Lunch Talk: Philipp Wiesner „Exploring the Trade-Off Between Carbon Footprint and User Experience of ML Inference Services“

During this Lunch Talk, Philipp Wiesner from the Distributed and Operating Systems research group at TU Berlin will dive into the important topic of sustainability and carbon footprint with regard to highly energy-intense Machine Learning inference services.

Abstract: Modern machine learning inference services, such as LLM serving, are highly energy-intensive and, hence, have a significant carbon footprint. However, such services can often be offered in varying levels of quality, such as different model sizes, quantized, or pruned versions. This talk explores the potential for dynamically adjusting service quality based on the current energy mix, to lower the reported Scope 2 emissions of service providers, as well as Scope 3 emissions reported by service users.

The BIFOLD Lunch Talk series gives BIFOLD members and external partners the opportunity to engage in dialogue about their research in Machine Learning and Big Data. Each Lunch Talk offers BIFOLD members, fellows and external researchers and guests the chance to present their research and to network with each other.
The Lunch Talk takes place at the TU Berlin. For further information on the Lunch Talks and registration, contact Dr. Laura Wollenweber via email.