.

Robert R. Tucci's Blog (32)

Causal AI dictum: A dataset is model-free

capcha-bugs

Time and time again, Judea Pearl makes the point on Twitter to neural net advocates that they are trying to do a provably impossible task, to derive a model from data. I could be wrong, but this is what I think he means.

When Pearl says "data", he is referring to what is commonly called a dataset. A dataset is a table of data, where all the entries of each column have the same units, and measure a single feature, and each row refers to one…

Continue

Added by Robert R. Tucci on March 24, 2021 at 4:32am — No Comments

Discontinuities in Causal AI. Classical & Quantum Detectors

leaping-chasm

A classic classical discontinuity: "Leaping the Chasm" (1886) by Ashley Bennett, son of photographer Henry Hamilton Bennett, jumping to "Stand Rock". See www.wisconsinhistory.org

I am happy to announce that today I added 2 new chapters to Bayesuvius (my free, open source book on Bayesian Networks). The titles of the 2 new chapters…

Continue

Added by Robert R. Tucci on February 7, 2021 at 6:30am — No Comments

New chapter on (in) Rubin’s Theory of Potential Outcomes

pearl-rubin Judea Pearl (left) and Donald Rubin (right). The 2  fathers of PO. (Photo shot in 2014 and pilfered by me from "The Book of Why", by Judea Pearl.)

I am happy to announce that today I added 3 new chapters to Bayesuvius (my free, open source book on Bayesian Networks). The titles of the 3 new chapters are

  1. Potential Outcomes
  2. Instrumental Variables
  3. Instrumental …
Continue

Added by Robert R. Tucci on January 26, 2021 at 11:30pm — No Comments

Quantum d-separation and quantum belief propagation à la Pearl

I've just written a new paper entitled: "Quantum d-separation and quantum belief propagation". Here is the abstract:

The goal of this paper is to generalize classical d-separation and classical Belief Propagation (BP) to the quantum realm. Classical d-separation is an essential ingredient of most of Judea Pearl's work. It is crucial to all 3 rungs of what Pearl calls the 3 rungs of Causation. So having a quantum version of d-separation and BP probably implies that most of…
Continue

Added by Robert R. Tucci on December 9, 2020 at 7:00pm — 1 Comment

Causal AI Kaggle Competition, Hosted by Big Pharma

Recently, someone (not me) suggested in a tweet to Judea Pearl, let's have a Kaggle** competition on the topic of Causal AI. I love the idea, so I am trying to promote it with this blog post. Big Pharma could fund such a competition for much less than what they pay for a TV commercial, and the technical insights they might gain from this exercise might prove to be invaluable to the pharmaceutical industry.…

Continue

Added by Robert R. Tucci on November 29, 2020 at 1:01pm — No Comments

When Daphne Koller met Judea Pearl. When Feynman met Dirac. When The Beatles met Bob Dylan.

Daphne Koller is the leader of a mega-startup (Insitro) that uses Machine Learning (do they use Causal Bayesian Networks???) to do drug research. She also co-founded Coursera with Andrew Ng, and she co-wrote with  Nir Friedman a 1200 page book about Probabilistic Graphical Models (e.g., Bayesian Networks)

Judea Pearl won a Turing award (commonly…

Continue

Added by Robert R. Tucci on November 29, 2020 at 1:00pm — No Comments

No Causation without representation!

My free book Bayesuvius now has 39 chapters. I had been postponing writing the chapters on Pearl causality until now, because I consider them to be the most important chapters in the whole book, and I wanted to nail them, to the best of my limited abilities. Well, I finally bit the bullet and wrote them. Please check them out, and send me feedback. I would especially like your opinion on  the chapters…

Continue

Added by Robert R. Tucci on October 9, 2020 at 9:00am — No Comments

Belief Propagation (Message Passing) for Classical and Quantum Bayesian Networks

My FREE book about Bayesian Networks, Bayesuvius, continues to grow. It currently has 33 chapters. The purpose of this blog post is to announce the release of a new Bayesuvius chapter on Belief Propagation (BP).

Belief Propagation (BP) (aka Message Passing)was first proposed in 1982 by Judea Pearl to simplify the exact evaluation of probability marginals of…

Continue

Added by Robert R. Tucci on September 1, 2020 at 10:30am — No Comments

Decision Trees & Bayesian Networks

third-man

Decision Trees (DTs) are certainly cool and it is not my intention to belittle them here. Compared to Bayesian Networks (bnets), they seem easier to construct. In fact, I just wrote a poem dedicated to Decision Trees. It starts "I think that I shall never see an AI as lovely as a decision tree". DTs do have some drawbacks, but, like I said, the purpose of this blog post is not to criticize them.

The real purpose of this post is to point out that there is a secret…

Continue

Added by Robert R. Tucci on August 16, 2020 at 6:30am — No Comments

How complicated projects like building a skyscraper or a rocket or a bridge manage so many details?

manhattan

It's a question I often wondered about as a child, as I marveled at NYC skyscrapers during my first visit there, or when I visited Cape Canaveral or the Golden Gate Bridge. I don't claim I will provide a full answer to this question in this short post. However, recently, I've been studying two types of tools that are very useful in handling the mindbogglingly numerous details of a complicated project. And I would like to tell you about those 2 tools here. So this is going to…

Continue

Added by Robert R. Tucci on August 15, 2020 at 9:30am — No Comments

Causal Reinforcement Learning under Bareinboim at Columbia Univ.

Elias Bareinboim

I recently wrote a popular article on Causal AI. Note that Causal AI is almost exclusively the province of Bayesian networks (bnets), not of their more puny brethren, neural networks (NNs). Indeed, Causal AI is based on Judea Pearl's theory of causality which is expressed via bnets, not NNs.

Something that I didn't mention in my…

Continue

Added by Robert R. Tucci on August 7, 2020 at 3:00am — No Comments

Prion Computer Viruses For China, With Love



Check out: When China Sees All, Chinese AI Is Creating an Axis of Autocracy, by Ross Andersen (The Atlantic, Sept 2020 print edition) And…

Continue

Added by Robert R. Tucci on July 31, 2020 at 1:06pm — No Comments

Quantum Neural Net, an Oxymoron

oxymoron

Venn diagram showing overlap between "quantum" and "neural network"

The term "quantum neural networks" is being used with increasing frequency by the quantum computing community. Maybe as a dishonest, bait-and-switch advertising strategy, this makes sense. However, from a scientific standpoint, "quantum neural network" is a very poor name choice for what is being alluded to here.

Artificial Neural Networks, often called "Neural Nets" for short, are supposed to…

Continue

Added by Robert R. Tucci on July 15, 2020 at 4:30am — No Comments

Simpson’s Paradox, the Bane of Clinical Trials

(This article is now a chapter of my github proto-book Bayesuvius)

Simpson's paradox is a recurring nightmare for all statisticians overseeing a clinical trial for a medicine. It is possible that if they leave out a certain "confounding" variable from a study, the study's conclusion on whether a medicine is effective or not, might be, without measuring that confounding variable, the opposite of what it would have been had that variable been measured. Statisticians have to enlist expert…

Continue

Added by Robert R. Tucci on July 9, 2020 at 7:00pm — No Comments

Added Chapter on Reinforcement Learning to my book “Bayesuvius” on Bayesian Networks

I just uploaded a new chapter to my github proto-book "Bayesuvius". This chapter deals with Reinforcement Learning (RL) done right, i.e., with Bayesian Networks :)

My chapter is heavily based on the excellent course notes for CS 285 taught at UC Berkeley by Prof. Sergey Levine. All I did was to translate some of those lectures into B net lingo.

During a recent conversation that I had on LinkedIn with some very smart Machine Learning experts, the experts opined that the fields…

Continue

Added by Robert R. Tucci on July 7, 2020 at 12:00pm — No Comments

QC Paulinesia, !(yet another book about quantum computing)

Today I uploaded to github the first version of my book “QC Paulinesia” about Quantum Computing. The title is a play on the surname of the famous quantum scientist Wolfgang Pauli and the word "Polynesia". The book is based on a paper with the same name that I wrote in 2004. 

Here is  its github repo. Here is the pdf for the book. Let me quote the current…

Continue

Added by Robert R. Tucci on June 27, 2020 at 1:00pm — No Comments

Generative Adversarial Networks (GANs) & Bayesian Networks

gan-asian Generative Adversarial Networks (GANs) software is software for producing forgeries and imitations of data (aka synthetic data, fake data). Human beings have been making fakes, with good or evil intent, of almost everything they possibly can, since the beginning of the human race. Thus, perhaps not too surprisingly, GAN software has been widely used since it was first proposed in this amazingly recent 2014 paper. To gauge how widely GAN software has been used so far, see, for…

Continue

Added by Robert R. Tucci on June 24, 2020 at 11:30pm — No Comments

Today I uploaded to github the first version of my book "Bayesuvius" about Bayesian Networks

vesuvius-from-pompei View of Mount Vesuvius form Pompeii

In the last week, I started a new book on classical (not quantum) Bayesian Networks. Today, I uploaded the first installment to github. Here is its github repo. Here is the pdf for the book. Let me quote the current…

Continue

Added by Robert R. Tucci on June 23, 2020 at 7:12pm — No Comments

Causal AI & Bayesian Networks



Bullet ripping through Jack of Hearts. High speed photo by MIT's Harold (Doc) Edgerton.











Artist Harry Clarke's 1919 illustration for "A Descent into the Maelström"



Introduction



We are all familiar with the dictum that "correlation does not imply causation". Furthermore, given a data file with samples of two variables…

Continue

Added by Robert R. Tucci on June 10, 2020 at 5:00pm — No Comments

Information Theory (Turbo Codes) & Bayesian Networks

An error correcting code (ECC) is a way of controlling errors in data that is being transmitted over an unreliable or noisy communication channel. In an ECC, the sender encodes the message with redundant information. The receiver is able to detect a limited number of errors, and to correct these errors without retransmission.

Turbo codes are a class of ECC that approach very closely the theoretical…

Continue

Added by Robert R. Tucci on June 4, 2020 at 4:04pm — No Comments

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service