languages. UMR also adds features to AMR that are critical to semantic interpretation and enhances AMR by proposing a companion document-level representation that captures linguistic phenomena such as coreference as well as temporal and modal dependencies that potentially go beyond sentence boundaries.
How can human-level artificial intelligence be achieved? What are the potential consequences? This book describes a research approach toward achieving human-level AI, combining a doctoral thesis and research papers by the author.
The research approach, called TalaMind, involves developing an AI system that uses a 'natural language of thought' based on the unconstrained syntax of a language such as English; designing the system as a collection of concepts that can create and modify concepts to behave intelligently in an environment; and using methods from cognitive linguistics for multiple levels of mental representation. Proposing a design-inspection alternative to the Turing Test, these pages discuss 'higher-level mentalities' of human intelligence, which include natural language understanding, higher-level forms of learning and reasoning, imagination, and consciousness. Dr. Jackson gives a comprehensive review of other research, addresses theoretical objections to the proposed approach and to achieving human-level AI in principle, and describes a prototype system that illustrates the potential of the approach.
This book discusses economic risks and benefits of AI, considers how to ensure that human-level AI and superintelligence will be beneficial for humanity, and gives reasons why human-level AI may be necessary for humanity's survival and prosperity.
