Artificial General Intelligence

Name: Artificial General Intelligence | 4th International Conference, AGI 2011, Mountain View, CA, USA, August 3-6, 2011, Proceedings
Brand: Springer
Price: 53.49 EUR
Availability: OnlineOnly

4th International Conference, AGI 2011, Mountain View, CA, USA, August 3-6, 2011, Proceedings

Jürgen Schmidhuber Kristinn R. Thorisson Moshe Looks(Herausgeber*in)

Springer (Verlag)

Erschienen am 19. Juli 2011

XIV, 416 Seiten

E-Book

PDF mit Wasserzeichen-DRM

Systemvoraussetzungen

978-3-642-22887-2 (ISBN)

53,49 €inkl. 7% MwSt.

Systemvoraussetzungen

für PDF mit Wasserzeichen-DRM

E-Book Einzellizenz

Als Download verfügbar

Beschreibung

Weitere Details

Weitere Ausgaben

Inhalt

Title
Preface
Organization
Table of Contents
Long Papers
Self-Modification and Mortality in Artificial Agents
Introduction
Universal Agents Ax
Various Universal Agents
Self-Modifiable Agents
Optimality of self-modifiable Agents
Embedded, Mortal AI
Conclusions
References
Delusion, Survival, and Intelligent Agents
Introduction
UniversalAgentsA?x
The Delusion Box
Discussion
Survival Machines
Discussion and Conclusion
References
Coherence Progress: A Measure of Interestingness Based on Fixed Compressors
Introduction
Interestingness
Fixed, Adaptive and Ideal Compressors
Coherence Progress
Qualitative Correctness
Oversimplified Alternatives
Coherence Progress Versus Learning Progress
Experiments
Discussion
Conclusion
References
Sequential Constant Size Compressors for Reinforcement Learning
Introduction
Sequential Recurrent Auto-associative Memory
SERVO: SCSC Assisted RL
UL Layer: sRAAM
Internal State Layer: VQ
RL Layer: SARSA()
Experiments and Results
Partially Observable Vision Maze
Learning to Wait
Discussion
References
Planning to Be Surprised: Optimal Bayesian Exploration in Dynamic Environments
Introduction
Preliminaries
Learning from Sequential Interactions
Information Gain as Learning Progress
Optimal Bayesian Exploration
Results for Finite Time Horizon
Non-triviality of the Result
Extending to Infinite Horizon
Experiment
Related Work
Conclusion
References
Optimal Direct Policy Search
Introduction
Reinforcement Learning in POMDPs
MDPs and POMDPs
Objectives of Learning
Direct Search in Policy Space
Optimal Direct Policy Search
Formal Optimality
Discussion
Experimental Test
Conclusion
References
Nonlinear-Dynamical Attention Allocation via Information Geometry
Introduction
Brief Review of OpenCog
Brief Review of Economic Attention Networks
Brief Review of Information Geometry
From Information Geometry to Mind Geometry
Information-Geometric Learning for Recurrent Networks: Extending the ANGL Algorithm
Information Geometry for Economic Attention Allocation: A Detailed Example
Conclusion
References
An Information Theoretic Representation of Agent Dynamics as Set Intersections
Introduction
Interaction as Intersection
Non-cooperative Games
Cybernetic Agent Model
Background in Algorithmic Information Theory
Player Strategy Learning
Player Approximation and Interaction Complexity
Future Work: Universal Interaction
Conclusions
References
On More Realistic Environment Distributions for Defining, Evaluating and Developing Intelligence
Introduction
Artificial Life, Biology and Intelligence
Darwin-Wallace Distribution
Approximating the Distribution through Testing
Discussion
References
Structural Emergence in Partially Ordered Sets Is the Key to Intelligence
Introduction
Theory
Small Systems
Experimental Evidence
Concluding Remarks
References
Integrating Perception and Cognition for AGI
Introduction
ACT-R
Partial Matching
Blending
Modeling Object Identification, Tracking and Action Prediction in ACT-R
Checkpoint Scenario
ACT-R Model
Recovering from Object Identification Errors
Recovering from Object Location Errors
Predicting Actions
Conclusion and Future work
References
Systematically Grounding Language through Vision in a Deep, Recurrent Neural Network
Introduction
Background
Systematicity
Past Neural Network Models of Grounded Language Learning
Methods
Task Description
Scenes, Objects, and the Visual Stream.
Sentences, Phonemes, and the Auditory Stream.
Meanings, Predicates, and the Intention Stream.
A Complete Example Trial.
Network Architecture
Experimental Evaluation
Results
Discussion
References
Comparing Humans and AI Agents
Introduction
Measuring Intelligence Universally
Test Setting and Administration
Agents and Interfaces
An AI Agent: Q-Learning
A Biological Agent: Homo Sapiens
Results
Discussion
References
The LIDA Framework as a General Tool for AGI
Introduction
Frameworks
Frameworks and Cognitive Architectures
The LIDA Model and Its Architecture
The LIDA Framework
Framework Outline
Framework Tools
Underlying Assumptions of the LIDA Framework
Conclusions
References
From Memory to Problem Solving: Mechanism Reuse in a Graphical Cognitive Architecture
Introduction
Problem Solving in Soar
The Graphical Memory Architecture
Extension to Problem Solving
Conclusion
References
Rational Universal Benevolence: Simpler, Safer, and Wiser Than "Friendly AI"
Introduction
"Friendship Structure" and "Coherent Extrapolated Volition"
Rational Universal Benevolence
Motivations and Distinctions
Defend Humans, the Future of Humankind, and the Destiny of Humankind
Humane Nature and Moral Growt h
Conclusion
References
The Collection of Physical Knowledge and Its Application in Intelligent Systems
Introduction
Background
Representing Physical Knowledge
Collecting Physical Knowledge
Applying Physical Knowledge
Conclusion
References
Rationality and General Intelligence
Introduction
Rationality Concepts and Challenges
Rationality
Well-Known Challenges
Non-standard Interpretations of Wason and the Linda Problem
Rationality and AGI Systems
Modeling Rationality
Coherence and Analogies
Conclusions
References
Wagging the Dog: Human vs. Machine Inference of Causality in Visual Sequences
Introduction
Historical Terms of Debate: Intention, Purpose, Cause, Will
Causality in Machine Learning
Causality in Artificial Intelligence
Causality in Cognitive Science, Neuroscience and Psychology
Experimental Methods
Human vs. Automatic Causal Inference: Results
Discussion
References
Learning Problem Solving Skills from Demonstration: An Architectural Approach
Introduction
Architecture
Learning Problem Solving Skills
Bootstrapping the Learning Process: Masterplan
Case Study: Sokoban
Results: Motor Babbling
Results: Imitation
Conclusions
References
Compression and Intelligence: Social Environments and Communication
Compression, Inference, Prediction and Intelligence
Social Environments and Communication
Lossless and Lossy Compression
The Elusive Model Paradox and (Human) Unpredictability
Detecting and Assessing Intelligence
Conclusions
References
Open Psi: Realizing D¨orner's "Psi" Cognitive Model in the OpenCog Integrative AGI Architecture
Introduction
Motivation and Emotion in Psi
OpenCog Prime
Psi Versus OpenCogPrime
OpenPsi
Conclusion
References
Extending Cognitive Architectures with Semantic Resources
Introduction
Cognitive Architectures as Knowledge Systems
Hybrid Semantics for Declarative Memory
HOMinE and ACT-R: An Integrated Cognitive Model
Design and Implementation of HOMinE
Mapping HoMinE to ACT-R
Conclusions
References
A Motivational System for Cognitive AI
Introduction: The Quest for General Intelligence
What Is Cognitive Artificial Intelligence?
Prerequisites for Defining a Motivational System
An Outline of a Motivational System, According to the Psi Theory
Needs
Behavior Control and Action Selection
Summary
References
On Fast Deep Nets for AGI Vision
Introduction
Neural Network ReNNaissance
Outlook
Conclusion
References
Considerations for a Neuroscience-Inspired Approach to the Design of Artificial Intelligent Systems
Introduction
Insights from Embodied Cognition
Neuroscience for Artificial Agents: The Mirror System Example
Conclusion
References
Brain Anatomy and Artificial Intelligence
Introduction
Brain Information Architecture
Cortex and Clustering
Indirect Activation of Receptive Fields
Basal Ganglia and Behaviour Selection
Reward Behaviours
Thalamus and Cortical Information Management
Cerebellum and Behaviour Sequence Management
Lessons for Design of General Intelligence Systems
References
Short Papers
Information, Utility and Bounded Rationality
Introduction
Bounded Rationality
Applications
Conclusion
References
A Family of G¨odel Machine Implementations
Introduction
Gödel Machine Specification
Discussion and Conclusion
References
Reinforcement Learning and the Bayesian Control Rule
Introduction
Setup
Critic
Actor
Experimental Results
Discussion and Conclusion
References
Societies of Intelligent Agents
Introduction
Language
Societies of Intelligent Agents
References
Towards a General Vision System Based on Symbol-Relation Grammars and Bayesian Networks
Introduction
Representation and Recognition
Visual Grammar
Transforming a SR-Grammar to a Bayesian Network
Examples
Visual Grammar for Face Detection
Low-Level Features of a Visual Grammar for Eyes
Conclusions and Future Work
References
Reasoning in Non-Axiomatic Logic: A Case Study in Medical Diagnosis
Problem and Background
NAL Overview
Testing Cases and Results
Comparison and Discussion
Conclusions
References
Measuring Agent Intelligence via Hierarchies of Environments
Introduction
Hierarchies of Environments for Measuring Intelligence
A Hierarchy of Finite State Machines
Discussion and Conclusion
References
Learning What to Value
Agents and Implementations
Reinforcement Learning
Learning What to Value
Conclusion
References
Learning the States: A Brain Inspired Neural Model
Introduction
Two Stages of the Architecture
Illustrative Simulations on Natural Images
RPCA Based Hierarchical Architecture
Discussion
Searching for States and Top-Down Encoding.
Conclusions and Outlook
References
AGI Architecture Measures Human Parameters and Optimizes Human Performance
Introduction
Architecture
Experiments
Results and Conclusions
References
Multigame Playing by Means of UCT Enhanced with Automatically Generated Evaluation Functions
Introduction
General Game Playing Competition
UCT
Guided UCT
Evaluation Function
Features Generation
Evaluation Function Generation
Experiment
Conclusions and Future Research Plans
References
From Sensorimotor Graphs to Rules: An Agent Learns from a Stream of Experience
Introduction
Learning Context-Sensitive Partitioning of Sensorimotor Space
Handling a Stream of Experience
Learning Productive Rules
Conclusion and Future Work
References
Three Hypotheses about the Geometry of Mind
Introduction
Metrics on Memory Spaces
Three Hypotheses about the Geometry of Mind
References
Imprecise Probability as a Linking Mechanism between Deep Learning, Symbolic Cognition and Local Feature Detection in Vision Processing
Introduction
Visual Attention Focusing
Imprecise Probabilities
Using Imprecise Probabilities to Guide Vision Processing
Conceptual Justification
References
Generalization of Figure-Ground Segmentation from Binocular to Monocular Vision in an Embodied Biological Brain Model
Introduction
Materials and Methods
CU3D-100 Dataset
Structure of the Models
Results and Discussion
References
The Illusion of Internal Joy
Introduction
Cortex Organization
Behaviors are Goal-Directed
Coordination of Pairs of Cortical Maps
Illusions
The Illusion of '$Joy Driven$ ' Behaviors
The Illusion of '$Internal Joy$ '
Discussion and Conclusion
References
Philosophically Inspired Concept Acquisition for Artificial General Intelligence
Introduction
The Philosophical Debate about Concepts
The Building Blocks Model
Empiricism vs. Nativism about Lexical Concepts
The Debate Suggests that Lexical Concepts are Simple Yet Acquired
Baptizing Meanings for Concepts and a Softbot Implementation
Conclusion
References
Machine Lifelong Learning: Challenges and Benefits for Artificial General Intelligence
Introduction
A Framework for Machine Lifelong Learning
Related Work
Current Computational and Data Storage Capacity
Challenges and Benefits
References
A Demonstration of Combining Spatial and Temporal Perception
Introduction
Overview of the Model
Experiment 1
Application of the Model
Results
Experiment 2
Application of the Model
Results
Conclusion
References
Towards Heuristic Algorithmic Memory
Introduction
Stochastic Context-Free Grammar Updates
Modifying Production Probabilities
Re-using Previous Solutions
Learning Programming Idioms
Frequent Sub-program Mining
Experiments
Conclusion and Future Work
References
Complex Value Systems in Friendly AI
No Ghost in the Machine
Complex Boundaries of Value
Completeness of Value
Detailed Inheritance of Humane Values
Conclusion
References
Real-World Limits to Algorithmic Intelligence
Introduction
Universally Optimal Intelligence
Conditions for Knowledge of the World
Knowledge
A World of Objects
Conclusion
References
AGI and Neuroscience: Open Sourcing the Brain
Introduction
Perspective
AGI and the Human Brain
High-Level Insight from Psychology and Cognitive Science vs Neuroscience
Should AGI Learn from the Human Brai n?
Brain Emulation as a Route to AGI
Can We Produce a SIM without Understanding the Mind?
Concluding Remarks
References
What Makes a Brain Smart? Reservoir Computing as an Approach for General Intelligence
Introduction
Reservoir Computing
Critical Branching and Metastability
A Critical Branching Reservoir Model
Reservoir Computing Applications
Conclusion
References
Author Index

Systemvoraussetzungen

Als PDF speichern Als Link merken