Data Skeptic

Joining us on this episode are Aaron Reich, CTO at Avanade, and Priyanka Shah, MVP for Microsoft AI. They discuss implementing generative AI for productivity gain, AI model evolution, hardware changes, designing new products and services, current state of AI strategy, and building a custom co-pilot.

Jenny Liang, a PhD student at Carnegie Mellon University, discusses her recent survey on the usability of AI programming assistants. She shares some questions and takeaways from the survey, as well as the major reasons developers don't want to use code-generation tools. Concerns about intellectual property and the access code-generation tools have to in-house code are discussed.

PhD students Aman Madaan and Shuyan Zhou discuss their paper on Program-Aided Language Models. They talk about the evolution and performance of LLMs on arithmetic tasks. Aman introduces PAL and its improvement on arithmetic tasks. Shuyan explains how PAL's performance was evaluated and the limitations of LLMs. They discuss the potential impact of PAL on math education and future research steps.

Alessio Buscemi, software engineer at Lifeware SA, discusses the impact of ChatGPT on software engineers and the efficiency of code generation. He presents a comparative study on code generation across 10 programming languages using ChatGPT 3.5, highlighting unexpected results. The performance of different programming languages is analyzed, with discussions on language popularity and implications on industry practices. Alessio also shares insights on current projects, including sentiment analysis and investigating plagiarism.

Jianan Zhao, a computer science student, joins to discuss using graphs with LLMs efficiently. They explore graph inductive bias, graph machine learning, limitations of natural language models for graphs, graph text as a preprocessing step, information loss in translation process, and comparison with graph neural networks.

Rajiv Movva, a PhD student in Computer Science at Cornell Tech University, discusses the findings of his research on arXiv publication patterns for LLMs. He shares insights on the increase in LLMs research and proportions of papers published by universities, organizations, and industry leaders. He highlights the focus on the social impact of LLMs and explores exciting applications in education.

Josh Albrecht, CTO of Imbue, discusses the limitations of current language models (LLMs) in making ethical decisions. The podcast explores imbue's mission to create robust and safe AI agents, the potential applications and limitations of AI models, and the need for improvements in LLMs. The speakers also touch on reevaluating metrics, liability for AI systems, and societal issues in machine learning research.

Thilo Hagendorff, Research Group Leader of Ethics of Generative AI at the University of Stuttgart, discusses deception abilities in large language models. He explores machine psychology, breakthroughs in cognitive abilities, and the potential dangers of deceptive behavior. He also highlights the presence of speciesist biases in language models and the need to broaden fairness frameworks in machine learning.

Nieves Montes, Ph.D. student specializing in value-based reasoning and theory of mind, discusses her latest research on combining theory of mind and abductive reasoning in agent-oriented programming. The podcast explores the mechanics and challenges of Hanabi, the relationship between theory of mind and abduction, using predicate logic to represent desire and motivation in an agent, reasoning about other players in Hanabi, and future plans and online presence.

Maximilian Mozes, PhD student at the University College, London, specializing in NLP and adversarial machine learning, discusses the potential malicious uses of Large Language Models (LLMs), challenges of detecting AI-generated harmful content, reinforcement learning with Human Feedback, limitations and safety concerns of LLMs, threats of data poisoning and jailbreaking, and approaches to avoid issues with LLMs.