Exploration of Neural Net Systems and Transformer Architectures in AI Evolution

Welcome to episode #932 of Six Pixels of Separation - The ThinkersOne Podcast.

Nicholas Mattei is an Associate Professor of Computer Science at Tulane University. Nick has dedicated his career to exploring the theory and practice of artificial intelligence, with a focus on decision-making systems for both autonomous agents and humans. He’s been into AI since he was young, influenced by his mother (a program coordinator at a science center, and his father, a professor). His early fascination with computers and programming laid the foundation for a career that has spanned prestigious institutions and cutting-edge projects. Before joining Tulane University, Nick worked as a Research Staff Member at the IBM TJ Watson Research Laboratory, where he tackled complex problems in question answering, natural language inference, decision making, and the ethical implications of AI on society. Prior to his time at IBM, he contributed to the AI & Algorithmic Decision Theory Group at Data61 in Australia. His work there and his involvement in organizing workshops on computational social choice highlight his commitment to advancing AI research. He also spent time at NASA in engineering where he was responsible for the design and implementation of control systems software for multiple small satellites. All cool stuff. At Tulane, Nick is not only advancing AI research but also emphasizing the importance of interdisciplinary collaboration. He works closely with the law school and sociology departments to study the societal impacts of AI. His leadership at the Tulane Center of Excellence for Community-Engaged AI demonstrates his dedication to ensuring that technological advancements benefit the wider community. Nick is passionate about education and mentorship, continually striving to inspire the next generation of AI researchers. His ability to bridge the gap between theoretical AI and practical applications makes him a valuable asset to both the academic and broader communities... and business thinkers like you. If you’re trying to understand the distinctions between traditional coding and AI, the nuances of machine learning, and the evolving landscape of AI technologies, this one is for you. Nick reflects on the challenges and opportunities presented by Generative AI, as well as his thoughts on the ethical deployment of AI systems, offering a nuanced perspective on the future of technology. His pragmatic approach to AI, coupled with his optimism about its potential to create value, provides a balanced view that is both insightful and thought-provoking at a time when we need it most. Enjoy the conversation...

• Running time: 1:06:08.
• Hello from beautiful Montreal.
• Subscribe over at Apple Podcasts.
• Please visit and leave comments on the blog - Six Pixels of Separation.
• Feel free to connect to me directly on Facebook here: Mitch Joel on Facebook.
• Check out ThinkersOne.
• or you can connect on LinkedIn.
• ...or on Twitter.
• Here is my conversation with Nicholas Mattei.
• Tulane University.
• Tulane Center of Excellence for Community-Engaged AI.
• Follow Nick on LinkedIn.
• Follow Nick on X.
• This week's music: David Usher 'St. Lawrence River'.

Takeaways

• Interdisciplinary work, particularly between computer science, law, and sociology, is crucial to understanding the societal impacts of AI.
• AI is broadly defined as making computers appear intelligent, covering tasks like Google Maps navigation and Netflix recommendations, while machine learning involves training computers to learn from data and make predictions without explicit programming.
• Traditional AI integrates data across sources, whereas generative AI predicts the next word in a sequence based on large datasets.
• Generative AI excels at repetitive writing tasks but struggles with truly innovative creativity, as it mimics patterns from its training data.
• Significant concerns exist about the ethical deployment of AI, particularly regarding data privacy, copyright issues, and potential societal impacts
• The transformer architecture, developed by Google, revolutionized AI by enabling faster and more efficient training of large language models.
• Skepticism exists about achieving artificial general intelligence (AGI), but aligning AI tools with human values and ensuring they support human decision-making is valuable.
• AI's integration into various domains, from criminal justice to creative industries, poses both opportunities and challenges, necessitating continuous adaptation and training.\
• Ongoing dialogue and exploration are essential to understanding and shaping the future of AI responsibly.

Chapters:

• 00:00 - Introduction and Nicholas's Background
• 06:00 - Defining AI and Machine Learning
• 12:00 - Generative AI Capabilities
• 20:00 - Ethical and Societal Impacts of AI
• 30:00 - Transformers and AI Development
• 35:00 - AI in Practice and Real-World Applications
• 45:00 - Future of AI and AGI
• 55:00 - Alignment and Ethical AI
• 58:00 - Conclusion and Future Discussions