All Tags
AWS
algorithm-design
architecture
cloud-principles
cost-reduction
data-centric
data-compression
data-processing
deployment
design
edge-computing
energy-footprint
hardware
libraries
locality
machine-learning
management
measured
migration
model-optimization
model-training
performance
queries
rebuilding
scaling
services
strategies
template
workloads
Tactic: Use Data Projection
Tactic sort:
Awesome Tactic
Type: Architectural Tactic
Category: green-ml-enabled-systems
Title
Use Data Projection
Description
Data projection means transforming data into a lower-dimensional embedding and using data to optimize projection parameters. Reducing the dimensionality of input data shrinks the dimensionality of the deep neural network (DNN), which leads to improved performance of the model. Using data projection as a preprocessing step can result in energy improvements without sacrificing performance or accuracy.
Participant
Data Scientist
Related software artifact
Data
Context
Machine Learning
Software feature
< unknown >
Tactic intent
Improve performance (and consequently energy efficiency) by projecting data into lower-dimensional embeddings
Target quality attribute
Performance
Other related quality attributes
Accuracy, Energy Efficiency
Measured impact
< unknown >
Source
Bita Darvish Rouhani, Azalia Mirhoseini, and Farinaz Koushanfar. 2016. DeLight: Adding Energy Dimension To Deep Neural Networks. In ISLPED ’16: Proceedings of the 2016 International Symposium on Low Power Electronics and Design. Association for Computing Machinery, New York, NY, USA, 112–117. (DOI: https://doi.org/10.1145/2934583.2934599)Graphical representation
![]( /AwesomeAndDarkTactics/assets/diagrams/use-data-projection.png)