A STDP-based Encoding Algorithm for Associative and Composite Data

TL;DR

This paper introduces a practical memory model based on STDP that encodes and retrieves high-dimensional associative and composite data, demonstrating effective auto-associative and semantic memory tasks.

Contribution

It presents a novel STDP-based encoding scheme capable of handling multiple composite data types in a continuous, associative memory framework.

Findings

Successful retrieval of streamed images with cues in auto-associative tasks.

Words can recall multiple or single sentences based on grammatical relations.

Model effectively encodes and retrieves semantic memories from sentences.

Abstract

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.