Profile picture
Judea Pearl
@yudapearl
, 3 tweets, 2 min read Read on Twitter
1/ Communication between CI and ML folks will improve drastically if we can translate sentences such as: "Bottou trains his NN under conditions ABC" into sentences of the form: "Given the conditional probabilities P(y|x,do(z)...)". After all, what do we get from "training" if ..
2/ if not conditional probabilities, both observational and interventional. Another benefit for the translation: theorems of impossibility. CI has developed a theory that tells us if certain tasks can be accomplished given information in the form of probabilities P(y|x, do(z)...
3/3
 We can use this theory to prevent disappointments from "training" schemes that lead to impossibilities. As far as I know, theories of what's possible or not possible were not developed (yet) for training schemes. Why not use what we have? eg,ucla.in/2Jc1kdD #Bookofwhy
Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to Judea Pearl
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Related hashtags

#Bookofwhy

More from @yudapearl see all

Profile picture
Judea Pearl
@yudapearl
1/4 I believe this set of slides reinforces what I tweeted earlier: -- its hard to cut the embilical chord to Mother-Stat. It occurred to me that this urgency to stay in Stat-womb was also the motivation behind the potential-outcome framework. The benefits
2/4 were obvious, nothing is new, Y_1 and Y_0 are ordinary variables, with some missing values, so what? Everything else is ordinary statistics. The price, of course, was (1) Everything was tied to experimental "treatments", not to "events" or absence of events, and (2) we need
3/4 to express knowledge in the language of {Y_1, Y_0} , namely, in the formidable language of "conditional ignorability". Some would argue: What's wrong in letting statisticians broaden the scope of statistics and then believe that "it is all statistics"? I believe PO is a good
Read 4 tweets
Profile picture
Judea Pearl
@yudapearl
@jd_wilko 1/n It is not really "provocative", just a gentle way of luring statisticians to cut the umbilical chord from Mother Stat. Instead of surgical do-operator, you condition on a variable F (force) which does the surgery for you. I used it in 1993 ucla.in/2pJqtW3 when
@jd_wilko 2/n I thought that statistician are not prepared for a surgery. Other researchers, too, labor to create the illusion of remaining in the stat-womb. E.g., Heckman etal created a fix-operator to enforce this illusion ucla.in/2L8OCyl. The folks at @harvardEpi still teach CI
@jd_wilko @HarvardEpi 3/3 by "imitating RCT's". It is a tough umbilical chord to cut. Denis Lindley was the only statistician I met who said: CUT! In all these schemes we still need to import information from outside the data, which is the key to realizing that we are out of the stat-womb. #Bookofwhy
Read 3 tweets
Profile picture
Judea Pearl
@yudapearl
1/3 Thanks for posting, Bruce, and thanks for offering an honest assessment of #Bookofwhy from the viewpoint of an enlightened economist. As you can probably guess, I am particularly interested in your comment: "His [Pearl] grasp of what economists, for example, understand ..
2/3 and don't understand about causal relationships is incomplete,". What is it that economists DO understand and that I assumed they DON'T? Has this part of "econ. understanding" been expressed formally in the econ. literature since Haavelmo and Cowell's Com.? Can economists
3/3 solve the toy problems posed to them here: ucla.in/2mhxKdO? I am genuinely trying to understand what they know that they are laboring to hide from us. E.g., Do they know which parameters can be identified by OLS? Which models have testable implications? etc. etc,...
Read 3 tweets

Related threads

Profile picture
afroRoboticist
@AfroRoboticist
Why Python?

A thread as requested by @hackSultan

Will try to make this as unbiased as possible. 😁

👇🏾
Python is a high-level programming language whose structure mimics the way humans think.
It abstracts away low-level stuff like memory management, pointers etc.. which are the main stay of other languages like C++, Java
It is dynamically typed. Don’t be perturbed. Basically when u say a language is statically typed, it means you need to declare your variables upfront. You need to tell the compiler how much memory to set aside to handle a variable in advance.
Read 24 tweets
Profile picture
Nunca vi 1 cientista
@_NV1C
#Ciência Básica é importante?

Vem com a gente nessa thread!

#UniversidadeePesquisa
Vamos pensar em um bolo de #chocolate 🎂
(2/6)
E os ingredientes? E a receita?
🍫🥚🥛📃
(3/6)
Read 6 tweets
Profile picture
Umar Alani
@umar_alani
There's so much more that health is missing through these few years of development.

So much more can be done through #communication and #data mining.

What exactly is drawing health back when other discipline have skyrocketed their fields in tech?

#HealthForAll #WHD2019
While Physicians, Nurses, Pharmacies and other health professionals need to do more through accumulating capabilities outside the core health discipline to take courses starting with maybe critical thinking or design thinking. Learn tech skills in analytics etc.

#HealthForAll
Preventive health can be finally the major health market when predictive analytics is capitalized on.

However, we still need a deep EHR system to gather enough data. These data are everywhere but putting them in a place is important for growth.

#HealthForAll #WHD2019 @WHO
Read 14 tweets
Profile picture
Dr. Murielle Ålund
@MurielleAlund
At our latest @MSU_IBIO postdoc meeting @HelenMcCreery presented her recent paper with @AmandaHund , @LizScordato et al. on the importance of #mentorship in #STEM & how to remediate to the lack of guidance on how to be a good mentor. Thread! 👇 (1/n)

onlinelibrary.wiley.com/doi/10.1002/ec…
I loved every bit of that paper! This is something that should really be spread to all institutions and read by grad students, postdocs and PIs alike! Read it, spread it, get discussions started at your department! #AcademicTwitter (2/n)
Starting with a survey of mentors & mentees at US research unis: 70% of mentees experienced breakdowns in mentoring relationships at some point, affecting their #mentalhealth , 39% reported frequent poor mentoring, but 70% of the mentors say that they rarely mentor poorly! (3/n)
Read 13 tweets
Profile picture
ALI
@AliSalari1965
Iranians are sharing this map of Iran to show how much of the Caspian Sea (in the north) the mullahs' are forfeiting at the 5th #CaspianSummit.
To the #world: Iranians do not recognize any such disgraceful deals signed by the mullahs' regime.
#ABC #NBC #FoxNews #CBS #BV
#Reuters
What the hell are you doing?!
Selling our #CaspianSea to Russia?!
You #Khamenei are doing any kinds of betrayal to save your regime but it's To the #world: Iranians do not recognize any such disgraceful deals signed by the mullahs' regime.
#AP #AFP #UPI #WashTimes #WashingtonPost
An updated map of the #CaspianSeaSellOut by Mullahs' regime.
Kazakhstan: around 30%
Turkmenistan: 21%
Russia: around 19%
Azerbaijan: around 19%
Iran: 11%
This goes to show how desperate the Iranian regime has become in the face of ongoing Iran Protests & U.S. sanctions.
Read 11 tweets
Profile picture
Sarah Jamie Lewis
@SarahJamieLewis
Given an adversarial environment I don't believe you can maintain strong - decentralized - consistency AND have timely confirmation between arbitrary parties AND scale up to a large number of nodes/transactions.

Pick 2.

(fwiw I hope this is wrong, but I don't see how)
Maybe some of those qualifications don't matter...maybe communication between parties are largely not arbitrary, maybe weaker models of consensus are practically viable...
But there is an inherent tradeoff when decentralizing, the more nodes/transactions the more processing/communication between nodes, the longer it takes for any party to be aware of global state.
Read 5 tweets

Trending hashtags

#LeParisien #fixedincome #CNBC #Journalism #FOS #appeal #Tech #WashingtonPost #regulations #ensino #borrowing #bbcpapers #NowWatching #world #LegalNews #manchester #Conservatives #DailyTelegraph #Modi #sales #mentaltoughness #5news #RecruitmentNews #Halifax #Journalist
Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!