Features of Distributed Representations for Tree-structures: A Study of RAAM (HS-IDA-TR-95-001) Presented at the 2nd Swedish Conference on Connectionism
نویسندگان
چکیده
This paper presents an in-depth analysis of properties of patterns, generated by the Recursive Auto-Associative Memory, based on the idea that representational features can be detected by a classification network. The intension of this analysis is to examine the actual reasons for the success of connectionist processes acting on super-positional activity vectors generated in the fashion described. We show that the structure supplied during training is maintained and is extractable from the generated pattern. Further, we show that the influence of the actual constituents in the structures supplied during training is not necessarily available, in the generated patterns, for holistic processing. The outlook for holistic processing is therefore limited unless new forms can be found which take into account, what Sharkey and Jackson call, ‘whole net’ representations.
منابع مشابه
Tail-Recursive Distributed Representations and Simple Recurrent Networks
Representation poses important challenges to connectionism. The ability to structurally compose representations is critical in achieving the capability considered necessary for cognition. We are investigating distributed patterns that represent structure as part of a larger effort to develop a natural language processor. Recursive Auto-Associative Memory (RAAM) representations show unusual prom...
متن کاملLinear recursive distributed representations
Connectionist networks have been criticized for their inability to represent complex structures with systematicity. That is, while they can be trained to represent and manipulate complex objects made of several constituents, they generally fail to generalize to novel combinations of the same constituents. This paper presents a modification of Pollack's Recursive Auto-Associative Memory (RAAM), ...
متن کاملCan Connectionist Models Exhibit Non-Classical Structure Sensitivity? (HS-IDA-TR-94-002) Presented at The 16th Annual Conference of the Cognitive Science Society
Several connectionist models have been supplying non-classical explanations to the challenge of explaining systematicity, i.e., structure sensitive processes, without merely being implementations of classical architectures. However, lately the challenge has been extended to include learning related issues. It has been claimed that when these issues are taken into account, only a restricted form...
متن کاملDistributed Patterns as Hierarchical Structures
Recursive Auto-Associative Memory (RAAM) structures show promise as a general representation vehicle that uses distributed patterns. However training is often difficult, which explains, at least in part, why only relatively small networks have been studied. We show a technique for transforming any collection of hierarchical structures into a set of training patterns for a sequential RAAM which ...
متن کاملScaling Connectionist Compositional Representations
The Recursive Auto-Associative Memory (RAAM) has come to dominate connectionist investigations into representing compositional structure. Although an adequate model when dealing with limited data, the capacity of RAAM to scale-up to real-world tasks has been frequently questioned. RAAM networks are difficult to train (due to the moving target effect) and as such training times can be lengthy. I...
متن کامل