Session
The Next Decade of Software Is About Climate - What Is the Role of ML?
Please note: there will not be a live Q&A for this session.
Climate action has risen to the top of many technology companies' agenda. But how does one get started? Green software engineering is an emerging discipline and being a part of the climate change solution is a relatively new part of many software companies' strategy. With accuracy hungry algorithms ML software is consuming more and more computational resources, largely benefiting from the increasingly better hardware. Are the results worth the environmental cost? This talk introduces the field of green software engineering, showing options to estimate the carbon footprint and discuss ideas on how to make Machine Learning greener, giving you the tools to take an active part in the climate solution.
Main Takeaways
1 Hear about green software engineering, how software impacts climate change.
2 Learn how ML can be made greener.
What's the focus of your work these days?
I'm the chair of the writers group in the Green Software Foundation, where we are working to create a trusted ecosystem of people, standards, tools and best practices for green software. We have a 26 member organization and over 700 individual members. One part of my work is making sure that the carbon footprint decreases, that we are part of the solution. And then I'm also a software engineer at Microsoft, as backend engineer on Microsoft 365 products. I work to incorporate the characteristics of green software into Microsoft 365 products as well.
What's the motivation for your talk?
This is where things get dire, rapid climate change. As an industry we are growing fast. We used to have a small amount of the global carbon footprint, but it is increasing and there are trends saying that by 2040, we will account for 14 percent of the global emissions, which is half the size of the transportation sector. Which is ironic because we talk a lot about transportation, don't fly, buy local, all these things, but with software we don't even talk about it. And we are past the time to talk about it. We need to act on a massive scale.
How would you describe the persona and level of your target audience for this talk?
This will be in the machine learning track, so you should have some kind of relationship with machine learning, if you're a data scientist or if you're a person who's working in a machine learning team as a software engineer or a non-technical role as well. And if you're curious about green software and if you want to learn more about this very nascent area
What would you like those who attend your session to walk away with?
I want them to understand that in order to meaningfully solve the climate problem, we need massive action at massive scale. This is not the time where we can improve our tech by one, two percent and then call it the day. We need to start having this conversation at an entirely different scale.
Speaker
Sara Bergman
Software Engineer at the Green Software Foundation
Sara is a Software Engineer at Microsoft Development Center Norway working in a team which owns several backend APIs powering people experiences in the Microsoft eco-system. She is an advocate for sustainable software practices at MDCN and M365. She is a member of the Green Software Foundation...
Read more
From the same track
How to Operationalize Transformer Models on the Edge
Thursday May 12 / 10:00AM EDT
Sometimes the hardest part of building ML solutions is deployment. In this session, we will look at different model deployment architectures, how to deploy with edge devices and inference in different programming languages. By the end of the demo-heavy session, you will see how to optimize large...
Cassie Breviu
Senior Program Manager @Microsoft
Open Machine Learning
Thursday May 12 / 11:10AM EDT
In this talk, Omar will talk about trends in the ML ecosystem for Open Science and Open Source. Omar will talk about the power of creating interactive demos using Open Source libraries, BigScience, a one-year long research workshop involving over 700 researchers, and other community-led efforts...
Omar Sanseviero
Machine Learning Engineer @huggingface
Operationalizing Responsible AI in Practice
Thursday May 12 / 01:30PM EDT
Enabling responsible development of artificial intelligent technologies is one of the major challenges we face as the field moves from research to practice. Researchers and practitioners from different disciplines have highlighted the ethical and legal challenges posed by the use of machine...
Mehrnoosh Sameki
Senior Technical Program Manager & Tech Lead @Microsoft