Cortese @ Stanford University Computer Science Department & aaai March 25, 2014
Stanford University

(Stanford University, Palo Alto, California - March 25 2014)

IMPLEMENTING SELVES WITH SAFE MOTIVATIONAL SYSTEMS AND SELF-IMPROVEMENT

While artificial (general) intelligence most often focuses on tools/systems for collecting knowledge and solving problems or achieving goals rather than self-reflecting entities, this implementation-oriented symposium will focus instead on guided self-creation and improvement — particularly as a method of achieving human-level intelligence in machines through iterative improvement ("seed AI").

In I Am a Strange Loop, Douglas Hofstadter argues that the key to understanding selves is the "strange loop", a complex feedback network inhabiting our brains and, arguably, constituting our minds. Further, humans have both conscious and unconscious minds, attention, emotions, partial self-reflection, a moral sense and many other aspects that are rarely addressed — yet seem critical for the creation of a safe self-sufficient autonomous system. This symposium will focus on the integration of these components into a coherent self-improving self. Ideally, the ultimate end result will be a successful entity with extensive self-knowledge and a safe or moral/ethical motivational system that functions with discrimination to promote cooperation with and contribution to community via iterative improvement of self, tools and theoretical constructs of relational dynamics and resource utilization/allocation/sharing.

Topics

Potential topics include (but aren't limited to) the following:

  • Integrative architectures w/explicit motivations implementing "self" as
  • Operating system with "plug-ins"
  • Society of mind (minsky)/economy of idiots (baum)
  • Global workspace/consciousness (baars/franklin)
  • Authorship (dennett/wegner)
  • Safe/moral/ethical motivational systems
  • Value sets versus goal hierarchies
  • "safe"/moral values/goal content
  • Evaluation schemes
  • Reflection
  • Self-examination
  • Self-modeling and self-knowledge
  • Goal-based self-evaluation for self-improvement
  • Attention and emotions
  • As knowledge/rules of thumb/"actionable qualia"
  • As (un)helpful biases (and intelligent improvement)
  • As evaluation/enforcement mechanisms
  • Integrating different knowledge/action representation schemes
  • Coordination and translation between schemes
  • Analyzing trade-offs/knowing when to switch
  • Self-improvement
  • Via automated tool/method incorporation and theory-inductive heuristics
  • Via learning/knowledge incorporation
  • Discovery (refactoring, modularization, encapsulation and scale-invariance)

While solutions need to be grounded and extensible, approaches starting with some initial structure rather than a tabula rasa with the lowest level bootstrapping approaches or first causes explanations (except where they are fully extended to initial structures and/or used to justify such structures) are preferred. While autopoiesis and "functional consciousness" are obviously key topics, phenomenal consciousness is, preferably, off-topic.

Primary Contact Mark Waser (Digital Wisdom Institute, MWaser@DigitalWisdomInstitute.org)

For More Information

For more information, please consult the supplemental symposium website.