Judea Pearl: Causal Reasoning, Counterfactuals, and the Path to AGI | Lex Fridman Podcast #56

1. 0:00 – 15:00

   The following is a…

   1. LFLex Fridman0:00

      The following is a conversation with Judea Pearl, professor at UCLA, and the winner of the Turing Award that's generally recognized as the Nobel Prize of computing. He's 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 Beijing Networks, and profound ideas in causality in general. These ideas are important not just to AI, but to our understanding and practice of science. But in the field of AI, the idea of causality, cause and effect, to many, lie 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. I recommend his most recent book, called Book of Why, that presents key ideas from a lifetime of work in a way that is accessible to the general public. This is the Artificial Intelligence podcast. If you enjoy it, subscribe on YouTube, give it five stars on Apple Podcast, support it on Patreon, or simply connect with me on Twitter, @lexfridman, spelled F-R-I-D-M-A-N. If you leave a review on Apple Podcast especially, but also Castbox, or comment on YouTube, consider mentioning topics, people, ideas, questions, quotes in science, tech, and philosophy that you find interesting, and I'll read them on this podcast. I won't call out names, but I love comments with kindness and thoughtfulness in them, so I thought I'd share them with you. Someone on YouTube highlighted a quote from the conversation with Noam Chomsky, where he said that, "The significance of your life is something you create." I like this line as well. On most days, the existentialist approach to life is one I find liberating and fulfilling. I recently started doing ads at the end of the introduction. I'll do one or two minutes after introducing the episode, and never any ads in the middle that break the flow of the conversation. I hope that works for you and doesn't hurt the listening experience. This show is presented by Cash App, the number one finance app in the App Store. I personally use Cash App to send money to friends, but you can also use it to buy, sell, and deposit Bitcoin in just seconds. Cash App also has a new investing feature. You can buy fractions of a stock, say $1 worth, no matter what the stock price is. Brokerage services are provided by Cash App Investing, a subsidiary of Square and member SIPC. I'm excited to be working with Cash App to support one of my favorite organizations called FIRST, best known for their FIRST robotics and LEGO competitions. They educate and inspire hundreds of thousands of students in over 110 countries, and have a perfect rating on Charity Navigator, which means the donated money is used to the maximum effectiveness. When you get Cash App from the App Store or Google Play, and use code LEXPODCAST, you'll get $10, and Cash App will also donate $10 to FIRST, which again is an organization that I've personally seen inspire girls and boys to dream of engineering a better world. And now, here's my conversation with Judea Pearl. You mentioned in an interview that science is not a collection of facts, but a constant human struggle with the mysteries of nature. What was the first mystery that you can recall that hooked you, that captivated your curiosity?

   2. JPJudea Pearl3:30

      Oh, the first mystery. That's a good one. Yeah, I remember that.

   3. LFLex Fridman3:37

      What was it?

   4. JPJudea Pearl3:38

      I had a fever for three days, uh, when I learned about Descartes' analytic geometry, and I found out that you can do all the construction in geometry using algebra. And I couldn't get over it. I simply couldn't get out of bed. (laughs)

   5. LFLex Fridman3:58

      So, what- what kind of world does analytic geometry unlock?

   6. JPJudea Pearl4:02

      Well, it connects algebra with geo- geometry. Okay, so Descartes had the idea that, um, (clears throat) geometrical construction and geometrical theorems and the assumptions can be articulated in the language of algebra, which means that all the proof that we did in high school in trying to prove that the three bisectors meet at one point, and that... (laughs) Okay, uh, all this can be proven by just-

   7. LFLex Fridman4:36

      Through algebra.

   8. JPJudea Pearl4:36

      ... shuffling around notation. Uh, that was a-

   9. LFLex Fridman4:41

      The connection-

   10. JPJudea Pearl4:42

       ... traumatic experience.

   11. LFLex Fridman4:42

       (laughs) The tr- traumatic experience.

   12. JPJudea Pearl4:44

       For me, it was. I'm telling you, right?

   13. LFLex Fridman4:45

       So, it's the connection between the different mathematical disciplines, that they all-

   14. JPJudea Pearl4:50

       No, in between two diff- two different languages.

   15. LFLex Fridman4:51

       Just even... Languages?

   16. JPJudea Pearl4:53

       Yeah.

   17. LFLex Fridman4:54

       So, which mathematic discipline is the most beautiful? Is geometry it for you?

   18. JPJudea Pearl4:58

       Both are beautiful. They have, uh, almost the same power.

   19. LFLex Fridman5:02

       But there's a visual element to geometry, being a-

   20. JPJudea Pearl5:05

       Visually, (laughs) it's more transparent, but, uh, once you get over to algebra, then, uh, y- a linear equation is a straight line. This translation is easily absorbed, uh, and, um, the f- to pass a tangent to a circle, uh, you know, (laughs) you have the basic theorems and you can do it with algebra. So, but, uh, the transition from one to another was really ... I thought that Descartes was the greatest mathematician of all times. (laughs)

   21. LFLex Fridman5:37

       So, you have been at the ... if you think of engineering and mathematics as a spectrum-

   22. JPJudea Pearl5:43

       Yes.

   23. LFLex Fridman5:43

       ... uh, you have been ... you have walked casually along this spectrum throughout your- throughout your life. You know, you had a little bit of engineering, and then, you know, uh, you have b- done a little bit of mathematics here and there. (laughs)

   24. JPJudea Pearl5:58

       Well, not a little bit. I mean, we got a very solid background in mathematics because our teachers were geniuses.

   25. LFLex Fridman6:06

       Yeah.

   26. JPJudea Pearl6:07

       Our teachers came from Germany in the 1930s, running away from Hitler. Uh, they left their careers in Heidelberg and Berlin, and came to teach high school in Israel. And we were the beneficiary of that experiment. So, I al- when they taught us math, the good way.

   27. LFLex Fridman6:25

       What's the good way to teach math?

   28. JPJudea Pearl6:26

       Chronologically.

   29. LFLex Fridman6:29

       The people?

   30. JPJudea Pearl6:29

       The people behind the theorems, yeah. Their cousins, and their nieces (laughs) and their faces, and how they jumped from the bathtub when they scream, "Eureka!" (laughs) and ran naked in town. (laughs)
2. 15:00 – 30:00

   Right. So that's ...…

   1. JPJudea Pearl15:00

      when one of them, uh, stays the same. Now, staying the same means that I have chosen to look only at those incidents where the guy has the same value as previous one. It's my choice as an experimenter. So, things that are not correlated before could become correlated. Like, for instance, if I have two coins which are uncorrelated, okay, and I choose only those flippings experiments in which a bell rings, and the bell rings when at least one of them is a tail, okay, then suddenly, I see correlation between the two coins, because I only look at the cases where the bell rang. You see, it's my design with my ignorance, essentially, with my, uh, audacity to ignore certain incidents, I suddenly create a correlation where it doesn't exist physically.

   2. LFLex Fridman16:07

      Right. So that's ... You just outlined one of the flaws of observing the world and, and trying to infer something fundamental about the world from looking at the correlation.

   3. JPJudea Pearl16:17

      I don't look at it as a flaw. The world works like that. Which mean- but the flaws comes if you try to impose-

   4. LFLex Fridman16:25

      Hmm.

   5. JPJudea Pearl16:25

      ... um, causal logic on correlation. It doesn't work too well.

   6. LFLex Fridman16:34

      I mean, but that's exactly what we do. That's what ... That has been the majority of science, is you're r-

   7. JPJudea Pearl16:40

      No, no, no. Majority of, of naive science. Statisticians know it. The statisticians know that if you condition on a third variable, then you can destroy or create correlations among two other variables.

   8. LFLex Fridman16:55

      Right.

   9. JPJudea Pearl16:55

      They know it. It's in the data.

   10. LFLex Fridman16:58

       Right.

   11. JPJudea Pearl16:58

       There's nothing to surprise them. That's why they all dismiss the Simpson paradox. "Ah, we know it." They don't know anything about it. (laughs)

   12. LFLex Fridman17:07

       Well, there's, uh, there's disciplines, like psychology, where all the variables are hard to ge- to account for. And so, uh, oftentimes, there is a leap between correlation to causation. You're, you're imposing-

   13. JPJudea Pearl17:18

       What do you mean a leap?

   14. LFLex Fridman17:21

       Uh-

   15. JPJudea Pearl17:21

       Who, who is trying to get causation from correlation? No one.

   16. LFLex Fridman17:25

       Not, not ... You're not proving causation-

   17. JPJudea Pearl17:27

       (laughs)

   18. LFLex Fridman17:28

       ... but you're sort of, uh, um, discussing it, implying, sort of hypothesizing without ability to prove.

   19. JPJudea Pearl17:35

       Well, which discipline we have in mind? I'll tell you if they are obsolete-

   20. LFLex Fridman17:40

       (laughs)

   21. JPJudea Pearl17:40

       ... or if they are outdated, or they're about to get outdated.

   22. LFLex Fridman17:44

       Yes.

   23. JPJudea Pearl17:44

       Or (laughs) -

   24. LFLex Fridman17:45

       Yes.

   25. JPJudea Pearl17:45

       Yeah, tell me which one you have in mind.

   26. LFLex Fridman17:46

       Well, psychology, you know? Uh-

   27. JPJudea Pearl17:48

       Psychology, what is it? SEM? Structural equation modeling?

   28. LFLex Fridman17:50

       No, no. I was thinking of applied psychology, studying, um ... for example, we work with human behavior in semi-autonomous vehicles, how people behave. And you have to conduct these studies of people driving cars.

   29. JPJudea Pearl18:04

       Everything starts with a question. What is a research question?

   30. LFLex Fridman18:07

       What is a research question?
3. 30:00 – 45:00

   Yeah. …

   1. LFLex Fridman30:00

      important research question.

   2. JPJudea Pearl30:01

      Yeah.

   3. LFLex Fridman30:01

      This is an important question. Then you- you d-

   4. JPJudea Pearl30:04

      No, no, I didn't construct a model yet. I just said it's important question.

   5. LFLex Fridman30:07

      It's important question.

   6. JPJudea Pearl30:07

      And the first exercise is express it mathematically. What do you want to prove? Like, if I tell you, "What's the eff- what will be the effect of taking this drug?" Okay? You have to say that in mathematics. How do you say that?

   7. LFLex Fridman30:22

      Yes.

   8. JPJudea Pearl30:23

      Can you write down the question? Not the answer.I want to find the effect of the drug on my headache.

   9. LFLex Fridman30:32

      Right.

   10. JPJudea Pearl30:32

       W- write down, right? Write it down.

   11. LFLex Fridman30:34

       That's where the do calculus comes in. (laughs)

   12. JPJudea Pearl30:35

       Yes. Do operator, what is do operator?

   13. LFLex Fridman30:38

       Do operator, yeah.

   14. JPJudea Pearl30:39

       Yeah. You have to have a-

   15. LFLex Fridman30:40

       Which is nice. It's the difference between association and intervention.

   16. JPJudea Pearl30:43

       Correct.

   17. LFLex Fridman30:43

       Very beautifully, sort of constructed.

   18. JPJudea Pearl30:45

       Yeah. So, we co- we have a do operator, so the do calculus connected on the do operator itself, connects the operation of doing to something that we can see.

   19. LFLex Fridman30:57

       Right. So, as opposed to the purely observing, you're making the choice to change a variable.

   20. JPJudea Pearl31:05

       Yeah. That's what it, it expresses.

   21. LFLex Fridman31:07

       Hm. Gotcha.

   22. JPJudea Pearl31:08

       And then, the way that we interpret it, and the mechan- mechanism by which we take your query, and we translate it into something that we can work with, is by giving it semantics. Saying that you have a model of the world, and you cut off all the incoming arrow into X, and you're looking now in the modified, mutilated model, you ask for the probability of Y. That is interpretation of doing X, because by doing things, you've liberated them from all influences that acted upon them earlier, and you subject them to the tyranny of your muscles.

   23. LFLex Fridman31:50

       So you... (laughs) You remove all the questions about causality by doing them.

   24. JPJudea Pearl31:55

       So, because-

   25. LFLex Fridman31:55

       You're now-

   26. JPJudea Pearl31:56

       ... it's one level of questions.

   27. LFLex Fridman31:58

       Yeah.

   28. JPJudea Pearl31:59

       Answer the questions about what will happen if you do things.

   29. LFLex Fridman32:01

       If you do, if you drink the coffee, if you take the aspirin.

   30. JPJudea Pearl32:04

       Right.
4. 45:00 – 1:00:00

   Mm-hmm. …

   1. JPJudea Pearl45:00

   2. LFLex Fridman45:00

      Mm-hmm.

   3. JPJudea Pearl45:01

      An expert... it's, it's mapping problem with which you are not familiar to a problem with which you are familiar.

   4. LFLex Fridman45:12

      Mm-hmm.

   5. JPJudea Pearl45:13

      Like, I'll give you a good example. The Greek believed that the sky is an opaque shell. It's not really outsp- in infinite space. It's an opaque shell, and the stars are holes-

   6. LFLex Fridman45:30

      Yeah.

   7. JPJudea Pearl45:30

      ... poked in the shell through which you see the eternal light. Okay? That was a metaphor. Why? Because they u- they understand how you poke holes in shells, okay? They're not f- they were not familiar with infinite space, okay? And it's, so, a- and, and we are walking on a shell of a turtle, and if you get too close to the edge, you're going to fall down to Hades, or wherever.

   8. LFLex Fridman45:56

      Yeah.

   9. JPJudea Pearl45:57

      Yeah? Um, that's a metaphor. It's not true. But this kind of metaphor enabled Aristarchus to measure the radius of the Earth.

   10. LFLex Fridman46:10

       Hm.

   11. JPJudea Pearl46:10

       Because he said, "Come on. If the w- we are walking on a turtle shell, then the ray of light coming through this angle will be different, um, this place, will be a different angle that's coming to this place. I know the distance. I'll measure the two, uh, angles, and then I have the radius of the shell of the, of the turtle."

   12. LFLex Fridman46:32

       Mm-hmm.

   13. JPJudea Pearl46:32

       Okay? And he did. And he found his measurement very close to the measurements we have today through the, uh, what, 6,000 and 700, 700 kilometers (laughs) of the Earth. That's something that would not h- occur to Babylonian astronomer, even though the Babylonian experiments were the machine learning people of the time. They fit curves, and they could predict the, um, eclipse of the moon much more accurately than the Greek, because they fit curve. Okay? Uh, so that's the difference in metaphor.

   14. LFLex Fridman47:18

       Mm-hmm.

   15. JPJudea Pearl47:19

       Something that you're familiar with. Again, a turtle shell. Okay? What does it mean if you are familiar? Familiar means that answers to certain questions are explicit. You don't have to derive them.

   16. LFLex Fridman47:35

       And they were made explicit, because somewhere in the past, you've constructed a model of that. Uh-

   17. JPJudea Pearl47:43

       Before, yeah, you- you're familiar with-

   18. LFLex Fridman47:44

       Yeah.

   19. JPJudea Pearl47:45

       So the child is familiar with billiard balls.

   20. LFLex Fridman47:47

       Yes.

   21. JPJudea Pearl47:48

       So the child could predict that if you let loose of one ball, the other one will bounce off. These are... You, you obtain that by, um, familiarity. Familiarity is answering questions, and you store the answer explicitly. You don't have to derive them. So, this is ideal for metaphor. All our life, all our intelligence is built around metaphors. Mapping from the unfamiliar to the familiar, but the, um, marriage between the two is a tough thing, which I, which we haven't yet been able to algorithmatize.

   22. LFLex Fridman48:24

       So, you think of that process of cas- of using metaphor to leap from one place to another as, we can call it reasoning? Is it a kind of reasoning?

   23. JPJudea Pearl48:35

       It is, uh, reasoning by metaphor, metaphorical reason-

   24. LFLex Fridman48:38

       Reasoning by metaphor.

   25. JPJudea Pearl48:38

       ... yeah.

   26. LFLex Fridman48:39

       Do you think of that as learning?So, learning is a popular terminology today in a narrow sense.

   27. JPJudea Pearl48:47

       It is, it is. It is definitely a form-

   28. LFLex Fridman48:49

       So you may not... Uh, okay, right.

   29. JPJudea Pearl48:51

       It's one of the most important learning, taking something which theoretically is derivable-

   30. LFLex Fridman48:57

       Yeah.
5. 1:00:00 – 1:15:00

   If there is test…

   1. LFLex Fridman1:00:00

      a good test? Is a Turing test a reasonable test?

   2. JPJudea Pearl1:00:04

      If there is test of free will, doesn't exist yet, uh, there's no-

   3. LFLex Fridman1:00:08

      How would you test free will? Uh, that's a-

   4. JPJudea Pearl1:00:11

      So far, we know only one thing. I mean (laughs) , if robots can communicate with reward and punishment among themselves, and hitting each other on the wrists and say, "You shouldn't have done that."

   5. LFLex Fridman1:00:27

      Mm-hmm.

   6. JPJudea Pearl1:00:27

      Okay? Um, playing better soccer because they can do that.

   7. LFLex Fridman1:00:34

      What do you mean, because they can do that?

   8. JPJudea Pearl1:00:36

      Because they can communicate among themselves. It's ƒ-

   9. LFLex Fridman1:00:38

      Because of the communication, they can do the soccer?

   10. JPJudea Pearl1:00:40

       Because they communicate like us.

   11. LFLex Fridman1:00:42

       Yeah.

   12. JPJudea Pearl1:00:42

       Reward and punishment. Yes, you didn't pass the ball the right, the right time and so forth, therefore you're going to sit on the bench for the next two... If they start communicating like that, the question is, will they play better soccer?As opposed to what? As opposed to what they do now, without this ability to reason about, uh, uh, reward and punishment, responsibility.

   13. LFLex Fridman1:01:06

       And a lot of factions.

   14. JPJudea Pearl1:01:08

       So far, I can only think about communication.

   15. LFLex Fridman1:01:11

       Communication is... And, and, and not necessarily natural language, but just communication.

   16. JPJudea Pearl1:01:16

       Just communication. And that's important to have a quick and effective means of communicating knowledge. If the coach tells you, "You should have passed the ball." Ping. He conveys so much knowledge to you, as opposed to, what? Go down and change your software. Right. That's the alternative. But the coach doesn't know your software.

   17. LFLex Fridman1:01:37

       Right.

   18. JPJudea Pearl1:01:37

       So, how can the coach tell you, "You should have passed the ball?" But that... (laughs) Our language is very effective. "You should have passed the ball." You know your software, you tweak the right module, okay? And next time, you don't do it.

   19. LFLex Fridman1:01:52

       Now, that's for playing soccer where the rules are well defined.

   20. JPJudea Pearl1:01:55

       No, no, no. Well, they're not well defined. When you should pass the ball and when you should stop-

   21. LFLex Fridman1:01:58

       Is not well defined.

   22. JPJudea Pearl1:02:00

       No, it's a-

   23. LFLex Fridman1:02:01

       Yeah.

   24. JPJudea Pearl1:02:02

       It's very soft world, very noisy.

   25. LFLex Fridman1:02:04

       It's... Yeah.

   26. JPJudea Pearl1:02:04

       Yes. You have to do it under pressure. (laughs)

   27. LFLex Fridman1:02:06

       An industry. It's art. Uh, but, uh, in terms of aligning values between computers and humans, do you think this cause and effect, uh, type of thinking is important to align the values? Values, morals-

   28. JPJudea Pearl1:02:24

       Mm-hmm.

   29. LFLex Fridman1:02:24

       ... ethics, under which the machines make decisions.

   30. JPJudea Pearl1:02:26

       Yeah.
6. 1:15:00 – 1:23:00

   So, s- seven years…

   1. JPJudea Pearl1:15:00

      we have an economical crisis, okay, you are capable of doing it too, and that worries me. I (laughs) , I want to believe that I'm not capable.

   2. LFLex Fridman1:15:12

      So, s- seven years after Daniel's death, you wrote an article, uh, at the Wall Street Journal, titled, Daniel Pearl and the Normalization of Evil.

   3. JPJudea Pearl1:15:19

      Yes.

   4. LFLex Fridman1:15:20

      What was your mess- uh, message back then, and how did it change today, over, over the years?

   5. JPJudea Pearl1:15:27

      I, I lost.

   6. LFLex Fridman1:15:30

      What was the message?

   7. JPJudea Pearl1:15:31

      The message was that, uh, we are not treating terrorism as a taboo. We are treating it as a bargaining device that is accepted, and people have grievance, and they go and sh- and, uh, bomb m- restaurants, okay? It's normal. Look, you're even not, not surprised when I tell you that. (laughs)

   8. LFLex Fridman1:15:59

      Yeah.

   9. JPJudea Pearl1:15:59

      20 years ago, you'd say, "What? For grievance, you go and blow up a restaurant?" (laughs) Today, it's become normalized. The banalization of evil. And we have created that to ourselves by normalizing, by all- uh, by making it part of, uh, political life. It's a political, uh, debate. Every terrorist yesterday becomes a freedom fighter today, and tomorrow, he'll become a terrorist again. It is switchable, okay?

   10. LFLex Fridman1:16:38

       Right, and so-

   11. JPJudea Pearl1:16:39

       Yeah.

   12. LFLex Fridman1:16:39

       ... you, we should call out evil when there's evil.

   13. JPJudea Pearl1:16:43

       We- if we don't want to be part of it.

   14. LFLex Fridman1:16:46

       Becoming.

   15. JPJudea Pearl1:16:47

       If you want... Yeah, if we want to separate good from evil, that's one of the first things that, uh, uh, what was it? In the Garden of Eden, remember, the first thing (laughs) that God tell them, uh, was, "Hey, you want some knowledge? Here is a tree of good and evil."

   16. LFLex Fridman1:17:08

       So, this evil touched your life personally. Does your heart have anger, sadness, or is it hope?

   17. JPJudea Pearl1:17:20

       Look, I see some beautiful people coming from Pakistan. I see beautiful people everywhere, but I see horrible propagation of evil in this country too. It shows you how populistic slogans can catch the mind of the best intellectuals.

   18. LFLex Fridman1:17:48

       Today is Father's Day.

   19. JPJudea Pearl1:17:50

       I didn't know that.

   20. LFLex Fridman1:17:51

       Yeah.

   21. JPJudea Pearl1:17:51

       Oh, no, I heard it, but-

   22. LFLex Fridman1:17:52

       What's a, what's, what's a fond memory you have of Daniel?

   23. JPJudea Pearl1:17:58

       Oh, many good memories, uh, immense. He was my mentor, really. He had, um, a sense of balance that I didn't have. (laughs)

   24. LFLex Fridman1:18:13

       (laughs) Yeah.

   25. JPJudea Pearl1:18:15

       He saw the beauty in every person. He was not as emotional as I am, and more looking at things in perspective. He really liked every person. He really grew up with the idea that a foreigner is a reason for curiosity, not for fear. (clears throat) This one time, we went in Berkeley, and, uh, homeless came out from some dark alley, and said, "Hey, man, could you spare a dime?" I retreated back, you know, two feet back, and Danny just hugged him and said, "Here's a dime. Enjoy yourself. Maybe he wants some (laughs) some, um, money to take a bus or whatever." "Where did you get it? Not from me." (laughs)

   26. LFLex Fridman1:19:08

       (laughs) Do you have advice for young minds today dreaming about creating, as you have dreamt, creating intelligent systems? What is the best way to arrive at new breakthrough ideas and carry them through the fire of criticism and, uh, and past conventional ideas?

   27. JPJudea Pearl1:19:28

       Ask your questions (laughs) freely. Your questions are never dumb.

   28. LFLex Fridman1:19:37

       (laughs)

   29. JPJudea Pearl1:19:37

       And solve them your own way, (laughs) okay? And don't take no for an answer. Look, if they are really dumb, you will find out quickly by trying an arrow to see that they're not leading any place. But follow them and try to understand things your way. That is my, uh, advice. I don't know if it- it's going to help anyone.

   30. LFLex Fridman1:20:04

       No, that brilliantly put.

Episode duration: 1:23:01