Transfer Learning through Analogy in Games

Abstract

Articles70 AI MAGAZINE E ffectively transferring previously learned knowledge to a new domain is one of the hallmarks of human intelligence. This is the objective of transfer learning, in which transferred knowledge guides the learning process in a broad range of new situations. In near transfer, the source and target domains are very similar and solutions can be transferred almost verbatim. In far transfer, the domains may appear quite different and the knowledge to be transferred involves deeper shared abstractions. Thus transfer learning can be viewed as a compromise between the delayed generalization of instance-or casebased reasoning and the early generalization of classical machine learning.We have explored the use of analogy as a general approach to near and far transfer learning in domains ranging from physics problem solving to strategy games (Klenk and Forbus 2007;Hinrichs and Forbus 2007). Using the same basic analogical mechanism, we have found that the main differences between near and far transfer involve the amount of generalization that must be performed prior to transfer and the way that the matching process treats nonidentical predicates. We present here two extensions of our analogical matcher, minimal ascension and metamapping, that enable far transfer between representations with different relational vocabulary. Evidence for the effectiveness of these techniques is provided by a large-scale external evaluation, involving a substantial number of novel distant analogs.

Theory-Based Causal Transfer:Integrating Instance-Level Induction and Abstract-Level Structure Learning

“…More complex theories could potentially account for delayed effects; e.g., when an agent could not find a causal attribute for a particular event, the agent could examine attributes jointly to best explain the causal effect observed. Prior work has examined structural analogies (Hinrichs and Forbus 2011;Zhang et al 2019a;2019b) and object mappings (Fitzgerald, Goel, and Thomaz 2018) to facilitate transfer; these may also be useful to acquire transferable causal knowledge.…”

The architecture challenge: Future artificial-intelligence systems will require sophisticated architectures, and knowledge of the brain might guide their construction

“…For example, the Companion cognitive architecture can use rich relational representations and analogy to perform distant transfer. Learning games with a previously learned analogous game led to more rapid learning than learning without such an analog (Hinrichs & Forbus 2011). This and many other experiments suggest that analogy not only can explain human transfer learning, but also can provide new techniques for machine learning.…”

Evidence from machines that learn and think like people

“…For example, the Companion cognitive architecture can use rich relational representations and analogy to perform distant transfer. Learning games with a previously learned analogous game led to more rapid learning than learning without such an analog (Hinrichs & Forbus 2011). This and many other experiments suggest that analogy not only can explain human transfer learning, but also can provide new techniques for machine learning.…”