AYS Webinar Summary - Session 3: How will AI Transform the Landscape of Precision Oncology?

Session Summary

The Biggest Question for My Science event series fosters a deeper understanding of the fundamental questions that drive scientific discovery. Each session features an AYS Fellow and a prominent scientist from related fields discussing the most critical question in their research and the scientific community's approach to it.

The third session, held on September 26, 2024 (CST), focused on Artificial intelligence (AI) and its application in precision oncology. With presentations by Professors Hao Chen (The Hong Kong University of Science and Technology), Katsushi Ikeuchi (University of Tokyo), and Ming Li (University of Waterloo), they explored recent progress on computational pathology with multimodal data integration for precision oncology through the lens of information theory and foundation model, with versatile applications to disease diagnosis, treatment response prediction, prognosis, etc.

During this session, our speakers shared their insights, suggestions, and expectations for junior researchers. Along with the biggest questions in their respective research fields, we hope these invaluable perspectives aim not only to inspire young scientists to think boldly, act purposefully, and achieve remarkable success, but also to equip them with the mindset to push boundaries, pursue impactful discoveries, and excel in their academic and professional endeavour.

"The Biggest Question" raised by speakers

Professor Katsushi Ikeuchi: How will AI Transform the Landscape of My Science? (Focusing on general robotics/computer vision questions)

Professor Katsushi Ikeuchi discussed the transformative impact of AI on various sectors, emphasizing the shift from AI as a mere component to an underlying framework that influences all aspects of life and industry. He highlighted the need for new strategies in AI development due to challenges with infrastructure and energy consumption. During his presentation, he identified several critical questions that remain unanswered:

  1. What are the fundamental differences between the current AI cycle and previous ones?

  2. How can we ensure the developed AI is human-friendly instead of Terminator-type?

  3. Should a human-friendly Doramenon type AI (Augmenting Intelligence) be developed instead of Terminator-type AI (Autonomous Intelligence)?

Professor Katsushi Ikeuchi discussed the significant transformation AI is bringing to various industries, moving from a component to a foundational framework with broad impacts. He addressed the need for innovative strategies in AI development to overcome challenges related to infrastructure and energy consumption, advocating for human-centric AI design.

Professor Hao Chen: Towards Multimodal Data Integration for Precision Oncology via Foundation Model

Professor Hao Chen focused on the use of AI in precision oncology, discussing the challenges and potential of developing generalized AI models in the medical field. He emphasized the importance of multi-modal data integration and the need for explainable AI to enhance trust and reliability in medical applications. Based on this importance, he posed several significant questions of the filed:

  1. In terms of data, how to get large-scale high-quality medical data for foundation model training while it is facing the ethical issues, heterogeneity, cost, and other challenges?

  2. In terms of algorithms, how to construct powerful enough AI algorithms for medical knowledge learning while it is facing the challenges including adaptability, capability, reliability and responsibility and so on.

  3. In terms of computing infrastructures, how to widespread deploy AI models? And how to sustainably learn the large AI models?

Emphasizing the importance of AI in precision oncology, Professor Hao Chen also highlighted the need to develop AI models that can effectively generalize across various medical tasks. He also stressed the value of integrating multi-modal data and the necessity for AI explainability to establish trust in medical applications.

Professor Ming Li: Recognizing Self and Nonself (for personalized cancer treatment as well as understanding autoimmune diseases)

Professor Ming Li explored the use of AI in understanding self and non-self-recognition, crucial for cancer and autoimmune disease research. He discussed the potential of modelling the immune system's selection process with AI and the implications for personalized cancer treatments. During his presentation, he raised the following questions in his research field:

  1. Can developed AI systems be considered human-friendly instead of Terminator-type?

  2. Why is AI-based modeling more effective in mice than in humans, and how can more human data improve results?

  3. How can AI tools be used in future clinical decision-making, especially for personalized medicine?

Message from the Speakers - To the junior researchers

Professor Ikeuchi: Think big, start small and take the first step

Professor Ikeuchi encourages young researchers to start small, emphasizing that the first step, however minor, is essential. He notes that students with grand ideas often become paralyzed by inaction. By beginning small, they gain momentum and fresh insight, which leads to the next step. He also advises understanding the history of their field and staying updated on current developments. He also highlights the value of interdisciplinary thinking and a broad perspective, acknowledging that research progress is typically incremental, requiring both vision and actionable steps.

Professor Chen: Finding the rights question is always half the journey to success

Prof. Chen stressed that finding the right question is key, often constituting half the journey to success. The rest lies in dedicating time and efforts to solving it. His message underscores the importance of curiosity and persistence, urging young scientists to focus on asking the right questions, which can lead to impactful and transformative solutions. He also suggested that students and young researchers should seek a strategy that balances the pursuit of important, challenging questions with the practical need to complete their degree in a timely manner, aiming to find a middle ground where they can work on substantial research that aligns with the academic progress.

Prof. Li: Pursue good science, not trends

Prof. Li’ s central message is to "believe in yourself" and pursue core scientific questions, even when they are not fashionable. He emphasized the importance of focusing on meaningful scientific problems rather than following popular trends in research, warning against producing papers that merely modify existing models. Using examples like Jeff Hinton’s students, who persevered with neural networks despite initial skepticism, and Jinbo Xu’s foundational work on AlphaFold, Prof. Li illustrated that sticking to core scientific issues can lead to significant breakthroughs.

Conclusion

The webinar on AI's impact on oncology featured experts covering AI's broad industry influence, precision oncology challenges, and immune system modelling. They highlighted AI's transformative potential in medicine, emphasizing human-centric design, data integration, and personalized treatment strategies. We list the questions here to encourage deep investigation and uncover new possibilities.

The Asian Young Scientist Webinar series is tailored for students and young faculty interested in science, research, and grappling with scientific inquiries across disciplines. Each session features a research topic raised by an AYS Fellow. It will be held monthly via online Zoom meetings and is open to participation.

With special thanks to our event network partners:

Next
Next

AYS Webinar Summary - Session 2: Is it Practically Possible to Slow Brain Aging?