|
Listen to this story
|
Leading AI research lab DeepMind is hiring research scientists who would be working towards the organisation’s efforts in climate tech and sustainability. The research lab claims to be creating ‘an exciting new team’ which will build AI solutions focused on reducing greenhouse gas emissions, starting with the carbon footprint of energy generation.
AGI has been the Northstar of DeepMind. The company strongly believes that advances in this direction would help humanity solve many real-world challenges, including climate change. In a 2021 DeepMind blog, author Koray Kavukcuoglu wrote, “Advances in AGI research will supercharge society’s ability to tackle and manage climate change – not least because of its urgency but also due to its complex and multifaceted nature.”
Interestingly, when DeepMind was acquired by Alphabet (Google’s parent company) in 2014, one of the first successful projects executed by the former was reducing the energy consumption of Google’s data centre cooling units by as much as 40%.
DeepMind’s climate tech solutions
Broadly, DeepMind has taken two approaches to tackle the climate change problem – prediction and control. While prediction models learn about patterns (e.g. weather patterns) and understand how they may evolve, control models prompt agents to take appropriate action in that direction. Two of the most important solutions DeepMind has built in the domain are weather prediction and plasma control for fusion.
In 2021, DeepMind open-sourced a dataset and trained a model for Deep Generative Models of Rainfall (DGMR). An AI system for short-term precipitation forecasting, DGMR bested several existing models in terms of accuracy and usefulness in 89% of test cases. The DeepMind team collaborated with the UK Met Office to perform short-term and high-resolution predictions of precipitation. DGMR uses generative models to generate ‘radar movies’ which predict the amount and location of future precipitation.
Earlier this year, the AI research lab collaborated with the Swiss Plasma Centre to develop a deep reinforcement learning system to successfully contain plasma in a Tokamak. This solution will advance nuclear fusion research. The research proposes building and running controllers on the Variable Configuration Tokamak in Lausanne, Switzerland, to control nuclear fusion plasma. The proposed learning architecture, claimed to be the first of its kind, combines deep reinforcement learning and a simulated environment to keep the plasma steady and accurately sculpt it into different shapes. This solution will ensure a source of ‘clean, limitless energy’.
The duality
Despite the great strides by DeepMind in the climate tech space, in 2020, it was reported that the company had shifted focus from tackling climate change to other areas of science. While the company had denied any such development, many key climate change researchers had left the organisation and the number of developments in this domain from DeepMind were far and few in between in recent years. A DeepMind spokesperson told CNBC, “Over time we’ve moved away from a narrower focus on domains, and cross-functional teams across DeepMind now contribute to our growing set of climate and sustainability projects.”
In 2017, DeepMind was gung-ho about solving the global warming issue, forming a new team called DeepMind Energy, led by Jim Gao, a former Google technical lead. At its peak, the team grew to 14 people who were building AI technologies to fight climate change. Gao exited the team and the company in 2019 to co-found Phaidra, an industrial AI company.
Climate tech in focus
Climate change is the hottest topic (pun unintended) which has been deservedly receiving much attention. At the current pace of deterioration, scientists believe that we are heading towards a climatic apocalypse in the near future. Against this backdrop, policy and government-level intervention aside, many big companies have announced intents and initiatives to combat climate change, and there has been an emergence of startups in the climate tech domain.
In recent years, climate startups have attracted record investments. According to reports, USD1.4 billion was invested in climate and carbon-focused startups in the second quarter of 2022 alone. This is a major achievement considering other industries are struggling to tap funds.
