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

Fast approximations to matrix multiplication have the potential todramatically reduce the cost of neural network inference. Recent work onapproximate matrix multiplication proposed to replace costly multiplicationswith table-lookups by fitting a fast hash function from training data. In thiswork, we propose improvements to this previous work, targeted to the deeplearning inference setting, where one has access to both training data andfixed (already learned) model weight matrices. We further propose a fine-tuningprocedure for accelerating entire neural networks while minimizing loss inaccuracy. Finally, we analyze the proposed method on a simple imageclassification task. While we show improvements to prior work, overallclassification accuracy remains substantially diminished compared to exactmatrix multiplication. Our work, despite this negative result, points the waytowards future efforts to accelerate inner products with fast nonlinear hashingmethods.

Look-ups are not (yet) all you need for deep learning inference