.png "ACSGCI"){kind=logo}
By Ashley Baker, Scientific Content Manager (Contractor) at the ACS Green Chemistry Institute
As part of the upcoming GC&E Conference’s re-vamped Student Experience, three postgraduate students are organizing an AI Hackathon, offering a new, experiential update to the annual student workshop. Sixty students will participate all day on Sunday, June 2, preceding the main conference. Working in teams to develop predictive models for chemical toxicity, the event will bring together experts and enthusiasts on AI and green chemistry. Participants will gain valuable hands-on experience in emerging chemistry methodologies that represent increasingly important career skills.
We had the opportunity to interview organizers Henrique Marçon, Paula Teeuwen, and Michael Zhou about their experience in a unique research program at the University of Cambridge, the value of hackathons, and their excitement about kickstarting collaborations.
Computational Tools in Student Research
Henrique, Paula, and Michael are all Ph.D. students at the EPSRC Centre of Doctoral Training in Automated Chemical Synthesis Enabled by Digital Molecular Techn.... There, they receive cross-disciplinary training to prepare for next-generation molecule-making problems, particularly the opportunity to engage with computational tools. While many students at the Centre experiment with machine learning to improve organic syntheses, there is also a required course load and an emphasis on collaborating across disciplines and cohorts.
“We are each in a different cohort year of SynTech and do research on different topics,” said Paula, whose research is in supramolecular chemistry. “I use computational methods for modeling metal-organic cages.”
Michael, meanwhile, is using machine learning to make predictions about chemical reactions and processes, specifically to increase the efficiency of drug discovery, optimize reactions, and increase reaction yield.
Henrique's concentration is on leveraging AI for predicting reaction selectivity. “I have a background in both chemistry and engineering which led me towards green and flow chemistry,” Henrique said. “Now I’m working to leverage machine learning, especially in cases where it’s unclear in what direction a reaction will go.”
While the Centre is not specifically focused on sustainability, Paula noted the connection to green chemistry by increasing the efficiency of chemical processes, building in circularity, and factoring in inherent hazards of reagents and intermediates.
“If you want to screen everything by hand, it requires a lot of solvents, time, and effort,” Paula said. “So, I think automation and computational methods such as AI can be used to accelerate processes and in that way reduce the need for huge amounts of trial and error. This could save a lot of resources.”
Henrique also points out that AI can evaluate large data sets, including experimental results and references, helping scientists see the bigger picture.
“Analyzing a vast collection of chemical data is very hard for humans," he said. “AI and machine learning can enhance our research, whether by predicting reactions before going to the lab or looking at the outcome of a molecule in nature and preventing the manufacture of chemicals that persist in the environment. Computational models help with challenging problems that we haven’t previously been able to solve.”
The Hackathon Experience
Paula, Michael, and Henrique all bring experience in hackathons – whether as participants or organizers – to the event at GC&E. They emphasized the value of this unique opportunity to work in interdisciplinary teams where machine learning and chemistry knowledge are combined to create novel solutions.
For the June hackathon, participants will build AI models to predict the toxicity of particular chemicals and determine which features contribute to their toxicity. This would allow them to propose safer alternatives.
“We played around with different data sets and tried to find ways to make people think out of the box,” he said. “We wanted to go beyond just plugging and playing and have people use their chemical knowledge and experience to make it fun and engaging.”
All three organizers said students shouldn't be apprehensive of the term "hackathon." Unlike a course, there is no right or wrong answer, but rather an infinite array of possible creative approaches to addressing the problem. To make the hackathon accessible to beginners, the organizers sent a tutorial introducing machine learning before the event. They emphasized that while the hackathon is beginner-friendly, it’s still highly valuable to those with more machine-learning experience.
In addition to using hard skills like Python and various machine learning methods, the organizers see the hackathon as an opportunity for students to build their interdisciplinary collaboration and presentation skills.
“It’s very important for participants to be able to explain what they did and identify how they can improve it,” Paula said. Michael, Paula, and Henrique all hope to provide participants with a great experience, create an environment for interdisciplinary collaboration, and inspire creative, hands-on learning that helps enhance students' future research. Being an international group themselves – hailing from Brazil, the Netherlands, and the UK – the organizers are excited about traveling to Atlanta and the many perspectives the hackathon and conference are bringing together.
“It’s not just one country, one region, one reaction. We can see the whole of chemistry becoming more sustainable,” Henrique said. “All while making new connections and learning new skills – it’s pretty hard to beat that!“
____
You can still register for the 28th Annual Green Chemistry and Engineering Conference, taking place June 2-5 in Atlanta, GA. This year's programming features numerous new and special events, including an escape room, innovation hub, networking receptions, a community service project, and more. Can't make it to Atlanta? Register as a virtual participant for exciting hybrid symposia.