When Can Self-Attention Be Replaced by Feed Forward Layers?
May 28, 2020 Β· Declared Dead Β· π arXiv.org
"No code URL or promise found in abstract"
Evidence collected by the PWNC Scanner
Authors
Shucong Zhang, Erfan Loweimi, Peter Bell, Steve Renals
arXiv ID
2005.13895
Category
eess.AS: Audio & Speech
Cross-listed
cs.CL,
cs.SD
Citations
2
Venue
arXiv.org
Last Checked
3 months ago
Abstract
Recently, self-attention models such as Transformers have given competitive results compared to recurrent neural network systems in speech recognition. The key factor for the outstanding performance of self-attention models is their ability to capture temporal relationships without being limited by the distance between two related events. However, we note that the range of the learned context progressively increases from the lower to upper self-attention layers, whilst acoustic events often happen within short time spans in a left-to-right order. This leads to a question: for speech recognition, is a global view of the entire sequence still important for the upper self-attention layers in the encoder of Transformers? To investigate this, we replace these self-attention layers with feed forward layers. In our speech recognition experiments (Wall Street Journal and Switchboard), we indeed observe an interesting result: replacing the upper self-attention layers in the encoder with feed forward layers leads to no performance drop, and even minor gains. Our experiments offer insights to how self-attention layers process the speech signal, leading to the conclusion that the lower self-attention layers of the encoder encode a sufficiently wide range of inputs, hence learning further contextual information in the upper layers is unnecessary.
Community Contributions
Found the code? Know the venue? Think something is wrong? Let us know!
π Similar Papers
In the same crypt β Audio & Speech
R.I.P.
π»
Ghosted
R.I.P.
π»
Ghosted
LPCNet: Improving Neural Speech Synthesis Through Linear Prediction
R.I.P.
π»
Ghosted
VoiceFilter: Targeted Voice Separation by Speaker-Conditioned Spectrogram Masking
R.I.P.
π»
Ghosted
TERA: Self-Supervised Learning of Transformer Encoder Representation for Speech
R.I.P.
π»
Ghosted
Mockingjay: Unsupervised Speech Representation Learning with Deep Bidirectional Transformer Encoders
R.I.P.
π»
Ghosted
Utterance-level Aggregation For Speaker Recognition In The Wild
Died the same way β π» Ghosted
R.I.P.
π»
Ghosted
Federated Learning: Strategies for Improving Communication Efficiency
R.I.P.
π»
Ghosted
In-Datacenter Performance Analysis of a Tensor Processing Unit
R.I.P.
π»
Ghosted
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning
R.I.P.
π»
Ghosted