MASR: A Modular Accelerator for Sparse RNNs

Recurrent neural networks (RNNs) are becoming the de-facto solution for speech recognition. RNNs exploit long-term temporal relationships in data by applying repeated, learned transformations. Unlike fully-connected (FC) layers with single vector matrix operations, RNN layers consist of hundreds of...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Proceedings / International Conference on Parallel Architectures and Compilation Techniques s. 1 - 14
Hlavní autori: Gupta, Udit, Reagen, Brandon, Pentecost, Lillian, Donato, Marco, Tambe, Thierry, Rush, Alexander M., Wei, Gu-Yeon, Brooks, David
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 01.09.2019
Predmet:
Acceleration
Accelerator
automatic speech recognition
Computer architecture
deep neural network
Encoding
Hardware
Hidden Markov models
Load management
Recurrent neural networks
sparsity
ISSN:2641-7936
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Recurrent neural networks (RNNs) are becoming the de-facto solution for speech recognition. RNNs exploit long-term temporal relationships in data by applying repeated, learned transformations. Unlike fully-connected (FC) layers with single vector matrix operations, RNN layers consist of hundreds of such operations chained over time. This poses challenges unique to RNNs that are not found in convolutional neural networks(CNNs) or FC models, namely large dynamic activation. In this paper we present MASR, a principled and modular architecture that accelerates bidirectional RNNs for on-chip ASR. MASR is designed to exploit sparsity in both dynamic activations and static weights. The architecture is enhanced by a series of dynamic activation optimizations that enable compact storage, ensure no energy is wasted computing null operations, and maintain high MAC utilization for highly parallel accelerator designs. In comparison to current state-of-the-art sparse neural network accelerators (e.g., EIE), MASR provides 2×area 3×energy, and 1.6×performance benefits. The modular nature of MASR enables designs that efficiently scale from resource-constrained low-power IoT applications to large-scale, highly parallel datacenter deployments.
ISSN:2641-7936
DOI:10.1109/PACT.2019.00009