Towards Sustainability of AI

Appearance

Promote Data Sparseness in Data Collection ​

Overview ​

Sustainability DimensionEcological
ML Development PhaseData Collection and Preparation
ML Development StakeholdersBusiness Stakeholder, Domain Expert, ML Development

Description ​

First, “Promote Data Sparseness in Data Collection” describes the tradeoff between the collection of more data points (e.g., through additional sensors or external server calls), thus increasing the CO2-eq-Footprint and the performance increases associated with this data point (Schneider et al., 2019). This can be achieved by gauging the performance increase of each data point before designing the acquisition (Yu, 2017). On a technical level, the spareness can be embraced by using efficient data collection algorithms (Rohankar et al., 2015; Xiang et al., 2013).

Sources ​

  • Schneider, J., Basalla, M., & Seidel, S. (2019). Principles of Green Data Mining. Proceedings of the 52nd Hawaii International Conference on System Sciences. Hawaii International Conference on System Sciences. https://doi.org/10.24251/HICSS.2019.250
  • Rohankar, R., Katti, C. P., & Kumar, S. (2015). Comparison of Energy Efficient Data Collection Techniques in Wireless Sensor Network. Procedia Computer Science, 57, 146–151. https://doi.org/10.1016/j.procs.2015.07.399
  • Xiang, L., Luo, J., & Rosenberg, C. (2013). Compressed Data Aggregation: Energy-Efficient and High-Fidelity Data Collection. IEEE/ACM Transactions on Networking, 21(6), 1722–1735. https://doi.org/10.1109/TNET.2012.2229716