Soar - Reflexive Learning

Description

• Soar stands in opposition to those architectures that take pains to decide what knowledge is worth learning and what is not. Unlike agents based on these types of architectures, SOAR agents have no control over what they learn through their built-in learning mechanism, chunking.
• SOAR uses the chunking mechanism to create productions that are stored in long-term memory every time an impasse is resolved.
• This mechanism can lead to the storage of useless chunks, but SOAR deals with this by matching production rules very efficiently.