For the last 50 years, we've designed cars to be safe...
For the 50th-percentile male.
Well, that's actually not 100% correct.
According to Stanford's report**, we introduced "female" crash test dummies in the 1960s, but...
1/
#CausalSky #ScienceSky #MLSky #EpiSky #StatSky #AISky
For the 50th-percentile male.
Well, that's actually not 100% correct.
According to Stanford's report**, we introduced "female" crash test dummies in the 1960s, but...
1/
#CausalSky #ScienceSky #MLSky #EpiSky #StatSky #AISky
November 5, 2025 at 8:16 AM
For the last 50 years, we've designed cars to be safe...
For the 50th-percentile male.
Well, that's actually not 100% correct.
According to Stanford's report**, we introduced "female" crash test dummies in the 1960s, but...
1/
#CausalSky #ScienceSky #MLSky #EpiSky #StatSky #AISky
For the 50th-percentile male.
Well, that's actually not 100% correct.
According to Stanford's report**, we introduced "female" crash test dummies in the 1960s, but...
1/
#CausalSky #ScienceSky #MLSky #EpiSky #StatSky #AISky
{DAGassist} let's you process a DAG and estimate models giving {dagitty} input and baseline model #CausalSky #rstats
cran.r-project.org/web/packages... I think {ggdag} has more to offer for analysis of a DAG, while {DAGassist} is a one-stop package for classifying variables and getting estimates
cran.r-project.org/web/packages... I think {ggdag} has more to offer for analysis of a DAG, while {DAGassist} is a one-stop package for classifying variables and getting estimates
October 30, 2025 at 7:16 PM
{DAGassist} let's you process a DAG and estimate models giving {dagitty} input and baseline model #CausalSky #rstats
cran.r-project.org/web/packages... I think {ggdag} has more to offer for analysis of a DAG, while {DAGassist} is a one-stop package for classifying variables and getting estimates
cran.r-project.org/web/packages... I think {ggdag} has more to offer for analysis of a DAG, while {DAGassist} is a one-stop package for classifying variables and getting estimates
All of the previous Causal Inference Interest Group seminar recordings are also now available on our new website!
cls-data.github.io/CIIG/
#CIIG #EpiSky #CausalSky
cls-data.github.io/CIIG/
#CIIG #EpiSky #CausalSky
October 21, 2025 at 3:57 PM
All of the previous Causal Inference Interest Group seminar recordings are also now available on our new website!
cls-data.github.io/CIIG/
#CIIG #EpiSky #CausalSky
cls-data.github.io/CIIG/
#CIIG #EpiSky #CausalSky
This paper’s been popping as “evidence” that you can’t do real #causalinference w/ obs data. To me it shows you need rigorous pre-specified design (in addition to the willingness to fold when your hypothesis is not possible to answer with the data at hand). #EpiSky, #CausalSky, #AcademicSky
October 22, 2025 at 2:59 PM
This paper’s been popping as “evidence” that you can’t do real #causalinference w/ obs data. To me it shows you need rigorous pre-specified design (in addition to the willingness to fold when your hypothesis is not possible to answer with the data at hand). #EpiSky, #CausalSky, #AcademicSky
Building DAGs for supply chain: always fun, sometimes a challenge
If you have experience with causal analysis for supply chains -- what's one lesson that you'd share with your intern on day 1?
For me it'd be: be explicit about time order
#CausalSky #SupplyChain #StatSky #MLSky
If you have experience with causal analysis for supply chains -- what's one lesson that you'd share with your intern on day 1?
For me it'd be: be explicit about time order
#CausalSky #SupplyChain #StatSky #MLSky
October 22, 2025 at 8:26 AM
Building DAGs for supply chain: always fun, sometimes a challenge
If you have experience with causal analysis for supply chains -- what's one lesson that you'd share with your intern on day 1?
For me it'd be: be explicit about time order
#CausalSky #SupplyChain #StatSky #MLSky
If you have experience with causal analysis for supply chains -- what's one lesson that you'd share with your intern on day 1?
For me it'd be: be explicit about time order
#CausalSky #SupplyChain #StatSky #MLSky
Can 5g of creatine improve your IQ scores?
Spoiler: Probably not, but here’s one thing that can improve your causal inferences
When we design and analyze randomized controlled trials (A/B tests) using the traditional frequentist framework, we often focus on p-values
1/
#CausalSky #StatSky #ABTest
Spoiler: Probably not, but here’s one thing that can improve your causal inferences
When we design and analyze randomized controlled trials (A/B tests) using the traditional frequentist framework, we often focus on p-values
1/
#CausalSky #StatSky #ABTest
October 22, 2025 at 7:18 AM
Can 5g of creatine improve your IQ scores?
Spoiler: Probably not, but here’s one thing that can improve your causal inferences
When we design and analyze randomized controlled trials (A/B tests) using the traditional frequentist framework, we often focus on p-values
1/
#CausalSky #StatSky #ABTest
Spoiler: Probably not, but here’s one thing that can improve your causal inferences
When we design and analyze randomized controlled trials (A/B tests) using the traditional frequentist framework, we often focus on p-values
1/
#CausalSky #StatSky #ABTest
📜 Paper: arxiv.org/abs/2510.14582
-------
**Complete given our problem definition
***Powerful, but not perfect: conditional independence testing can be challenging in finite samples
****In the worst-case scenario, for real-world sparse graphs scaling can be more friendly
8/8
#CausalSky
-------
**Complete given our problem definition
***Powerful, but not perfect: conditional independence testing can be challenging in finite samples
****In the worst-case scenario, for real-world sparse graphs scaling can be more friendly
8/8
#CausalSky
Local Causal Discovery for Statistically Efficient Causal Inference
Causal discovery methods can identify valid adjustment sets for causal effect estimation for a pair of target variables, even when the underlying causal graph is unknown. Global causal discovery metho...
arxiv.org
October 20, 2025 at 8:44 AM
📜 Paper: arxiv.org/abs/2510.14582
-------
**Complete given our problem definition
***Powerful, but not perfect: conditional independence testing can be challenging in finite samples
****In the worst-case scenario, for real-world sparse graphs scaling can be more friendly
8/8
#CausalSky
-------
**Complete given our problem definition
***Powerful, but not perfect: conditional independence testing can be challenging in finite samples
****In the worst-case scenario, for real-world sparse graphs scaling can be more friendly
8/8
#CausalSky
Looking for a causality-related PhD program?
@tobigerstenberg.bsky.social might have something for you!
#CausalSky #StatSky
@tobigerstenberg.bsky.social might have something for you!
#CausalSky #StatSky
The Causality in Cognition Lab at Stanford University is recruiting PhD students this cycle!
We are a supportive team who happened to wear bluesky appropriate colors for the lab photo (this wasn't planned). 💙
Lab info: cicl.stanford.edu
Application details: psychology.stanford.edu/admissions/p...
We are a supportive team who happened to wear bluesky appropriate colors for the lab photo (this wasn't planned). 💙
Lab info: cicl.stanford.edu
Application details: psychology.stanford.edu/admissions/p...
October 18, 2025 at 7:41 AM
Looking for a causality-related PhD program?
@tobigerstenberg.bsky.social might have something for you!
#CausalSky #StatSky
@tobigerstenberg.bsky.social might have something for you!
#CausalSky #StatSky
NEW PAPER!!! "Causal Machine Learning Methods and Use of Cross‐Fitting in Settings With High‐Dimensional Confounding"
led by Susie Ellul, with Stijn Vansteelandt & John Carlin
Published in Stats in Med
Check it out 👇
onlinelibrary.wiley.com/doi/10.1002/...
#EpiSky #CausalSky
led by Susie Ellul, with Stijn Vansteelandt & John Carlin
Published in Stats in Med
Check it out 👇
onlinelibrary.wiley.com/doi/10.1002/...
#EpiSky #CausalSky
Causal Machine Learning Methods and Use of Cross‐Fitting in Settings With High‐Dimensional Confounding
Observational epidemiological studies commonly seek to estimate the causal effect of an exposure on an outcome. Adjustment for potential confounding bias in modern studies is challenging due to the p...
onlinelibrary.wiley.com
September 25, 2025 at 2:35 AM
NEW PAPER!!! "Causal Machine Learning Methods and Use of Cross‐Fitting in Settings With High‐Dimensional Confounding"
led by Susie Ellul, with Stijn Vansteelandt & John Carlin
Published in Stats in Med
Check it out 👇
onlinelibrary.wiley.com/doi/10.1002/...
#EpiSky #CausalSky
led by Susie Ellul, with Stijn Vansteelandt & John Carlin
Published in Stats in Med
Check it out 👇
onlinelibrary.wiley.com/doi/10.1002/...
#EpiSky #CausalSky
What are the best books on causality and causal ML you've ever read?
Why these ones?
In 2018, I started reading everything on causality I could possibly find.
At that time I had a few successful large-scale ML projects on my account both in business and academia.
1/
#CausalSky #MLSky #StatSky
Why these ones?
In 2018, I started reading everything on causality I could possibly find.
At that time I had a few successful large-scale ML projects on my account both in business and academia.
1/
#CausalSky #MLSky #StatSky
September 19, 2025 at 8:29 AM
What are the best books on causality and causal ML you've ever read?
Why these ones?
In 2018, I started reading everything on causality I could possibly find.
At that time I had a few successful large-scale ML projects on my account both in business and academia.
1/
#CausalSky #MLSky #StatSky
Why these ones?
In 2018, I started reading everything on causality I could possibly find.
At that time I had a few successful large-scale ML projects on my account both in business and academia.
1/
#CausalSky #MLSky #StatSky
He did it before Double Machine Learning
I met with professor Mark van der Laan because I think his work is pretty incredible and it sometimes feels like a secret that only a few people know about, especially in industry.
1/
#CausalSky #StatSky #CausalInference
I met with professor Mark van der Laan because I think his work is pretty incredible and it sometimes feels like a secret that only a few people know about, especially in industry.
1/
#CausalSky #StatSky #CausalInference
September 12, 2025 at 8:29 AM
He did it before Double Machine Learning
I met with professor Mark van der Laan because I think his work is pretty incredible and it sometimes feels like a secret that only a few people know about, especially in industry.
1/
#CausalSky #StatSky #CausalInference
I met with professor Mark van der Laan because I think his work is pretty incredible and it sometimes feels like a secret that only a few people know about, especially in industry.
1/
#CausalSky #StatSky #CausalInference
what is your take on the definition of proxy variable?
Is it X = X* + e, .....OR..... X* = X + e ? whereby X is measured/observed, and X* is latent/unobserved , e is noise.
#causalsky #econsky #causalinference
Is it X = X* + e, .....OR..... X* = X + e ? whereby X is measured/observed, and X* is latent/unobserved , e is noise.
#causalsky #econsky #causalinference
November 9, 2024 at 4:43 PM
what is your take on the definition of proxy variable?
Is it X = X* + e, .....OR..... X* = X + e ? whereby X is measured/observed, and X* is latent/unobserved , e is noise.
#causalsky #econsky #causalinference
Is it X = X* + e, .....OR..... X* = X + e ? whereby X is measured/observed, and X* is latent/unobserved , e is noise.
#causalsky #econsky #causalinference
What do have gloves in common with world models and decision-making?
Dr Juan Orduz has an answer (a new episode)
1/n
#causalsky #causality #marketing #podcast
Dr Juan Orduz has an answer (a new episode)
1/n
#causalsky #causality #marketing #podcast
November 16, 2023 at 10:56 AM
What do have gloves in common with world models and decision-making?
Dr Juan Orduz has an answer (a new episode)
1/n
#causalsky #causality #marketing #podcast
Dr Juan Orduz has an answer (a new episode)
1/n
#causalsky #causality #marketing #podcast
Amazing turnout for the first #CIIG seminar to hear @igelstorm.bsky.social talking about income as a causal exposure. People have joined from UK, USA, France, Germany, Switzerland, Austria, Norway, Netherlands, India, Korea, Malawi, Brazil, Saudi Arabia, Chile!
#CausalSky #EpiSky
#CausalSky #EpiSky
January 20, 2025 at 3:19 PM
Amazing turnout for the first #CIIG seminar to hear @igelstorm.bsky.social talking about income as a causal exposure. People have joined from UK, USA, France, Germany, Switzerland, Austria, Norway, Netherlands, India, Korea, Malawi, Brazil, Saudi Arabia, Chile!
#CausalSky #EpiSky
#CausalSky #EpiSky
In recent years, following all the progress happening at the intersection of causality and deep learning has become increasingly difficult.
Rapid growth of the field, further accelerated by the emergence of large foundation models, has resulted in...
🧵 1/
#CausalSky #MLSky
Rapid growth of the field, further accelerated by the emergence of large foundation models, has resulted in...
🧵 1/
#CausalSky #MLSky
February 28, 2025 at 9:29 AM
In recent years, following all the progress happening at the intersection of causality and deep learning has become increasingly difficult.
Rapid growth of the field, further accelerated by the emergence of large foundation models, has resulted in...
🧵 1/
#CausalSky #MLSky
Rapid growth of the field, further accelerated by the emergence of large foundation models, has resulted in...
🧵 1/
#CausalSky #MLSky
This is one of the best real-world examples for why causality matters that I encountered so far.
And it's extremely simple.
1/
#CausalSky #DecisionMaking #Business
And it's extremely simple.
1/
#CausalSky #DecisionMaking #Business
September 17, 2025 at 8:31 AM
This is one of the best real-world examples for why causality matters that I encountered so far.
And it's extremely simple.
1/
#CausalSky #DecisionMaking #Business
And it's extremely simple.
1/
#CausalSky #DecisionMaking #Business
What irony.
...
Yesterday, I got a message from Instagram stating that one of my reels violated their community policy.
1/n
#CausalSky #causalAI #machinelearning
#CausalBanditsPodcast #causality #causalinference
youtube.com/shorts/UV-PF...
...
Yesterday, I got a message from Instagram stating that one of my reels violated their community policy.
1/n
#CausalSky #causalAI #machinelearning
#CausalBanditsPodcast #causality #causalinference
youtube.com/shorts/UV-PF...
🤖 How are your decisions, AI? [Backstory in the description]
What irony....Yesterday, I got a message from Instagram stating that one of my reels violated their community policy.Pretty surprised, I started reading to u...
youtube.com
January 24, 2024 at 12:10 PM
What irony.
...
Yesterday, I got a message from Instagram stating that one of my reels violated their community policy.
1/n
#CausalSky #causalAI #machinelearning
#CausalBanditsPodcast #causality #causalinference
youtube.com/shorts/UV-PF...
...
Yesterday, I got a message from Instagram stating that one of my reels violated their community policy.
1/n
#CausalSky #causalAI #machinelearning
#CausalBanditsPodcast #causality #causalinference
youtube.com/shorts/UV-PF...
Dave Lagnado is looking for a post doc to work on causal inference! The BR-UK team is doing some very cool stuff, so if you're currently looking for a job, check this out: www.ucl.ac.uk/work-at-ucl/...
March 5, 2025 at 5:10 PM
In 2025, burn the pyramid.
Climb the ladder instead.
Pyramids of evidence are a popular device for conveying the idea that different types of studies provide different strengths of evidence for causal claims.
🧵 1/
#causality #causalSky
Climb the ladder instead.
Pyramids of evidence are a popular device for conveying the idea that different types of studies provide different strengths of evidence for causal claims.
🧵 1/
#causality #causalSky
January 1, 2025 at 10:00 AM
In 2025, burn the pyramid.
Climb the ladder instead.
Pyramids of evidence are a popular device for conveying the idea that different types of studies provide different strengths of evidence for causal claims.
🧵 1/
#causality #causalSky
Climb the ladder instead.
Pyramids of evidence are a popular device for conveying the idea that different types of studies provide different strengths of evidence for causal claims.
🧵 1/
#causality #causalSky
Benchmark Me a River: A New Causal Discovery Benchmarking Kit is Here
Causal discovery from observational data is often deemed one of the most challenging tasks in causal modeling.
The limited availability of real-world causal discovery benchmarks doesn't make it any easier.
1/ 👇🏼
#CausalSky
Causal discovery from observational data is often deemed one of the most challenging tasks in causal modeling.
The limited availability of real-world causal discovery benchmarks doesn't make it any easier.
1/ 👇🏼
#CausalSky
July 21, 2025 at 8:29 AM
Benchmark Me a River: A New Causal Discovery Benchmarking Kit is Here
Causal discovery from observational data is often deemed one of the most challenging tasks in causal modeling.
The limited availability of real-world causal discovery benchmarks doesn't make it any easier.
1/ 👇🏼
#CausalSky
Causal discovery from observational data is often deemed one of the most challenging tasks in causal modeling.
The limited availability of real-world causal discovery benchmarks doesn't make it any easier.
1/ 👇🏼
#CausalSky
Join us on Monday 12 May for a social epi CIIG seminar with @mkellyirving.bsky.social with a focus on health inequalities!
Register to get Zoom link here:
turing-uk.zoom.us/meeting/regi...
#CIIG #EpiSky #CausalSky
Register to get Zoom link here:
turing-uk.zoom.us/meeting/regi...
#CIIG #EpiSky #CausalSky
DON'T MISS: "Moving through the fog: considering #causality in health inequalities research"
The next Alan Turing Institute Causal Inference Interest Group seminar with @mkellyirving.bsky.social on 12th May 2025 at 3pm BST!
Register at: turing-uk.zoom.us/meeting/regi...
#CIIG #EpiSky #CausalSky
The next Alan Turing Institute Causal Inference Interest Group seminar with @mkellyirving.bsky.social on 12th May 2025 at 3pm BST!
Register at: turing-uk.zoom.us/meeting/regi...
#CIIG #EpiSky #CausalSky
May 6, 2025 at 10:56 AM
Join us on Monday 12 May for a social epi CIIG seminar with @mkellyirving.bsky.social with a focus on health inequalities!
Register to get Zoom link here:
turing-uk.zoom.us/meeting/regi...
#CIIG #EpiSky #CausalSky
Register to get Zoom link here:
turing-uk.zoom.us/meeting/regi...
#CIIG #EpiSky #CausalSky
Tomorrow, my colleagues will present our work on LLM-assisted causal discovery for mapping hidden patterns in crystalline solids.
If you're in Utah, come and visit them!
#CausalSky
If you're in Utah, come and visit them!
#CausalSky
July 31, 2025 at 8:29 AM
Tomorrow, my colleagues will present our work on LLM-assisted causal discovery for mapping hidden patterns in crystalline solids.
If you're in Utah, come and visit them!
#CausalSky
If you're in Utah, come and visit them!
#CausalSky
Not all real-life decisions are about one dimensional outcomes.
When applying for a new job, we might care about the compensation, but also the company's mission or their work-from-home policy.
1/ 👇🏼
#CausalSky
When applying for a new job, we might care about the compensation, but also the company's mission or their work-from-home policy.
1/ 👇🏼
#CausalSky
June 2, 2025 at 8:29 AM
Not all real-life decisions are about one dimensional outcomes.
When applying for a new job, we might care about the compensation, but also the company's mission or their work-from-home policy.
1/ 👇🏼
#CausalSky
When applying for a new job, we might care about the compensation, but also the company's mission or their work-from-home policy.
1/ 👇🏼
#CausalSky