Long before Campbell Watson began working on climate models and environmental impact reporting for IBM Research, he was an accountant. He was good with numbers and his parents told him that accounting was a good and stable profession, so he joined an accounting firm right after college. He spent his days commuting to work on the 35th floor of an office building in Melbourne. Watson wore a suit to work every day, where he did tax returns for wealthy families. Apparently, he was successfully making his way in the world.
There was just one problem: He hated being an accountant. He consoled himself in the early mornings and weekends with his favorite pastime, surfing, to counter the monotony of the scoreboards and leaderboards. It only took six months before he decided he needed to get out. Fortunately, going to university in Australia was relatively affordable at the time, so Watson went back to school, starting in the general sciences and ending up in atmospheric science. Today at IBM Research, he works on advanced atmospheric models, but at the time his interest in Earth was more for fun. Now dressed in colorful vintage shirts at work, surrounded by a virtual backdrop of crashing waves on video calls, Watson retired from the branch of science that made his favorite hobby possible.
“I thought the connection between atmospheres, oceans and surfing was really cool,” he says. When Watson was 10, a friend taught him to surf. He grew up in Melbourne, near some fantastic beaches, popular for water sports. “I started on a boogie boogie, but I was jealous that he was able to stand up, so I went and got a surfboard,” he says. In high school, he started surfing regularly with a group of friends. They rode the waves up and down Victoria’s so-called Surf Coast, at spots like the picturesque Bells Beach. New York City’s waves pale in comparison to Australia’s, and Watson, ever in search of the perfect wave, once wrote of Vice about how he surfed a perfect artificial wave in the Texas hinterland. His old group of surfing buddies still try to get together for regular surf excursions to this day.
As a climate and sustainability scientist at IBM Research, the team leading Watson focuses primarily on geospatial data, which includes a partnership with NASA to model climate change and predict weather. The rest of their time is spent on environmental social governance (ESG) reporting, using large-scale linguistic models (LLM) to help businesses more efficiently track and report their environmental impact, such as their emissions. greenhouse throughout the supply chain.
Watson moved to the United States in 2012 for a postdoctoral fellowship at Yale University, where he studied cloud physics over Dominica and gravity waves over New Zealand. He has worked at IBM Research for more than a decade, after moving from Connecticut to New York in 2014.
At school, Watson had always excelled in science. He liked English, although he tended to get poor grades in the subject. “I just found science easier than the humanities, but I was encouraged to go into business.” His mother worked as a secretary and his father was in business recruitment. However, family influence aside, Watson always felt like his brain was wired to study physics, biology, and other hard sciences. “Initially I wasn’t attracted to science as a discipline, it was just that the information was easier to understand,” he says. “I don’t understand grammar; it’s always been so difficult,” he says. For Watson, his struggles with language education made the sciences seem more intuitive by comparison.
Ironically, he now works with language all the time. But these days, it comes in the form of LLMs. A typical day for him starts with progress report meetings with several different teams working on different scientific modeling problems. These are linked to the geospatial models they are working on, as well as their LLM work on ESG reports.
Off-the-shelf LLMs aren’t very good with things like acronyms and the semantic relationships of specific fields in sustainability—things that ESG reports are full of. So, the Watson team has used thousands of existing reports to conduct continuous prior training on the IBM Granite LLM. They incorporated fine-tuning, guideline tuning, and RAG to power a chatbot that can perform ESG tasks such as generating report inputs and calculating carbon footprints from supplier transaction data.
One of the challenges they have faced is finding a model to read the tables, albeit very different from what Watson faced in his accounting days. It seems simple, but while LLMs are good at parsing a paragraph, most models can’t necessarily tell that a column in a table contains information related to information in another column, or that they are organized by labels at the top of each column. They are working on the scope to address that detail, but ultimately, they want to come up with a system that allows interaction between this LLM and geospatial models. “This could be CO2 emissions from industrial plants, methane emissions from cows, or deforestation from the procurement of palm oil,” says Watson. “This information should be represented in these reports and can be collected through geospatial data.”
For this AI issue, there are many complex problems to be solved along the way, but he sees the unification of these processes into some kind of agent system as the ultimate promise of using models for climate and sustainability. “I think it’s really cool that we’ve used continuous pre-training to make him aware of the durability of the Granite LLM, but we haven’t adapted it so much that it’s no longer a Granite model,” says Watson.
Beyond the ESG work, geospatial models take up most of Watson and his team’s time. This year they reached a milestone with this work, launching the Prithvi WxC in collaboration with NASA. It is a new general-purpose artificial intelligence model for weather and climate. “We were tired of building a new model every time we wanted to learn something new from geospatial data, so when foundation models came along, we saw them as a great opportunity to address a variety of problems ,” he says.
Fortunately, a team at NASA was interested in using foundational models for Earth science, weather data, and climate predictions. Pulling data from different satellites, which acquire images at different times, at different scales, and even at different wavelengths of light, has proven to be a challenging multimodal problem. “It’s a very interesting problem that, if solved, will help many other fields,” says Watson. Prithvi WxC, which is open source, is a prime example of how his team takes all their research from basic concept, to building blocks, and then into something that will actually have an impact on the world .
“We have researchers, developers, scientists and engineers who are doing this work, but then it has to be translated to be something useful for our partners, collaborators and the community,” he says. This translation step is key, as it distinguishes the fundamental research being done at IBM Research from similar work at your typical academic institution. Part of Watson’s day-to-day work is dedicated to mapping this process, making sure that research is moving in a direction that ensures it will be useful to businesses and other organizations, and that those who can benefit from its work his team either understand what the team is doing, or provide input to move it in that direction.
As Watson has morphed into a lab leader in recent years, he is in a position to set strategy. His role is to question whether a project is scalable, useful and impactful – and to do something about it if he thinks it isn’t. After all, you can’t grab your surfboard and head to the beach without checking the weather report, finding out which beach has the best conditions that day and consulting tide tables. Nor do you fall into a wave without a plan. Growing into the role, Watson was surprised to find she enjoyed learning leadership skills. “It’s been fun discovering my own leadership style,” he says. “Although I can think of times in the past when I found myself leading things, it has felt somewhat random.” These data points assured him that it was possible.
Some happened outside of work, in situations where Watson was genuinely surprised by his ability to draw people into a project just for fun. One of these projects was a live coding performance where he wrote code in Ruby and Python, which was translated into sound. The result was equal parts art and science: He installed a weather station on top of the Ace Hotel in New York City and performed his musical score while a friend did the live coding to produce visual accompaniments. To pull it all off, they needed other people to set up the sound equipment, hook up the microelectronics to the weather station’s air quality monitoring system, and make sure the station’s Internet connection worked. of the weather was stable.
“It was really interesting to see how willing people were to help,” he says. “What do you get out of this? Why are you saying yes?” he would ask. But at the same time, Watson admits he’s eager to lend a hand when asked.
In surfing, a perfect day means keeping your head above water and enjoying every wave the ocean throws at you. In his work at IBM Research, Watson has found joy and meaning in the journey, the daily effort it takes to build on yesterday’s wins and bring his team along on the journey—and of course, the fun of doing it. things work.