Demystifying the Ph.D. in AI...

Journey, Challenges, and Innovations in the Medical Field

Good morning fellow AI enthusiast! I'm thrilled to share a special episode of the "What's AI" podcast with you! In this iteration, we focus on AI research and the Ph.D.... my Ph.D. I've wanted to help future Ph.D. students and anyone curious about the world of artificial intelligence research.

p.s. read until the end! Auxane also shared her recent work in AI ethics with us!

1️⃣ Demystifying the Ph.D. in AI: Journey, Challenges, and Innovations in the Medical Field

Hey there, it's Louis Bouchard, and I'm thrilled to share a special episode of the "What's AI" podcast with you! Today, it's just you and me, and we're going on a journey through my Ph.D. in AI at Polytechnique Montreal and Mila. I will be honest, it's hard to find guests during summer, and I've wanted to help future Ph.D. students and anyone curious about the world of artificial intelligence research. If one of those two sounds like you, this episode is for you. I'll take you through my personal experience, the challenges, and the reasons behind choosing this path.

Back in the day, I had no clue what to do with my life, just like many students. But when I stumbled upon a class about AI, everything changed. I fell in love with the idea of being paid to do math and programming - the cool stuff, you know? From there, my passion for AI grew, and I decided to pursue a Master's degree to dive deeper into the field.

During my Master's, I started my YouTube channel to share my learning journey and connect with others interested in AI. It was fantastic to have a platform to explain complex concepts and learn from my viewers' feedback. This experience led me to realize that I not only loved AI, but I also enjoyed teaching and sharing knowledge.

Fast forward to my Ph.D., where I'm now tackling two exciting challenges: training AI models with sensitive hospital data and using AI to diagnose and monitor multiple sclerosis. It's no easy feat, but I'm determined to make a difference in the medical field through the power of AI.

Throughout this episode, I'll share insights on the admission process, the day-to-day life of a Ph.D. candidate (spoiler alert: it's a lot of reading papers and coding experiments), and the skills you develop along the way, like critical thinking, problem-solving, and effective communication.

I'll also delve into the concept of federated learning, which allows us to build robust AI models without compromising the privacy of sensitive data. It's an exciting approach that opens up new possibilities for AI development in healthcare.

Lastly, we'll explore how AI can revolutionize the diagnosis of multiple sclerosis. By automating lesion detection on brain scans, we can save time, reduce errors, and provide better insights for patients and healthcare professionals.

So, whether you're considering a Ph.D. in AI or simply curious about the intersection of AI and medicine, join me on this special episode. I'll share the ups and downs of my journey, the impact of AI on healthcare, and how you can get involved too.

Hit play now, and let's embark on this AI adventure together! And hey, if you have any questions or want to chat about AI, Biomedical Engineering, or anything else, feel free to reach out to me on LinkedIn, X or Twitter @What'sAI. Can't wait to hear from you!

Listen now on Spotify or Apple Podcasts!

2️⃣ [Sponsor] An AI-powered Investing Copilot

Pluto.fi’s AI-powered Investing Copilot helps you comprehend complex financial data and generate charts from data that is relevant to you. Pluto’s AI is designed to streamline and customize market research and investment strategies based on your own portfolio and investment goals. Whether you need data visualizations or insightful analysis, Pluto is along for your ride in the world of investing. I personally really love how it can help me understand more about my portfolio to further improve it. Basically a very good "research assistant" for investing.

Here’s a quick tutorial using Pluto.fi, your AI investing co-pilot! You can also use Pluto to create customizable dashboards to monitor specific stocks, track portfolio performance, and follow market trends. See an example of this here: in this tutorial video!

2️⃣ AI Ethics with Auxane - Integrating Bioethics and AI Ethics in Healthcare

Greetings AI aficionados!

As technology advances, we find ourselves at the crossroads of AI and healthcare, sparking crucial ethical discussions. Today, we're going a bit more theoretical, and I will be introducing an approach that brings together bioethics and AI ethics to address the ethical implications of AI in healthcare. I thought this would be a perfect fit with Louis’ work! 

AI Ethics - Understanding the Impact:

AI ethics centers around the ethical consequences of AI technology on societies, focusing on principles such as human autonomy, rights, non-discrimination, transparency, and privacy. Moreover, there's a concern that AI could amplify existing discrimination patterns, and that black-boxes AI could bring even worst consequences than anticipated. While ethical frameworks exist, the healthcare sector's specific implications often remain upractically addressed.

Bioethics - The Compassion of Healthcare:

Bioethics in healthcare has a rich history, prioritising patient-centricity and medical values. It emphasises beneficence (promoting well-being), non-maleficence (avoiding harm), autonomy (respecting individual decisions), and justice (fair distribution of resources). These principles guide medical professionals in making ethical decisions. Yet, complex situations can still arise, demanding careful consideration.

Reconciling Both Perspectives:

Integrating bioethics and AI ethics offers a multi-level view of the ethical implications of AI in healthcare. By understanding and respecting the traditional vision of bioethics while considering the societal and systemic impacts of AI, we can address ethical challenges comprehensively.

Beneficence: Striving for well-being is key. AI interventions must contribute to patient welfare while benefiting society as a whole. For instance, AI-powered assisting diagnostic tools can improve healthcare outcomes and support doctor in their endeavours, benefiting patients and the community. Imagine a remote healthcare system that utilises AI to analyse patient data, enabling early detection of life-threatening conditions. This technology could ensures timely intervention for doctors and medical staff when alarmed, saving lives and improving overall public health.

Non-Maleficence: First, do no harm. Both bioethics and AI ethics share the commitment to minimize harm. We ensure responsible AI development to prevent negative consequences, safeguarding patient privacy, and security. In the development of AI-powered surgical assistants, meticulous safety protocols are followed to prevent any harm to patients during procedures. The technology complements the surgeon's skills, enhancing precision and reducing the risk of complications. On the other hand, solutions such as those studied by Louis in his PhD regarding data privacy are at the heart of avoiding harm for individuals’ privacy. Go Louis! 

Autonomy: Respecting individual choices. Balancing human and machine decision-making, we empower both patients and AI tools. AI assists doctors in decision-making without overshadowing their expertise. Foe example, AI-driven treatment recommendation systems could collaborate with doctors to present personalised treatment options for patients. Physicians use these recommendations as valuable insights while respecting their expertise and patients' preferences. On the other hand, autonomy also calls for patients’ consent to use AI. This topic thus relates really closely to transparency and explainability, as enlightened consent includes, in this case, knowing what the AI is about before agreeing to use it. 

Justice: Fairness for all. Bioethics promotes equal access to healthcare, while AI ethics calls for equitable distribution of AI's benefits. We prioritise fairness in resource allocation and strive to eliminate biases in AI applications. For instance, AI-driven allocation systems in organ transplant centers consider multiple factors, ensuring fair distribution of organs based on medical urgency and patient compatibility. The technology here could help mitigate biases and maximise life-saving opportunities for all patients if tested properly against bias. 

Explainability: This principle is stated explicitly in AI ethics, but is ongoingly present in all other principles in both bio and AI ethics. It means shedding light on the "black box" nature of AI solutions. If this principle is embraced entirely, developers of AI solutions must provide clear explanations of how the AI system works, including its rationale, methodology, and the data used to train it. This is especially vital in healthcare, where patients rightly want to know how AI detects specific illnesses and what factors influence health recommendations for example. It relates to the question of enlightened consent which can’t be disregarded. Transparency in AI decision-making ensures patients can trust and understand the technology that impacts their well-being. 

Conclusion - Paving the Way for Ethical AI in Healthcare

The integration of bioethics and AI ethics is possible and necessary in the ever-changing landscape of AI-driven healthcare. By recognising shared values and objectives, we can navigate ethical complexities with empathy and foresight. Ensuring AI enhances patient well-being while considering societal welfare is our collective responsibility.

Let's move forward together, embracing a future where ethical AI contributes positively to the health and well-being of all.

If you are interested to know more about this framework, you can refer to our newly published paper on the topic! And I have to say, this one is very close to my heart as it’s my first time being published as first author, and it was done with very good friends that showed a lot of support and taught me a lot! 

Stay curious, and until next time, keep exploring the frontiers of AI, and have a fantastic weekend! 

- Auxane Boch (TUM IEAI research associate, freelancer)

We are incredibly grateful that the newsletter is now read by over 12'000+ incredible human beings counting our email list and LinkedIn subscribers. Reach out to contact@louisbouchard.ai with any questions or details on sponsorships or visit my Passionfroot profile. Follow our newsletter at Towards AI, sharing the most exciting news, learning resources, articles, and memes from our Discord community weekly.

If you need more content to go through your week, check out the podcast!

Thank you for reading, and we wish you a fantastic week! Be sure to have enough rest and sleep!

Louis