Large Patterns Make Great Symbols: An Example of Learning from Example

We look at distributed representation of structure with variable binding, that is natural for neural nets and allows traditional symbolic representation and processing. The representation supports learning from example. This is demonstrated by taking several instances of the mother-of relation implying the parent-of relation, by encoding them into a mapping vector, and by showing that the mapping vector maps new instances of mother-of into parent-of. 1 Introduction Distributed representation is used commonly with neural nets, as it is in ordinary computers, to encode a large number of attributes or things with a much smaller number of variables or units. In this paper we assume that the units are binary so that the encodings of things are bit patterns or bit vectors (`pattern' and `vector' will be used interchangeably in this paper). In normal symbolic processing the bit patterns are arbitrary identifiers, and it matters only whether two patterns are identical or different, whereas b..