Symbolic Anomaly Detection and Assessment Using Growing Neural Gas

Metacognitive architectures provide one solution to the brittleness problem for agents operating in complex, changing environments. The Metacognitive Loop, in which a system notes an anomaly, assesses the problem and guides a solution, is one form of such an architecture. This paper extends prior work on implementing the note phase of this process in symbolic planning domains using the A-distance. This extension uses a growing neural gas algorithm to construct a network which represents various normal and anomalous states. Testing shows that this technique allows for improved detection of anomalies in the note phase as well as categorization of anomalies by severity and type in the assess phase.

[1]  VARUN CHANDOLA,et al.  Anomaly detection: A survey , 2009, CSUR.

[2]  Shai Ben-David,et al.  Detecting Change in Data Streams , 2004, VLDB.

[3]  Paul A. Crook,et al.  A Robot Implementation of a Biologically Inspired Method for Novelty Detection , 2002 .

[4]  Donald Perlis,et al.  Logic, Self-awareness and Self-improvement: the Metacognitive Loop and the Problem of Brittleness , 2005, J. Log. Comput..

[5]  Tim Oates,et al.  The metacognitive loop I: Enhancing reinforcement learning with metacognitive monitoring and control for improved perturbation tolerance , 2006, J. Exp. Theor. Artif. Intell..

[6]  Michael T. Cox.  Perpetual Self-Aware Cognitive Agents , 2007, AI Mag..

[7]  David W. Aha,et al.  GOAL‐DRIVEN AUTONOMY FOR RESPONDING TO UNEXPECTED EVENTS IN STRATEGY SIMULATIONS , 2013, Comput. Intell..

[8]  Florian Metze,et al.  Generalized radial basis function networks for classification and novelty detection: self-organization of optimal Bayesian decision , 2000, Neural Networks.

[9]  Tim Oates,et al.  We’re Not in Kansas Anymore: Detecting Domain Changes in Streams , 2010, EMNLP.

[10]  Donald Perlis,et al.  Data-Driven Goal Generation for Integrated Cognitive Systems , 2013, AAAI Fall Symposia.

[11]  Darsana P. Josyula,et al.  Application of MCL in a dialog agent , .

[12]  Michael M. Cox,et al.  Noting Anomalies in Streams of Symbolic Predicates Using A-Distance , 2012 .

[13]  John Blitzer,et al.  Domain Adaptation with Structural Correspondence Learning , 2006, EMNLP.

[14]  Tim Oates,et al.  The Integration of Cognitive and Metacognitive Processes with Data-driven and Knowledge-rich Structures , 2013 .

[15]  Tim Oates,et al.  The Metacognitive Loop and Reasoning about Anomalies , 2011, Metareasoning.

[16]  Thomas Villmann,et al.  Supervised Neural Gas with General Similarity Measure , 2005, Neural Processing Letters.

[17]  Tom Fawcett,et al.  Activity monitoring: noticing interesting changes in behavior , 1999, KDD '99.

[18]  Michèle Basseville,et al.  Detection of abrupt changes: theory and application , 1993 .

[19]  Bernd Fritzke,et al.  A Growing Neural Gas Network Learns Topologies , 1994, NIPS.

[20]  Barry Smyth,et al.  Retrieval, reuse, revision and retention in case-based reasoning , 2005, The Knowledge Engineering Review.

[21]  Manuela Veloso.  Learning by analogical reasoning in general problem-solving , 1992 .

[22]  Helen Ashman,et al.  Anomaly Detection over User Profiles for Intrusion Detection , 2010 .

[23]  John Blitzer,et al.  Biographies, Bollywood, Boom-boxes and Blenders: Domain Adaptation for Sentiment Classification , 2007, ACL.