People of ACM - Sekou Remy

December 20, 2022

You have said you believe the promise of AI lies not in the power to replace human decision-making but in the ability of AI to enable superhuman decision-making. Will you explain what you mean by this statement?

We are currently at a point in science where computation is being leveraged in many significant and impactful ways. That said, we still have a very long way to go. Our team’s approach is to directly focus on users and use-cases which can leverage AI—and Machine Learning (ML) specifically—in principled, tractable ways to help decision-makers manage incredibly large search spaces, especially ones where the sequencing of actions is important. We’ve started in healthcare where there’s already a culture of data and model consumption, and we’ve been able to mine use-cases where decades of advances from Supervised Learning and Reinforcement Learning (as just two examples of ML) can be brought to bear. At the end of the day, we believe that these types of cutting-edge technologies can augment human decision-making, enabling us to do more in the future.

A recent paper you co-authored was titled “AI-Assisted Tracking of Worldwide Non-Pharmaceutical Interventions for COVID-19.” What have we learned from applying technology to address Covid-19 that may help us respond to future pandemics?

Our goal with that project was to highlight practical ways in which AI could be leveraged to support the COVID-19 response at multiple levels from personal to global. In that particular paper, our team was able to demonstrate how a Natural Language Processing (NLP) pipeline and a hybrid cloud deployment could help a small team scale their reach. Early in the pandemic, there were very few comprehensive sources of information about what responses governments were implementing within their borders. Such information was critical for those analyzing the pandemic. We observed that there were small teams—some in academic settings, some citizen scientists—who were trying to keep track of what was happening in their own geographies. These small teams struggled with the pace at which information was produced, the variety of formats, and venues from which information was channeled. The teams also struggled to contextualize the data sets which were being shared. Our project developed an open-source tool to help teams like these. The tools were also packaged within a platform which could be leveraged across borders and deployed in multiple contexts. COVID-19 was unlike many pandemics before it in that there were a lot more digital artifacts, but users were still struggling to process them in meaningful ways. We developed tools which helped them get a handle on what was happening.

These tools were useful in themselves, but I think that they were even more useful as concrete examples of AI which we could present to public health professionals and their colleagues. AI has a communication problem. Many times, the people who should be sponsor-users are lost because we aren’t able to effectively share how our tools can improve their workflows. Using these tools to develop communities of practice and advocating for the application of AI in a principled manner was a very rewarding experience for me.

You recently wrote an overview of how IBM Research is combating malaria in Africa and around the world. What has been the most important technological innovation in this effort?

We explore how AI and hybrid cloud-enabled efforts (from genomics to mechanistic modelling) can help with infectious diseases like malaria. Malaria is far from “solved” and the current COVID-19 pandemic has reversed progress in many places around the globe. These challenges notwithstanding, there has been significant progress in the sharing of computational resources between organizations. Cloud-native computing is the technological innovation which has been the most important as it has helped simplify the management and sharing of resources, and not just in the cloud. For example, over the past three years, IBM has lent its technical expertise to the Bill and Melinda Gates Foundation and some of their program-aligned partners to deploy ML and mechanistic disease models at scale. All these disease models are developed by experts in the field. Together, using cloud-native computing, we are amplifying the value which can be derived from such models (and available data) by applying ML in transparent and principled ways.

What is the computing and technology scene like in and around Nairobi in terms of factors such as students graduating from universities with computer science degrees, technology startups, and large technology companies? What trends will shape the computing landscape in that region in the coming years?

IBM Research-Africa has been in Nairobi for 10 years, and I’ve had the pleasure of being here for more than half that time. There’s a very vibrant and highly interconnected innovation ecosystem here in Nairobi. Universities play a pivotal role in this ecosystem, but I’d say that technology hubs are the key catalysts for the innovation culture. Known as the Silicon Savannah, Nairobi is important both regionally and within the African continent. It also plays a critical role for convening talent outside of Kenya. In terms of future trends in the region, I’ve noticed that universities are continuing to evolve and embrace more aspects of the “Hub Culture.” Many of these new hubs are domain specific and influenced by artificial intelligence (e.g., AI+Maternal Health and AI+Aerospace). Multinationals and local tech companies are also quite active and are continuously evolving. I think the future will be shaped by increased engagement from the African Diaspora investing time and talent into enterprises, and by extension universities in Kenya and other countries across the continent.