Latent Time-Adaptive Drift-Diffusion Model

TL;DR

The paper introduces the LTDDM, a model that learns timing tasks faster than traditional recurrent neural networks, aligning with animal learning behaviors and outperforming LSTMs in speed.

Contribution

The LTDDM extends the TDDM to better replicate animal timing behavior and demonstrates significantly faster learning compared to LSTM neural networks.

Findings

LTDDM learns timing tasks much faster than LSTMs.

LTDDM exhibits behavioral properties consistent with animal data.

LTDDM outperforms LSTMs across multiple timing tasks.

Abstract

Animals can quickly learn the timing of events with fixed intervals and their rate of acquisition does not depend on the length of the interval. In contrast, recurrent neural networks that use gradient based learning have difficulty predicting the timing of events that depend on stimulus that occurred long ago. We present the latent time-adaptive drift-diffusion model (LTDDM), an extension to the time-adaptive drift-diffusion model (TDDM), a model for animal learning of timing that exhibits behavioural properties consistent with experimental data from animals. The performance of LTDDM is compared to that of a state of the art long short-term memory (LSTM) recurrent neural network across three timing tasks. Differences in the relative performance of these two models is discussed and it is shown how LTDDM can learn these events time series orders of magnitude faster than recurrent neural networks.