Climate crisis in focus: AI lens
Welcome to the 60th edition of the?AI and Global Grand Challenges newsletter , where we explore?how AI is tackling the largest problems facing the world.
The aim:?To inspire AI action by builders, regulators, leaders, researchers and those interested in the field.
If you would like to support our continued work from ￡1 then click?here !
---
Packed inside
If you would like to support our continued work from ￡1 then click?here !
Graham Lane & Marcel Hedman
__________________________________
Key Recent Developments
---
Can AI help resuscitate Mozambique's mangrove forests?
What:?As a result of climate change, tropical storms and cyclones are becoming frequent in Mozambique. The government is working vigorously to make the country’s coastal communities resilient to the effects of global warming. One project with a partner called Blue Forest addresses coastal reforestation aiming to plant 100 million mangroves in the next 30 years offsetting 200,000 tonnes of CO2 annually. They are using AI and other technological advancements to identify key hot spots in coastal Mozambique where restoration is most needed.
Key Takeaways:?The project is an example of carbon offset in action. It will be financed by the sales of carbon credits generated through the reforestation activities over the 30-year period of the partnership. Any proceeds will be used for the development and building of the local communities. Crucially it demonstrates AI and its effectiveness in optimization problems.
---
How Meta is using AI to focus on more sustainable technology
What:?Meta (formerly Facebook) AI published a blog post and an accompanying video explaining how they are using AI to address the climate crisis. The blog provides the example of Meta’s collaboration in the Open Catalyst Project, which uses AI to model and discover new catalysts. These are crucial to developing better ways of storing electricity that is produced by intermittent renewable energy sources such as wind and solar.
Researchers at Meta are also developing computer vision systems to interpret satellite imagery and produce high-resolution maps of the world’s forests. The aim is to estimate the carbon stored in forested areas, which would improve the ability to assess the impact of reforestation efforts.
The posting also discusses steps within the company to improve energy efficiency even though they are already using 100% renewable energy.
Key Takeaways:?Large AI models and, more broadly, the big tech companies, have been widely criticised for pursuing a growth strategy that is environmentally unsustainable. It is encouraging that Meta are publicising their commitment to environmental sustainability even as the company battles with other ethical challenges.
---
Climate Change AI - interactive summaries and video tutorials
What:?Tackling climate change with machine learning ?is a major, recent ACM Computing Survey. It describes approximately 80 different applications of ML to address the climate crisis. The Climate Change AI website now provides a means to search the report for ML and thematic keywords and to view summaries of relevant sections.
The Climate Change AI website also provides?two 45 minute video tutorials :
- one is an introduction to climate change for ML practitioners;
- the other is an in-depth introduction about ML that would be suitable for climate change researchers.
__________________________________
AI Ethics
An MIT Technology Review series investigates how AI is enriching a powerful few by dispossessing communities that have been dispossessed before. With case studies from South Africa, Venezuela, Indonesia, and Aotearoa (New Zealand).
A thoughtful article exploring different concepts of bias and competing notions of fairness.
领英推荐
A report from the Stanford University Human-Centered AI spring conference. A sample quote from Rob Reich, Stanford professor of political science:?“AI scientists [are] like late-stage teenagers who have just come into a recognition of their powers in the world, but whose frontal lobes are not yet sufficiently developed to give them social responsibility.”
The?full conference video ?is now available.
Other interesting reads
Images of deforestation of the Harz forests in Germany, coral bleaching on the Great Barrier Reef in Australia and glacial melt in Sermersooq, Greenland.
Computer vision is being used to scan images to collect conservation information about animal populations at an unprecedented scale. In August 2021, 17 million images were analysed automatically.
A technical description of the use of ML in simulations that identify geological sites suitable for storing CO2 that has been captured from the air.
__________________________________
Cool companies found this week
Environment
Blue Forest ?- Has a mission to “Bring Innovation to Reforestation” and has announced a new project restoring mangrove forests in Mozambique by leveraging AI and carbon capture with the aim of planting between 50 – 100 million trees. Blue Forest projects are?financed through carbon credits .
Digital twins
Maxar ?and?blackshark.ai ?- announced a?collaboration to create a 3D digital twin of the entire planet ?that is continually updated. Maxar provides a global cloudless satellite imagery base map. Blackshark.ai will transform this into a semantically labelled digital twin, making it easier to use in various simulation use cases. The venture raised $20 million in seed funding in November 2021 but financial details of the new arrangements were not made public.
Smart buildings
PassiveLogic ?- has developed a platform that can generate a digital twin of a building and then “utilize new autonomous technologies, machine learning and easily configurable software to help streamline the implementation and management of building controls systems”. The company raised?$34 million in January ?and has now?raised an additional $15 million .
__________________________________
And Finally ...
Full Level 5 autonomous driving ?… on Mars
(so we can’t call it a “World First”)
__________________________________
AI/ML must knows
Foundation Models?- any model trained on broad data at scale that can be fine-tuned to a wide range of downstream tasks. Examples include BERT and GPT-3. (See also Transfer Learning)
Few shot learning?- Supervised learning using only a small dataset to master the task.
Transfer Learning?- Reusing parts or all of a model designed for one task on a new task with the aim of reducing training time and improving performance.
Generative adversarial network?- Generative models that create new data instances that resemble your training data. They can be used to generate fake images.
Deep Learning?- Deep learning is a form of machine learning based on artificial neural networks.
Thanks for reading and we'll see you next week!
If you are enjoying this content and would like to support the work then you can get a plan?here ?from ￡1/month!
___________________________________
Graham Lane and Marcel Hedman
This newsletter is an extension of the work done by?Nural Research , a group which explores AI use to inspire collaboration between those researching AI/ ML algorithms and those implementing them. Check out the website for more information?www.nural.cc
Feel free to send comments, feedback and most importantly things you would like to see as part of this newsletter by getting in touch?here .