Judea Pearl is a professor at UCLA and a winner of the Turing Award, that's generally recognized as the Nobel Prize of computing. He is one of the seminal figures in the field of artificial intelligence, computer science, and statistics. He has developed and championed probabilistic approaches to AI, including Bayesian Networks and profound ideas in causality in general. These ideas are important not just for AI, but to our understanding and practice of science. But in the field of AI, the idea of causality, cause and effect, to many, lies at the core of what is currently missing and what must be developed in order to build truly intelligent systems. For this reason, and many others, his work is worth returning to often. This conversation is part of the Artificial Intelligence podcast. If you would like to get more information about this podcast go to https://lexfridman.com/ai or connect with @lexfridman on Twitter, LinkedIn, Facebook, Medium, or YouTube where you can watch the video versions of these conversations. If you enjoy the podcast, please rate it 5 stars on Apple Podcasts or support it on Patreon. This episode is presented by Cash App. Download it (App Store, Google Play), use code "LexPodcast". Here's the outline of the episode. On some podcast players you should be able to click the timestamp to jump to that time. 00:00 - Introduction 03:18 - Descartes and analytic geometry 06:25 - Good way to teach math 07:10 - From math to engineering 09:14 - Does God play dice? 10:47 - Free will 11:59 - Probability 22:21 - Machine learning 23:13 - Causal Networks 27:48 - Intelligent systems that reason with causation 29:29 - Do(x) operator 36:57 - Counterfactuals 44:12 - Reasoning by Metaphor 51:15 - Machine learning and causal reasoning 53:28 - Temporal aspect of causation 56:21 - Machine learning (continued) 59:15 - Human-level artificial intelligence 1:04:08 - Consciousness 1:04:31 - Concerns about AGI 1:09:53 - Religion and robotics 1:12:07 - Daniel Pearl 1:19:09 - Advice for students 1:21:00 - Legacy
The host encourages listeners to leave reviews and recommendations for podcast episodes and introduces a special offer for Cash App users. The guest discusses the importance of causality in the development of AI and recommends his book, "Book of Why".
The idea that geometrical constructions and theorems can be expressed in terms of algebraic language was a revolutionary concept. Through analytic geometry, it's now possible to construct geometrical shapes using algebraic equations.
Philosophical discussion on the concepts of determinism and stochasticity in the universe, and the beauty and power of each approach. The question of free will and the ability to differentiate between machines and entities with free will is also explored.
The correlation between two variables can often lead to inaccurate assumptions about causation when observations are taken over a long period of time or involve multiple aggressive variables.
The hosts discuss the possibility of people falling asleep when their autonomous car is driving, and share stories related to sleeplessness and automobiles - from ancient Babylonian Kings to modern-day traffic.
The absence of mathematical equations to capture the idea of cause and effect relationships makes it challenging to handle models without them. Deep neural networks can be conceptualized as conditional probability estimators that help infer cause and effect relationships.
The podcast discusses the task of eliciting knowledge from an expert or through self-discovery, which can involve complex questions that ordinary cognition may not be able to answer. It also mentions the interpretation of doing X, where liberating things from earlier influences and subjecting them to the tyranny of muscles can help quantify initially unquantified knowledge parameters.
In this episode, the speaker discusses the role of models in understanding complex systems and how they can be used to infer causality without physically changing the system.
The discussion revolves around the complexity of learning causation in a world with millions of variables and how much causal information is required to understand a model.
Using the metaphor of a turtle shell and stars as holes, one can measure the radius of the Earth. This allows for a simple way to understand complex astronomical concepts.
The human ability to reason by metaphor is something that machines are not yet capable of, and it is a valuable tool for problem-solving, particularly in diverse populations and conditions where existing models may not be applicable.
In this podcast episode, the speaker discusses the limitations of machine learning in identifying shots in sports. While machines can be trained to recognize patterns based on factual data, they lack the ability to interpret subtle nuances and context that human analysts can pick up on.
This podcast discusses how communication in soccer can improve players' performance and the future potential of artificial intelligence to understand and convey sophisticated human concepts.
The key to aligning human and computer values is developing a model of the recipient based on empathy. This can be achieved by building a model of oneself and understanding how changes to that model will affect one's behavior.
The podcast discusses two key aspects of China's economic growth: how they managed to feed their population and triple it, and the lesser-known miracle of their technological advancements and innovations.
The speaker explores the idea of parenting a robot and the potential consequences of a lack of warmth and discipline, which could lead to negative behaviors. They also discuss the importance of parental figures in a robot's development due to the influence of metaphorical reasoning.
The best way to generate breakthrough ideas is to challenge conventional thinking and persevere through criticism. When dreaming of creating intelligent systems, it's important to remain open-minded and think critically about the potential impact of the technology.
In this podcast, Lex Fridman talks to Judea Pearl about the fundamental law of counterfactuals - understanding what 'if' means, and how it can help us comprehend the world.