From an environmental standpoint, there are a few crucial aspects of training a neural network that have a major impact on the quantity of carbon that it emits. These factors include: the location of the server used for training and the energy grid that it uses, the length of the training procedure, and even the make and model of hardware on which the training takes place. In order to approximate these emissions, we present our Machine Learning Emissions Calculator, a tool for our community to better understand the environmental impact of training ML models. We accompany this tool with an explanation of the factors cited above, as well as concrete actions that individual practitioners and organizations can take to mitigate their carbon emissions. * equal contribution Preprint. Under review.

The divide between attributional and consequential research perspectives partly overlaps with the long-standing methodological discussions in the lifecycle assessment (LCA) and input-output analysis (IO) research communities on the choice of techniques and models for dealing with situations of coproduction.The recent harmonization of LCA allocations and IO constructs revealed a more diverse set of coproduction models than had previously been understood. This increased flexibility and transparency in inventory modeling warrants a re-evaluation of the treatment of coproduction in analyses with attributional and consequential perspectives.In the present article, the main types of coproductions situations and of coproduction models are reviewed, along with key desirable characteristics of attributional and consequential studies. A concordance analysis leads to clear recommendations, which call for important refinements to current guidelines for both LCA/IO practitioners and database developers. We notably challenge the simple association between, on the one hand, attributional LCA and partition allocation, and on the one hand, consequential LCA and substitution modeling.

No abstract

No abstract

No abstract

No abstract

No abstract

No abstract