Environmental impact of machine learning
18/01/25 09:40
Mitigating the Environmental Impact of Machine Learning is my new research team with SFU computer scientist Stephen Makonin (collaborator on Tackling the Caarbon Footprint of Streaming Media), Makonin’s graduate student Kehui Li, and AI artist and SCA PhD student Jess MacCormack. We are researching the environmental impact of machine learning, aka artificial intelligence, and ways to mitigate it, such as by developing models that use much less electricity. Funded by Arne Eigenfeldt (PI), Jim Bizzochi, and my SSHRC Insight Grant, Small-File Generative Art
Read More…
Read More…
low-impact music streaming
30/07/24 12:37
An article on low-impact music streaming, quoting me, in the MIT Technology Review:
https://www.technologyreview.com/2024/07/17/1095024/music-streaming-climate-friendly-tips/
https://www.technologyreview.com/2024/07/17/1095024/music-streaming-climate-friendly-tips/
Marks and Przedpelski, "The Carbon Footprint of Streaming Media: Problems, Calculations, Solutions"
09/04/24 15:07
Marks and Przedpelski, "The Carbon Footprint of Streaming Media: Problems, Calculations, Solutions," in Film and TV Production in the Age of Climate Change.
A massive survey, extracted and expanded from our 65-page 2021 report, featuring a critique of the politics of calculating ICT's carbon footprint, the International Energy Agency's attack on The Shift Project, and the small-file solution. On Selected Writings page
A massive survey, extracted and expanded from our 65-page 2021 report, featuring a critique of the politics of calculating ICT's carbon footprint, the International Energy Agency's attack on The Shift Project, and the small-file solution. On Selected Writings page
AI's carbon footprint
09/04/24 14:33
At the Small File Media Festival we are not big fans of AI (or rather, machine learning, as we do not believe these systems are intelligent) which is now the largest contributor to the expansion of data centers and also has a whopping water footprint (though we surmise streaming is still the largest contributor to ICT’s energy use as a whole, given the strain it places on on devices). If you must artifice, please use a small vision-language model like TinyGPT-V, or train a homegrown bot with precise tasks, which will draw less energy.
A few resources:
"Data centers are sprouting up as a result of the AI boom," Business Insider, October 2023 https://www.businessinsider.com/ai-data-energy-centers-water-energy-land-2023-10
“Artificial Intelligence Threats to Climate Change,” Climate Action against Disinformation
Li et al., “Making AI Less ‘Thirsty’: Uncovering and Addressing the Secret Water Footprint of AI Models”
Patterson et al., “Carbon Emissions and Large Neural Network Training”
Schick and Schütze, “Small Language Models Are Also Few-Shot Learners”
Wang and Wang, “Small language models (SLMs) A cheaper, greener route into AI,” UNESCO
Read More…
A few resources:
"Data centers are sprouting up as a result of the AI boom," Business Insider, October 2023 https://www.businessinsider.com/ai-data-energy-centers-water-energy-land-2023-10
“Artificial Intelligence Threats to Climate Change,” Climate Action against Disinformation
Li et al., “Making AI Less ‘Thirsty’: Uncovering and Addressing the Secret Water Footprint of AI Models”
Patterson et al., “Carbon Emissions and Large Neural Network Training”
Schick and Schütze, “Small Language Models Are Also Few-Shot Learners”
Wang and Wang, “Small language models (SLMs) A cheaper, greener route into AI,” UNESCO
Read More…
Call for work, Fifth Annual Small File Media Festival!
09/04/24 14:29
Coming very soon at smallfile.ca! We're continuing our partnership with Vancouver's legendary The Cinematheque and streaming at low bitrate worldwide in October. Deadline June 15.
