Exploring AI Tools to Tackle Climate Change

Original article can be found here (source): Artificial Intelligence on Medium

Exploring AI Tools to Tackle Climate Change

Changemakers in AI Present at San Francisco’s Climate Action Summit

Aroshi presenting her research at the San Francisco Climate Action Summit

Guest post by Aroshi Ghosh, Berkeley AI4ALL ‘19

AI4ALL Editor’s note: Meet Aroshi Ghosh, a Berkeley AI4ALL 2019 alum and Changemaker in AI. In this post, she discusses how she became passionate about addressing climate change and how her experience speaking at the Climate Action Summit empowered her to share her work and influence others to action. She explains the research project she worked on with other Berkeley AI4ALL alumni, which seeks to address some of the climate change challenges we are already facing today.

Today, you turn on the TV and more likely than not, you are confronted with news clips of disasters around the world — unseasonal floods, snowstorms, wildfires, and droughts. Growing up I learned about climate change and its destructive potential to influence the earth and the future of humanity; however, it never registered as an issue that I could address, appearing to be a cause only for scientists or environmentalists with complicated graphs and pictures. After watching the United States withdraw from the Paris Agreement, I became disillusioned, tired of relying on governments and policymakers to take action. Feeling powerless, I brainstormed what I could to make an impact — even in a small way.

I decided to make my voice heard by applying to speak at the Climate Action Summit 2020, hosted at the Microsoft Reactor in San Francisco this past February. I wanted to share exactly how everyday people can use technology to be a part of the climate change solution. My partner Isita, a Stanford AI4ALL alum, and I started by introducing the fundamentals of AI, and followed with three case studies based on our own AI project experience. My case study featured a machine learning model I created last summer, in conjunction with other Berkeley AI4ALL alumni. The model uses Natural Language Processing (NLP) to process aid requests, providing disaster relief during large scale, and increasingly frequent, natural disasters.

We argued that while it is important to prevent the further progression of climate change, it is equally important to help solve existing problems that are the result of climate change.

Currently, natural disaster victims receive aid primarily by sending “SOS” messages to relief hotlines. However, not all messages sent to these hotlines are aid-related, and sorting the messages manually can take a lot of time and effort. To address this issue, my team and I created a simple neural network model which preprocesses the text messages into embeddings — or vectors — and runs them through a variety of models (Logistic Regression, Simple Neural Networks, and Recurrent Neural Networks), sorting the messages faster than a human can. The algorithm then performs a binary classification, labeling each message as “Aid” or “Not Aid.” We are still improving our algorithm, but our model demonstrates how simple it can be to leverage AI to make a difference in our community. We wanted our Climate Action Summit audience to understand that with a little training, anyone can learn the basics of AI.

Aroshi presented her aid relief model to demonstrate how simple AI can solve complex problems

Presenting at the Climate Action conference and interacting with people who were as passionate as I am about the environmental crisis was eye-opening. The forum was sponsored by the Rising Women’s Collective, and it was enriching to meet so many passionate environmentalists from such a diverse array of jobs and industries all with the desire to “collaborate [and] build solutions for a more resilient world”.

To actually realize that I could be an agent of change as a high school student, mobilizing adults into exploring the world of AI and data science, was an unforgettable experience.

Speaking at the Climate Action Summit helped me understand truly how important it is to empower the people around us into becoming a part of the fight against climate change. I realized that I can be a voice of this generation, and by sharing my knowledge and collaborating across disciplines, I can help identify innovative technological solutions that may ultimately help us overcome the biggest challenge facing our world — and our very existence.

About Aroshi

Aroshi Ghosh is a high school sophomore in Northern California. She is a Berkeley AI4ALL 2019 alum and is passionate about identifying AI technology to address the environmental crisis. Aroshi also believes that communication is as important as technical skill, competing in tournaments with her school’s nationally ranked speech and debate team. This summer Aroshi will be interning at the Naval Postgraduate School, where she plans to create a data analysis platform that will extract sonar data, allowing users to see 3D visualizations of undersea terrain.