Sustainable AI

In this episode, I spoke with Koustuv Saha about his paper called Modeling Stress with Social Media Around Incidents of Gun Violence on College Campuses. We spoke over the summer, on a day that neither of us expected would be in the immediate aftermath of one of the worst mass shootings in Illinois history.

You can find a link to the full paper here: https://dl.acm.org/doi/abs/10.1145/3134727?casa_token=HWXx8Y-udkAAAAAA:FvlSN_4HOAawImzQJOJ8r7HLt1cltAexUoglICyAAVy-1zKtXowHTY9spB4Cxbdsvc3-kIZnXQ

You can learn more about Koustuv here: https://koustuv.com/

Full transcript: https://tinyurl.com/4se6pjfm

Sustainable Development Goal #16: Peace, Justice & Strong Institutions

In this episode, Lauren Kahn, research fellow at the Council on Foreign Relations, talks about the impact of AI and emerging technologies on the future of warfare. She answers questions like, how is AI being used in conflicts like the one in Ukraine? Which players and institutions are driving defense innovation? And what risks are being considered as we traverse this new frontier? 

It's a grim reality that war is at times necessary to promote and protect peace. This interview seeks to understand how AI is being used and thought about in this context. 

References: 

A Force for the Future: A High-Reward, Low-Risk Approach to AI Military Innovation

In this episode, we dive into Google's Open Building's dataset with Abdoulaye, a research program manager at Google Research Ghana. Join us as we explore how this dataset is putting communities on the map. Informal settlements around the world often go unnoticed by traditional tools, making it challenging for governments and humanitarian groups to assess the population size and identify certain needs. Discover how the team at Google utilized machine learning to pinpoint the location and geometry of buildings in Africa and Asia, and learn why they made this valuable dataset accessible to the public. Abdoulaye shares insights on the project's implementation, its significance, and the cautious considerations involved in such an innovative initiative.

Correction: In this episode, Abdoulaye mentioned that he has been to Makoko. He has not actually been to Makoka. He lived in Lagos, and often drove past Makoko but never took the boat to actually visit the city.

COVID-19 patients that don't believe in COVID-19, hospitals at full capacity and burnout among healthcare professionals, are just a few of the challenges that put pressure on our hospital systems, on top of the age old question, what's wrong with this patient and how do we fix it?  To help solve these challenges, Dr. Karandeep uses, not only his background as a trained physician, but also his expertise in biomedical informatics.  In this conversation, Assistant Professor of Learning Health Sciences, Internal Medicine, Urology, and Information at the University of Michigan, Dr. Karandeep, and Lily Adelstein discuss applications of artificial intelligence in the hospital setting. 

Episode links: 

Dr. Karandeep's linkedin: linkedin.com/in/kdpsingh/

Dr. Karandeep's twitter: @kdpsinghlab

Early identification of patients admitted to hospital for covid-19 at risk of clinical deterioration: model development and multisite external validation study: https://www.bmj.com/content/376/bmj-2021-068576.full

Digitising the prediction and management of sepsis: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(22)00658-4/fulltext

External Validation of a Widely Implemented Proprietary Sepsis Prediction Model in Hospitalized Patients: https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2781307

Predicting 30-day hospital readmissions using artificial neural networks with medical code embedding: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0221606

This episode is a conversation with Director of Innovation Michael Kayemba and Regional Director, International Alex Nawar at GiveDirectly. GiveDirectly is a non-profit that helps individuals in extreme poverty by providing unconditional cash transfers via mobile phone.

In this episode we discuss extreme poverty, GiveDirectly’s approach to alleviating extreme poverty and how the organization has started using machine learning to effectively target the individuals most in need.

In this episode of the podcast, Lily Adelstein speaks with Ariel Frankel about building capacity among public health officials world-wide to prepare for and leverage AI and other edge technologies to advance public health priorities. 

Ariel Frankel is the Director of Public Health at TechChange, an all-in-one digital training and events firm. She is currently using human centered design methodologies to support the development of an online learning platform to facilitate the capacity building and knowledge sharing of key public health issues for a wide variety of audiences, from frontline health workers to Ministry of Health officials. 

In this episode of the podcast, Lily Adelstein and Emily Aiken discuss the paper, "Machine learning and phone data can improve targeting of humanitarian aid". 

Emily Aiken is a PhD student at the UC Berkeley School of Information advised by Joshua Blumenstock. Emily is interested in the application of computational methods to issues of economic development, global health, and sustainability. She is currently working on expanding machine learning methods for estimating poverty and other measures of well-being from digital trace data. Her past research has included deep learning methods for tracking disease outbreaks using Internet-based data sources and identifying missing people in social media streams using techniques from computer vision. She received her B.A. in computer science from Harvard University, and is the recipient of a 2022-2024 Microsoft Research PhD Fellowship.

Website: https://emilylaiken.github.io/

Welcome to Sustainable AI! This podcast brings together researchers and industry professionals using artificial intelligence and machine learning to tackle some of today's biggest challenges - 17 of today's biggest challenges to be exact. Why 17? In 2015, the United Nations laid out 17 ambitious goals to be achieved by 2030. In this podcast, we'll explore how creative researchers and professionals are using machine learning to achieve these goals. 

These goals include: no poverty; zero hunger; good health and well-being; quality education; gender equality; clean water and sanitation; affordable and clean energy; decent work and economic growth; industry, innovation and infrastructure; reduced inequalities; sustainable cities and communities; responsible consumption and production; climate action; life below water; life on land; peace, justice and strong Institutions; and partnerships for the goals.

In this episode of the podcast, Lily Adelstein and Sammy Mahdi discuss the ways WaterScope is using machine learning to help ensure safe drinking water worldwide. 

Sammy Mahdi, PhD, is the lead project engineer at WaterScope UK limited. He works closely with engineers, suppliers and end-users to develop and bring the product to market. He received his Master of Research and his PhD from the University of Cambridge in Sensor Technologies and Applications.

WaterScope's mission is to enable inclusive access to smart, easy-to-use, and reliable testing for bacterial presence in drinking water.

Since its conception in 2015, the team has designed and developed a water testing system specifically for and with end-users. The testing process maintains scientific integrity while remaining low-cost, user-friendly, and effective across various environments.

For more information visit www.waterscope.org or @WaterScope_org