Selected Papers of Pei Wang

All the following publications are authored by Pei Wang unless specified otherwise.

The on-line versions here may be slightly different from the published versions.

NARS Overview

• Intelligence: From Definition to Design
  [Proceedings of the Third International Workshop on Self-Supervised Learning, PMLR 192:35–47, 2022]
  A non-technical overview of the NARS project
• A Logical Model of Intelligence — an introduction to NARS
  [On-line document since 2009]
  A brief introduction of the project
• From NARS to a Thinking Machine
  [Artificial General Intelligence Research Institute Workshop, Washington DC, May 2006]
  A discussion of the development plan of NARS
• The Logic of Intelligence
  [Artificial General Intelligence, 31-62, Springer, 2006]
  A high-level description of the NARS project
• Toward a Unified Artificial Intelligence
  [AAAI Fall Symposium on Achieving Human-Level Intelligence through Integrated Research and Systems, 83-90, Washington DC, October 2004]
  AI should, and can, be treated as a whole

NARS Description

• Non-Axiomatic Logic: A Model of Intelligent Reasoning
  [World Scientific Publishing Co., 2013]
  The logic of NARS, formally defined and up-to-date. The Appendices contain full specifications of the grammar and inference rules.
• Rigid Flexibility: The Logic of Intelligence
  [also as eBook, with freely accessible Front Matter and Back Matter, Springer, 2006]
  A comprehensive description of the NARS project as of version 4.3
• Non-Axiomatic Reasoning System: Exploring the essence of intelligence
  [Ph.D. Dissertation, Indiana University, 1995]
  A comprehensive description of the NARS project as of version 3.0
• The OpenNARS implementation of the Non-Axiomatic Reasoning System (by Patrick Hammer, Tony Lofthouse, and Pei Wang)
  [AGI-16, New York City, July 2016]
  A description of the open-source implementation of NARS, OpenNARS, as of version 1.7.0
  
• Non-Axiomatic Reasoning System (Version 2.2)
  [Technical Report No. 75 of CRCC, 1993]
  A comprehensive description of the NARS project as of version 2.2, which has been long out-of-date, though has historical value.
  

General AI/AGI Issues

• A Constructive Explanation of Consciousness
  [Journal of Artificial Intelligence and Consciousness, 7(2):257-275, 2020]
  Consciousness is self-awareness and self-control
• On Defining Artificial Intelligence
  [Journal of Artificial General Intelligence, 10(2):1-37, 2019]
  A comprehensive discussion about the objectives of AI research
  Commentaries and Author’s Response
  [Journal of Artificial General Intelligence, 11(2), 2020]
• Conceptions of Artificial Intelligence and Singularity (by Pei Wang, Kai Liu, and Quinn Dougherty)
  [Information 2018, 9(4), 79; doi: 10.3390/info9040079]
  What AGI is and can do
• Safe baby AGI (by Jordi Bieger, Kristinn R. Thórisson, and Pei Wang)
  [AGI-15 Berlin, July 2015]
  Education is the most important way to make AGI safe
• What Should AGI Learn From AI & CogSci? (by Pei Wang, Bas R. Steunebrink, and Kristinn R. Thórisson)
  [AGI-14 Special Session on AGI and Cognitive Science, Quebec City, Canada, August 2014]
  What makes AGI different from the related fields
• Formal Models in AGI Research
  [Workshop in AGI-13, Beijing, China, August 2013]
  Why mathematical logic and computation theory cannot provide a foundation for AGI
• Motivation Management in AGI Systems
  [Proceedings of AGI-12, Oxford, UK, December 2012]
  Motivations are persistent, spontaneous, mutually restricting, and changing over time
• Theories of Artificial Intelligence: Meta-theoretical considerations
  [Chapter 16 of Theoretical Foundations of Artificial General Intelligence]
  A good AGI theory should be correct, concrete, and compact.
• Rationality-guided AGI as cognitive systems (by Ahmed Abdel-Fattah, Tarek R. Besold, Helmar Gust, Ulf Krumnack, Martin Schmidt, Kai-Uwe Kuhnberger, and Pei Wang)
  [Proceedings of CogSci 2012, Pages 1242-1247, Sapporo, Japan, August 2012]
  New models of rationality and their relation with AGI and cognitive science
• The Evaluation of AGI Systems
  [Proceedings of AGI-10, Lugano, Switzerland, March 2010]
  Evaluation and meta-evaluation, empirical vs. theoretical
• Insufficient knowledge and resources: a biological constraint and its functional implications
  [Papers from AAAI 2009 Fall Symposium on Biologically Inspired Cognitive Architectures, Pages 88-93, Arlington, Virginia, November 2009. An extended version is published as a journal article in 2011.]
  The assumption on insufficient knowledge and resources is crucial for AI
• Suggested Education for Future AGI Researchers
  [On-line document since 2008]
  The background knowledge needed for AGI research, a personal view
• What Do You Mean by "AI"? [presentation]
  [Proceedings of AGI-08, Pages 362-373, Memphis, Tennessee, March 2008]
  Analysis and comparison of five typical ways to define AI
• Artificial General Intelligence — A Gentle Introduction
  [On-line document since 2007]
  AGI: theoretical problems and representative answers
• Aspects of Artificial General Intelligence (by Pei Wang and Ben Goertzel)
  [Introduction chapter of Advances in Artificial General Intelligence: Concepts, Architectures and Algorithms, IOS Press, 2007]
  Clarification and justification of AGI research in general
• Three Fundamental Misconceptions of Artificial Intelligence
  [Journal of Experimental & Theoretical Artificial Intelligence, 19(3), 249-268, 2007]
  It is a mistake to always take an AI system as an axiomatic system, a Turing machine, or a system with a model-theoretic semantics
• Artificial General Intelligence and Classical Neural Network
  [The IEEE International Conference on Granular Computing, 130-135, Atlanta, Georgia, May 2006]
  The strength and weakness of neural networks as general-purpose intelligent systems
• Artificial Intelligence: What it is, and what it should be
  [The AAAI Spring Symposium on Cognitive Science Principles Meet AI-Hard Problems, 97-102, Stanford, California, March 2006]
  On the identity and methodology of AI
• On the Working Definition of Intelligence
  [Technical Report No. 94 of CRCC, 1994. The on-line version is a revision finished in 1995.]
  The general philosophical issues of artificial intelligence

Logic and Reasoning

• Comparative Reasoning for Intelligent Agents (by Patrick Hammer, Peter Isaev, Hugo Latapie, Francesco Lanza, Antonio Chella and Pei Wang)
  [AGI-23 Stockholm, June 2023]
  Comparative reasoning on measurements, comparitive relations, and relatively defined concepts
• Toward a Logic of Everyday Reasoning
  [In Blended Cognition: The Robotic Challenge Springer, 2019]
  Some basic issues in logic (draft version)
• Assumptions of decision-making models in AGI (by Pei Wang and Patrick Hammer)
  [AGI-15 Berlin, July 2015]
  NARS vs. decision theory and reinforcement learning
• Issues in Temporal and Causal Inference (by Pei Wang and Patrick Hammer)
  [AGI-15 Berlin, July 2015]
  Temporal inference and causal inference in NARS
• Analogy in a General-Purpose Reasoning System
  [Cognitive Systems Research, 10(3), 286-296, 2009]
  Comparing the analogy in NARS with that in Copycat and SME
• Cognitive Logic versus Mathematical Logic
  [The Third International Seminar on Logic and Cognition, Guangzhou, May 2004]
  The logic of mathematics is not the logic of cognition
• The Generation and Evaluation of Generic Sentences
  [Philosophical Trends, Supplement 2004, 35-44]
  Use NARS to handle generic sentences
• Wason's Cards: What is Wrong?
  [The Third International Conference on Cognitive Science, 371-375, Beijing, August 2001]
  A comparison of NARS and traditional logic in terms of their conception of "evidence"
• Abduction in Non-Axiomatic Logic
  [The IJCAI Workshop on Abductive Reasoning, 56-63, Seattle, Washington, August 2001]
  Introducing Higher-Order Non-Axiomatic Logic, and comparing it with other approaches on abduction
• Unified Inference in Extended Syllogism
  [Abduction and Induction, 117-129, Kluwer Academic Pub, 2000]
  A unified formalization of deduction, induction, abduction, and revision as extended syllogism
• A New Approach for Induction: From a Non-Axiomatic Logical Point of View
  [Philosophy, Logic, and Artificial Intelligence, 53-85, Zhongshan University Press, 1999]
  The induction capacity of NARS
• From Inheritance Relation to Non-Axiomatic Logic
  [International Journal of Approximate Reasoning, 11(4), 281-319, 1994]
  A detailed description of the logical kernel of NARS

Uncertainty

• Issues in Applying Probability Theory to AGI
  [Workshop in AGI-13, Beijing, China, August 2013]
  Why probability theory cannot provide a foundation for AGI
• On the validity of Dempster-Shafer theory (by Jean Dezert, Pei Wang, and Albena Tchamova)
  [Proceedings of Fusion 2012, Singapore, July 2012]
  Dempster rule is conjunctive, while evidence combination should be additive
• Formalization of Evidence: A Comparative Study
  [Journal of Artificial General Intelligence, 1, 25-53, 2009]
  It takes two numbers to properly measure evidential support for a belief
• The Limitation of Bayesianism
  [Artificial Intelligence, 158(1), 97-106, 2004]
  Bayesianism has no general rule to revise beliefs
• Confidence as Higher-Order Uncertainty
  [The Second International Symposium on Imprecise Probabilities and Their Applications, 352-361, Ithaca, New York, June 2001]
  A discussion about the confidence measurement defined in NARS, and its relation with probability-based approaches
• Heuristics and Normative Models
  [International Journal of Approximate Reasoning, 14(4), 221-235, 1996]
  How NARS can reproduce various "heuristics and biases" observed in human reasoning
• The Interpretation of Fuzziness
  [IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, 26(2), 321-326, 1996]
  NARS vs. fuzzy logic
• A Unified Treatment of Uncertainties
  [The Fourth International Conference for Young Computer Scientists, 462-467, Beijing, July 1995]
  A general description about the uncertainty representation in NARS, including brief comparisons with other approaches
• Reference Classes and Multiple Inheritances
  [A revised version appears in International Journal of Uncertainty, Fuzziness and Knowledge-based Systems, 3(1), 79-91, 1995]
  NARS vs. non-monotonic logics and probability theory
• A Defect in Dempster-Shafer Theory
  [The Tenth Conference of Uncertainty in Artificial Intelligence, 560-566, Seattle, WA, July 1994]
  NARS vs. D-S theory
• Belief Revision in Probability Theory
  [The Ninth Conference of Uncertainty in Artificial Intelligence, 519-526, Washington DC, July 1993]
  NARS vs. the Bayesian approach
  

Meaning and Truth

• The frame problem, the relevance problem, and a package solution to both (by Yingjin Xu and Pei Wang)
  [Synthese, DOI: 10.1007/s11229-012-0117-8, 2012]
  How NARS handles semantic relevance
• Embodiment: Does a laptop have a body?
  [Proceedings of AGI-09, Pages 174-179, Arlington, Virginia, March 2009]
  Being "embodiment" means to take experience into account
• Experience-Grounded Semantics: A theory for intelligent systems
  [Cognitive Systems Research, 6(4), 282-302, 2005] 
  To define "truth" and "meaning" according to experience

Categorization and Learning

• A Unified Model of Reasoning and Learning
  [Proceedings of the Second International Workshop on Self-Supervised Learning, PMLR 159:28-48, 2021]
  In NARS, reasoning, learning, and many other cognitive functions are aspects of the same underlying process
  
• Self in NARS, an AGI System (by Pei Wang, Xiang Li, and Patrick Hammer)
  [Frontiers in Robotics and AI, Volume 5, Article 20, March 2018]
  An extended version of the AGI-17 paper, with additional discussion on self-organization and consciousness
  
• Self-Awareness and Self-Control in NARS (by Pei Wang, Xiang Li, and Patrick Hammer)
  [AGI-17, Melbourne, August 2017]
  A "self" concept and the related beliefs and operations in NARS are partly innate and partly learned
  
• Different Conceptions of Learning: Function Approximation vs. Self-Organization (by Pei Wang and Xiang Li)
  [AGI-16, New York City, July 2016]
  Learning in NARS, compared to mainstream machine learning
  
• The Logic of Categorization
  [The Fifteenth FLAIRS Conference, 181-185, Pensacola, Florida, May 2002]
  A discussion of the categorization model in NARS, which is integrated with reasoning and learning
• The Logic of Learning
  [The AAAI workshop on New Research Problems for Machine Learning, 37-40, Austin, Texas, July 2000]
  A comparison of inference-based learning and algorithm-based learning 
• Comparing Categorization Models: A psychological experiment
  [Technical Report No. 79 of CRCC, 1993]
  Comparisons of NARS with some categorization models proposed by psychologists.

Resource Management

• An Architecture for Real-time Reasoning and Learning (by Pei Wang, Patrick Hammer, and Hongzheng Wang)
  [AGI-20, Online, June 2020]
  For an AGI to work in "real time": definition, necessity, and design
  
• The Emotional Mechanisms in NARS (by Pei Wang, Max Talanov, and Patrick Hammer)
  [AGI-16, New York City, July 2016]
  The emotion mechanism in NARS
  
• Solving a Problem With or Without a Program
  [Journal of Artificial General Intelligence, 3(3):43-73, 2013]
  Three ways to solve a problem: how can a computer be creative
• Case-by-case problem solving
  [Proceedings of AGI-09, Pages 180-185, Arlington, Virginia, March 2009]
  Directly solving a problem instance using available knowledge and resources
• Computation and Intelligence in Problem Solving
  [A revised version appears as "Problem solving with insufficient resources" in the International Journal of Uncertainty, Fuzziness and Knowledge-based Systems, 12(5), 673-700, 2004]
  How to solve problems without problem-specific algorithms

Experience and Interaction

• A Model of Unified Perception and Cognition (by Pei Wang, Christian Hahm and Patrick Hammer)
  [Frontiers in Artificial Intelligence, Vol. 11, DOI: 10.3389/frai.2022.806403, 2022]
  Arguments for the NARS approach toward perception and cognition
• Perception from an AGI Perspective (By Pei Wang and Patrick Hammer)
  [Proceedings of AGI-18, Prague, Czech, August 2018]
  Perception in AGI should be subjective, active, and unified with other cognitive processes
• Natural Language Processing by Reasoning and Learning
  [Proceedings of AGI-13, Beijing, China, August 2013]
  Some preliminary ideas and results of natural language processing in NARS

Application

• Neurosymbolic hybrid approach to driver collision warning (Kyongsik Yun, Thomas Lu, Alexander Huyen, Patrick Hammer, and Pei Wang)
  [Proc. SPIE 12101, Pattern Recognition and Tracking XXXIII, 2022]
  NARS is used with other techniques in a project of NASA/JPL
• A Metamodel and Framework for Artificial General Intelligence From Theory to Practice (by Hugo Latapie, Ozkan Kilic, Gaowen Liu, Ramana Kompella, Adam Lawrence, Yuhong Sun, Jayanth Srinivasa, Yan Yan, Pei Wang, Kristinn R. Thorisson)
  [Journal of Artificial Intelligence and Consciousness, Vol. 8, No. 2, Pages 205-227, 2021]
  NARS is used with other techniques in various projects of Cisco
• A Reasoning Based Model for Anomaly Detection in the Smart City Domain (By Patrick Hammer, Tony Lofthouse, Enzo Fenoglio, Hugo Latapie, and Pei Wang)
  [Proceedings of IntelliSys 2020, Pages 144-159, Online, September 2020]
  NARS+DL in a cooperative project with Cisco
• A prototype intelligent decision-support system with a unified planning and learning capabilities (by Nady Slam, Wushour Slamu, Pei Wang)
  [International Journal on Artificial Intelligence Tools, Vol. 26, No. 6, 1750025 (22 pages), 2017]
  Applying NAL in decision support
• On the logic of percutaneous coronary interventions (by Peter Lanzer and Pei Wang)
  [International Journal of Clinical Cardiology, Vol.2, No.3, 2015]
  Applying NAL in medical decision making
• Natural Language Processing by Reasoning and Learning
  [Proceedings of AGI-13, Beijing, China, August 2013]
  Some preliminary ideas and results of natural language processing in NARS
• Reasoning in Non-Axiomatic Logic: A Case Study in Medical Diagnosis (by Pei Wang and Seemal Awan)
  [Proceedings of AGI-11, 297-302, Mountain View, California, August 2011]
  Testing the expressive and inferential power of NAL in medical diagnosis
• Recommendation Based on Personal Preference
  [Computational Web Intelligence: Intelligent Technology for Web Applications, 101-115, World Scientific Publishing Company, 2004]
  Applying the evidence measurement of NARS in recommendation and on-line shopping