Neuromodulation Gated Transformer

05/05/2023
by   Kobe Knowles, et al.
0

We introduce a novel architecture, the Neuromodulation Gated Transformer (NGT), which is a simple implementation of neuromodulation in transformers via a multiplicative effect. We compare it to baselines and show that it results in the best average performance on the SuperGLUE benchmark validation sets.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
11/06/2021

Convolutional Gated MLP: Combining Convolutions gMLP

To the best of our knowledge, this is the first paper to introduce Convo...
research
05/12/2022

Supplementary Material: Implementation and Experiments for GAU-based Model

In February this year Google proposed a new Transformer variant called F...
research
01/31/2021

Classification Models for Partially Ordered Sequences

Many models such as Long Short Term Memory (LSTMs), Gated Recurrent Unit...
research
04/20/2023

An Introduction to Transformers

The transformer is a neural network component that can be used to learn ...
research
03/01/2023

Building High-accuracy Multilingual ASR with Gated Language Experts and Curriculum Training

We propose gated language experts to improve multilingual transformer tr...
research
10/26/2021

Geometric Transformer for End-to-End Molecule Properties Prediction

Transformers have become methods of choice in many applications thanks t...
research
11/25/2022

The Naughtyformer: A Transformer Understands Offensive Humor

Jokes are intentionally written to be funny, but not all jokes are creat...

Please sign up or login with your details

Forgot password? Click here to reset