English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Humans have a remarkably large capacity to store detailed visual informationin long-term memory even after a single exposure, as demonstrated by classicexperiments in psychology. For example, Standing (1973) showed that humanscould recognize with high accuracy thousands of pictures that they had seenonly once a few days prior to a recognition test. In deep learning, the primarymode of incorporating new information into a model is through gradient descentin the model's parameter space. This paper asks whether deep learning viagradient descent can match the efficiency of human visual long-term memory toincorporate new information in a rigorous, head-to-head, quantitativecomparison. We answer this in the negative: even in the best case, modelslearning via gradient descent require approximately 10 exposures to the samevisual materials in order to reach a recognition memory performance humansachieve after only a single exposure. Prior knowledge induced via pretrainingand bigger model sizes improve performance, but these improvements are not veryvisible after a single exposure (it takes a few exposures for the improvementsto become apparent), suggesting that simply scaling up the pretraining datasize or model size might not be a feasible strategy to reach human-level memoryefficiency.

Can deep learning match the efficiency of human visual long-term memory in storing object details?