2024 AI/ML Systems Conference LSU

Two weeks ago, I was fortunate to speak on the “AI In Energy” panel at the 2024 AI/ML Systems conference at Louisiana State University. I shared the panel with Dr. Frank Tsai and Dr. Z. George Xue, two innovative researchers in carbon sequestration and climatic oceanography, and was invited by Brad Ives with the LSU Institute for Energy Innovation.

I’d like to share three lessons from the conference and reflect on how it’s shaped my approach to ML with my team at MAIA Analytics.

First, while the climate concerns from scaling AI are legitimate, there exist many opportunities to apply AI to directly mitigate carbon emissions. For instance, numerical inverse modeling of hurricanes requires hours of processing time from a supercomputer center, but a well-trained ML model that can run on your laptop and complete the task in minutes. Rather than solely focusing on large foundation models, we should continue developing smaller, energy-efficient models fine-tuned for specific applications, which can deliver comparable results with a fraction of the computational cost.

Secondly, I was surprised by the momentum from my fellow conference attendees, both in computing and earth sciences, to address the climate crisis at its source: on the ground of the economy. I sometimes feel like this is the crux of our job at MAIA Analytics––uncovering the drivers of commercial and industrial solar installation and making this information actionable for key players. We need to continue strengthening the bridge between academia and industry in computational and earth sciences, ensuring that research insights translate into practical solutions for sustainable energy adoption.

Finally, the climate crisis presents a unifying challenge for the academy and industry alike. Louisiana has firsthand experience with extreme weather disasters, and my alma mater state of North Carolina has recently been affected by Hurricane Helene. I see a challenge in these tragedies, both to understand these rapidly shifting earth-systems processes and to address their root causes.

At MAIA, we are committed to expanding the use of machine learning to drive the renewables industry forward, ensuring that communities across the US can leverage nascent public-private partnerships to build a green, resilient economy in the face of increasing uncertainty. I am excited to see where the Institute for Energy Innovation continues to move its platform for public-private dialogues, and I hope to see more organizations committed to the use of AI as a force for good against our generation’s greatest challenge.