| 摘要 |
A biologically-inspired model for sequence representation, method of construction and application of such models, and systems incorporating same are provided. The model captures the statistical nature of sequences and uses that for sequence encoding, recognition, and recall. The model can be trained in real time, has few tunable parameters, and is highly parallelizable, which ensures that it can scale up to very large problems. Applications of the model to word and speech recognition, machine leaning, robotics, computational bioinformatics, genetics datasets, and other sequence processing pipelines are provided. |
