Interview with an ML Expert: Climate Change, AI, and Making an Impact
Climate change is one of the greatest threats to humanity. That’s why it is our primary focus in Data for Good projects. At Appsilon, we are committed to making a positive impact on the planet, looking for ways to tackle climate change and mitigate climate risks. We spoke to Jędrzej Świeżewski, our Machine Learning Lead about his perspective, plans, and hopes for the future.
Read more to find out about:
- AI and Ethics – 5 practical applications
- Our definitive guide to YOLO Object Detection
- Machine Learning and Sustainable Agriculture
- Seeding change with Computer Vision
- ML in Gabon – A Data for Good success story
- The future role of technology in tackling hunger and climate change
Interview QnA with Appsilon’s ML Lead
Machine Learning and Making an Impact
Appsilon is committed to using data for good. As a Machine Learning Lead, where do you think lies the potential for the biggest impact?
At Appsilon, we can contribute the most in two main areas. The first one is the automated processing of images – still images, videos, or similar forms of data. These are the types of projects we can aid with computer vision techniques. The second one is data presentation. By building data visualization dashboards we are allowing users to get insights and draw conclusions otherwise obscured by the vastness of information or their complexity.
At the same time, it’s worth noting there are many different ways in which we, or anyone, can make a positive impact. Even though sometimes it might be hard to know what the final result will be, as many factors influence it. However, I believe the key moment is when the decision is made to prioritize impact. Once you make a commitment, it shifts focus and lays the foundation for future success.
Can you give us some examples of how processing images can be used for good?
Computer vision has been used in quite a few biodiversity projects. Introducing this technology speeds up the process of data analysis, for example quickly providing information from images captured by trap cameras. Things that used to take weeks or months can now be done in a matter of hours. This supports the monitoring and hence, the preservation of wildlife.
It is also being used in the processing of medical imagery. One of our collaborators used it to introduce novel diagnostic measures based on the monitoring of subtle motions of the patients. Those motions are so subtle (think of minute trembling), that they may be missed by a doctor. However, using computer vision, they can be identified and linked to developing neurological problems. As a result, a potential treatment can be applied at an earlier stage, potentially helping the patient before their health issues intensify.
But it’s not only about efficiency and saving time. The troves of data we collect are already enormous and keep growing. Without automation, we have no hope of unlocking all the knowledge within.
Positive, Global Impacts of AI
Are there any other AI capabilities you see making a positive impact on people’s lives?
Machine translation has helped to connect the world, making it more accessible than ever before. Variants of reinforcement learning are automating mundane tasks (both in the private and professional lives of people) so that we can devote time to more interesting and fulfilling tasks. Various flavors of unsupervised machine learning have also been applied in research, e.g., to aid drug discovery.
What have been some of the best projects you have worked on so far?
A cornerstone project, which proved that we can make a significant social impact, was Mbaza. The important work done by ecologists and park rangers to preserve one of the most pristine ecosystems, the Gabonian rainforest, could clearly be enhanced with technology. To help them structure and visualize (geographically, temporally, and through direct inspection) the vast amounts of data collected in the studies, we developed an AI-powered analytics tool.
The machine learning models we have trained added a layer of automation to the data analysis. It cut the most time-consuming part of it from several weeks of manual work to several hours of our software running on a laptop. Accounts of the users of Mbaza are very rewarding, encouraging us to continue our efforts in improving the solution.
A more recent project, which I personally find very promising, is a collaboration with scientists from the Arctowski Polish Antarctic Station managed by IBB PAS. They use various methods of visually monitoring the Antarctic ecosystem, and we already see their processes can significantly benefit from the use of computer vision. One example case we studied was automating the way nests of Shags are counted, based on images collected by drones.
Driving Change with ML
What do you think will be the next big thing in technology for good? Something that could help solve some of the pressing global challenges.
Machine learning is an example of an enabler tool. We can already see it is being successfully used in so many different industries and areas of research. It transforms the way we approach problems. Every time I see the list of the biggest global challenges, I can recall machine learning being used to aid each of them. It’s simply amazing.
Of course, it is not the solution to the challenges in itself. That’s why my hope is that the biggest impact will take place on a meta-level. The technology will not only accelerate the process of solving individual challenges. But also, as importantly, it will encourage more people to get involved and be empowered by the current advancements.
Do you think we are making the most of AI’s potential when it comes to combating various societal and ecological issues? Or are there some obstacles slowing down the application of this technology?
From my point of view, the pace could be higher.
One of the reasons for this is the regulations, both in terms of acts of law and bottom-up best practices. They are definitely needed to make the process safe but are unfortunately lagging behind. I do appreciate the complexity of the task of regulating AI’s development. So I’m in favor of doing it gradually. At the same time, I believe being too cautious opens up more risks than it manages.
Another reason is that the field is still relatively fresh. As such, the potential of applying the methods to increase economical gain drives people in the industry away from using the methods for social good. I have high hopes that this will gradually change. Many people at the beginning of their careers that I speak to actively want to contribute to the greater good.
It is why we have our Data for Good program, to help shift the industry’s perspective and show the great things that we can do to bring a positive impact on the planet.
It is also a recurring theme among the reasons people joining Appsilon cite as drivers of their decision – the possibility of not only working on exciting projects but also doing their part in tackling global challenges.
The Future of ML and Positive Change
You say that what drives you the most in this field is the huge application potential of machine learning. When it comes to social domains, are there challenges you feel particularly strongly about and would like to focus on in your work?
For me personally, biodiversity and wildlife loss, in general, are important topics. From the global impact of habitat degradation on trophic chains to the very local struggles to keep cities green – planting trees, or creating wildflower meadows).
Machine learning skills can aid those efforts, by enhancing the insights we draw from data. Classical data processing and visualization techniques can also be very informative, e.g., raising awareness about the issues.
What are your hopes for the future, when it comes to applying artificial intelligence and machine learning for social good?
I believe the climate crisis is here. We need to act to mitigate the effects our past and current actions will have on this and the next generations. I hope leveraging data science techniques, machine learning, and computer vision, will contribute to that effort. I hope that by producing positive examples of such an impact we will also draw more people, possibly with different skill sets and approaches, into doing their part.