A STDP-based Encoding Algorithm for Associative and Composite Data
This work addresses memory encoding and retrieval for associative and composite data, which is an incremental advancement in neuromorphic computing.
The paper tackles the problem of storing and retrieving high-dimensional associative data by proposing a practical memory model based on spike-timing-dependent plasticity (STDP). The results show successful retrieval of images in an auto-associative memory task and flexible recall of sentences based on grammatical relations in a semantic memory task.
Spike-timing-dependent plasticity(STDP) is a biological process of synaptic modification caused by the difference of firing order and timing between neurons. One of the neurodynamical roles of STDP is to form a macroscopic geometrical structure in the neuronal state space in response to a periodic input. This work proposes a practical memory model based on STDP that can store and retrieve high-dimensional associative data. The model combines STDP dynamics with an encoding scheme for distributed representations and can handle multiple composite data in a continuous manner. In the auto-associative memory task where a group of images is continuously streamed to the model, the images are successfully retrieved from an oscillating neural state whenever a proper cue is given. In the second task that deals with semantic memories embedded from sentences, the results show that words can recall multiple sentences simultaneously or one exclusively, depending on their grammatical relations.