#causal — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #causal, aggregated by home.social.
-
The estimand framework, formalized in ICH E9(R1) regulatory guidance, provides a structured approach to define scientific objectives with precision. We apply the estimand framework to dose–exposure–response analyses. ... strategy to improve exposure–response analyses for dose selection, particularly when the relevant evidence includes data from multiple studies.
#estimand #exposure-response #dose-response #causal #pharmacometrics #pmx
-
The estimand framework, formalized in ICH E9(R1) regulatory guidance, provides a structured approach to define scientific objectives with precision. We apply the estimand framework to dose–exposure–response analyses. ... strategy to improve exposure–response analyses for dose selection, particularly when the relevant evidence includes data from multiple studies.
#estimand #exposure-response #dose-response #causal #pharmacometrics #pmx
-
The estimand framework, formalized in ICH E9(R1) regulatory guidance, provides a structured approach to define scientific objectives with precision. We apply the estimand framework to dose–exposure–response analyses. ... strategy to improve exposure–response analyses for dose selection, particularly when the relevant evidence includes data from multiple studies.
#estimand #exposure-response #dose-response #causal #pharmacometrics #pmx
-
The estimand framework, formalized in ICH E9(R1) regulatory guidance, provides a structured approach to define scientific objectives with precision. We apply the estimand framework to dose–exposure–response analyses. ... strategy to improve exposure–response analyses for dose selection, particularly when the relevant evidence includes data from multiple studies.
#estimand #exposure-response #dose-response #causal #pharmacometrics #pmx
-
The estimand framework, formalized in ICH E9(R1) regulatory guidance, provides a structured approach to define scientific objectives with precision. We apply the estimand framework to dose–exposure–response analyses. ... strategy to improve exposure–response analyses for dose selection, particularly when the relevant evidence includes data from multiple studies.
#estimand #exposure-response #dose-response #causal #pharmacometrics #pmx
-
The vast datasets on which LLMs are trained are repositories of human linguistic activity, imbued with the collective referential histories of countless speakers. Even if an #LLM does not "know" these histories in a human-like conscious sense, the statistical patterns it learns from this data implicitly encode these #causal-historical links.
https://jordivitria.substack.com/p/meaningful-language-understanding
-
The vast datasets on which LLMs are trained are repositories of human linguistic activity, imbued with the collective referential histories of countless speakers. Even if an #LLM does not "know" these histories in a human-like conscious sense, the statistical patterns it learns from this data implicitly encode these #causal-historical links.
https://jordivitria.substack.com/p/meaningful-language-understanding
-
The vast datasets on which LLMs are trained are repositories of human linguistic activity, imbued with the collective referential histories of countless speakers. Even if an #LLM does not "know" these histories in a human-like conscious sense, the statistical patterns it learns from this data implicitly encode these #causal-historical links.
https://jordivitria.substack.com/p/meaningful-language-understanding
-
The vast datasets on which LLMs are trained are repositories of human linguistic activity, imbued with the collective referential histories of countless speakers. Even if an #LLM does not "know" these histories in a human-like conscious sense, the statistical patterns it learns from this data implicitly encode these #causal-historical links.
https://jordivitria.substack.com/p/meaningful-language-understanding
-
The vast datasets on which LLMs are trained are repositories of human linguistic activity, imbued with the collective referential histories of countless speakers. Even if an #LLM does not "know" these histories in a human-like conscious sense, the statistical patterns it learns from this data implicitly encode these #causal-historical links.
https://jordivitria.substack.com/p/meaningful-language-understanding
-
Meta buried 'causal' evidence of social media harm, US court filings allege
#HackerNews #Meta #Causal #Evidence #SocialMedia #Harm #CourtFilings #News
-
Meta buried 'causal' evidence of social media harm, US court filings allege
#HackerNews #Meta #Causal #Evidence #SocialMedia #Harm #CourtFilings #News
-
Meta buried 'causal' evidence of social media harm, US court filings allege
#HackerNews #Meta #Causal #Evidence #SocialMedia #Harm #CourtFilings #News
-
Meta buried 'causal' evidence of social media harm, US court filings allege
#HackerNews #Meta #Causal #Evidence #SocialMedia #Harm #CourtFilings #News
-
Meta buried 'causal' evidence of social media harm, US court filings allege
#HackerNews #Meta #Causal #Evidence #SocialMedia #Harm #CourtFilings #News
-
CW: Long HiveMind request: Guidelines for causal DAGs in Bayesian modelling
Fediverse #HiveMind request: Guidelines for causal DAGs in Bayesian modelling
One of my many personal peeves is sloppy conceptual diagramming. I have been exposed to far too many box and arrow diagrams where there is no clear and consistent interpretation of what types of things the boxes are supposed to represent (and ditto for the arrows).
I am starting to engage with a Bayesian modelling project (this I know from nothing - https://tomlehrersongs.com/wp-content/uploads/2018/12/lobachevsky.pdf) and am being exposed to causal DAGs. I presume these are reasonably consistent and interpretable. Nonetheless, they are typically presented as though the interpretation is self-evident and I keep coming up with multiple possible interpretations/constraints.
>>>>> Do people have recommendations for *detailed* interpretations of causal DAGs as used in Bayesian modelling? <<<<<
Issues covered should include things like:
* Constraints on node types: If nodes represent variables can they have multivariate values or are they constrained to be univariate?
* Interpretation with respect to "cases": Should the DAG be interpreted as referring to relationships that hold on a per-case basis?
* Interpretation with respect to "populations": If DAGs represent per-case relations, do they also represent populatiopns of those cases?
* Interpretation with respect to data matrices: How does a causal DAG map onto a matrix of data to be modelled?
* Interpretation with respect to multiple types of "case": Repeat all of the issues above but when there are multiple types of case not necessarily in one-to-one relationships (e.g. multi-level modelling).
* What is the relationship between the DAG and the equations/statements implementing that DAG in the probabilistic programming language of choice (e.g. STAN, PyMC)?
* What are the constraints on the temporal relationships of the variables at the nodes? Assuming the variables are measures at points in time you (presumably) don't want causes coming after their effects. Different variables may be measured at different points in time and a measurement at some point in time may be a measure of some cumulative process over prior times - so I would expect some nuanced consideration of temporal constraints.#Bayesian #modelling #statistics #causal #DAG #DirectedAcylicGraph #diagram #interpretation
-
CW: Long HiveMind request: Guidelines for causal DAGs in Bayesian modelling
Fediverse #HiveMind request: Guidelines for causal DAGs in Bayesian modelling
One of my many personal peeves is sloppy conceptual diagramming. I have been exposed to far too many box and arrow diagrams where there is no clear and consistent interpretation of what types of things the boxes are supposed to represent (and ditto for the arrows).
I am starting to engage with a Bayesian modelling project (this I know from nothing - https://tomlehrersongs.com/wp-content/uploads/2018/12/lobachevsky.pdf) and am being exposed to causal DAGs. I presume these are reasonably consistent and interpretable. Nonetheless, they are typically presented as though the interpretation is self-evident and I keep coming up with multiple possible interpretations/constraints.
>>>>> Do people have recommendations for *detailed* interpretations of causal DAGs as used in Bayesian modelling? <<<<<
Issues covered should include things like:
* Constraints on node types: If nodes represent variables can they have multivariate values or are they constrained to be univariate?
* Interpretation with respect to "cases": Should the DAG be interpreted as referring to relationships that hold on a per-case basis?
* Interpretation with respect to "populations": If DAGs represent per-case relations, do they also represent populatiopns of those cases?
* Interpretation with respect to data matrices: How does a causal DAG map onto a matrix of data to be modelled?
* Interpretation with respect to multiple types of "case": Repeat all of the issues above but when there are multiple types of case not necessarily in one-to-one relationships (e.g. multi-level modelling).
* What is the relationship between the DAG and the equations/statements implementing that DAG in the probabilistic programming language of choice (e.g. STAN, PyMC)?
* What are the constraints on the temporal relationships of the variables at the nodes? Assuming the variables are measures at points in time you (presumably) don't want causes coming after their effects. Different variables may be measured at different points in time and a measurement at some point in time may be a measure of some cumulative process over prior times - so I would expect some nuanced consideration of temporal constraints.#Bayesian #modelling #statistics #causal #DAG #DirectedAcylicGraph #diagram #interpretation
-
CW: Long HiveMind request: Guidelines for causal DAGs in Bayesian modelling
Fediverse #HiveMind request: Guidelines for causal DAGs in Bayesian modelling
One of my many personal peeves is sloppy conceptual diagramming. I have been exposed to far too many box and arrow diagrams where there is no clear and consistent interpretation of what types of things the boxes are supposed to represent (and ditto for the arrows).
I am starting to engage with a Bayesian modelling project (this I know from nothing - https://tomlehrersongs.com/wp-content/uploads/2018/12/lobachevsky.pdf) and am being exposed to causal DAGs. I presume these are reasonably consistent and interpretable. Nonetheless, they are typically presented as though the interpretation is self-evident and I keep coming up with multiple possible interpretations/constraints.
>>>>> Do people have recommendations for *detailed* interpretations of causal DAGs as used in Bayesian modelling? <<<<<
Issues covered should include things like:
* Constraints on node types: If nodes represent variables can they have multivariate values or are they constrained to be univariate?
* Interpretation with respect to "cases": Should the DAG be interpreted as referring to relationships that hold on a per-case basis?
* Interpretation with respect to "populations": If DAGs represent per-case relations, do they also represent populatiopns of those cases?
* Interpretation with respect to data matrices: How does a causal DAG map onto a matrix of data to be modelled?
* Interpretation with respect to multiple types of "case": Repeat all of the issues above but when there are multiple types of case not necessarily in one-to-one relationships (e.g. multi-level modelling).
* What is the relationship between the DAG and the equations/statements implementing that DAG in the probabilistic programming language of choice (e.g. STAN, PyMC)?
* What are the constraints on the temporal relationships of the variables at the nodes? Assuming the variables are measures at points in time you (presumably) don't want causes coming after their effects. Different variables may be measured at different points in time and a measurement at some point in time may be a measure of some cumulative process over prior times - so I would expect some nuanced consideration of temporal constraints.#Bayesian #modelling #statistics #causal #DAG #DirectedAcylicGraph #diagram #interpretation
-
CW: Long HiveMind request: Guidelines for causal DAGs in Bayesian modelling
Fediverse #HiveMind request: Guidelines for causal DAGs in Bayesian modelling
One of my many personal peeves is sloppy conceptual diagramming. I have been exposed to far too many box and arrow diagrams where there is no clear and consistent interpretation of what types of things the boxes are supposed to represent (and ditto for the arrows).
I am starting to engage with a Bayesian modelling project (this I know from nothing - https://tomlehrersongs.com/wp-content/uploads/2018/12/lobachevsky.pdf) and am being exposed to causal DAGs. I presume these are reasonably consistent and interpretable. Nonetheless, they are typically presented as though the interpretation is self-evident and I keep coming up with multiple possible interpretations/constraints.
>>>>> Do people have recommendations for *detailed* interpretations of causal DAGs as used in Bayesian modelling? <<<<<
Issues covered should include things like:
* Constraints on node types: If nodes represent variables can they have multivariate values or are they constrained to be univariate?
* Interpretation with respect to "cases": Should the DAG be interpreted as referring to relationships that hold on a per-case basis?
* Interpretation with respect to "populations": If DAGs represent per-case relations, do they also represent populatiopns of those cases?
* Interpretation with respect to data matrices: How does a causal DAG map onto a matrix of data to be modelled?
* Interpretation with respect to multiple types of "case": Repeat all of the issues above but when there are multiple types of case not necessarily in one-to-one relationships (e.g. multi-level modelling).
* What is the relationship between the DAG and the equations/statements implementing that DAG in the probabilistic programming language of choice (e.g. STAN, PyMC)?
* What are the constraints on the temporal relationships of the variables at the nodes? Assuming the variables are measures at points in time you (presumably) don't want causes coming after their effects. Different variables may be measured at different points in time and a measurement at some point in time may be a measure of some cumulative process over prior times - so I would expect some nuanced consideration of temporal constraints.#Bayesian #modelling #statistics #causal #DAG #DirectedAcylicGraph #diagram #interpretation
-
CW: Long HiveMind request: Guidelines for causal DAGs in Bayesian modelling
Fediverse #HiveMind request: Guidelines for causal DAGs in Bayesian modelling
One of my many personal peeves is sloppy conceptual diagramming. I have been exposed to far too many box and arrow diagrams where there is no clear and consistent interpretation of what types of things the boxes are supposed to represent (and ditto for the arrows).
I am starting to engage with a Bayesian modelling project (this I know from nothing - https://tomlehrersongs.com/wp-content/uploads/2018/12/lobachevsky.pdf) and am being exposed to causal DAGs. I presume these are reasonably consistent and interpretable. Nonetheless, they are typically presented as though the interpretation is self-evident and I keep coming up with multiple possible interpretations/constraints.
>>>>> Do people have recommendations for *detailed* interpretations of causal DAGs as used in Bayesian modelling? <<<<<
Issues covered should include things like:
* Constraints on node types: If nodes represent variables can they have multivariate values or are they constrained to be univariate?
* Interpretation with respect to "cases": Should the DAG be interpreted as referring to relationships that hold on a per-case basis?
* Interpretation with respect to "populations": If DAGs represent per-case relations, do they also represent populatiopns of those cases?
* Interpretation with respect to data matrices: How does a causal DAG map onto a matrix of data to be modelled?
* Interpretation with respect to multiple types of "case": Repeat all of the issues above but when there are multiple types of case not necessarily in one-to-one relationships (e.g. multi-level modelling).
* What is the relationship between the DAG and the equations/statements implementing that DAG in the probabilistic programming language of choice (e.g. STAN, PyMC)?
* What are the constraints on the temporal relationships of the variables at the nodes? Assuming the variables are measures at points in time you (presumably) don't want causes coming after their effects. Different variables may be measured at different points in time and a measurement at some point in time may be a measure of some cumulative process over prior times - so I would expect some nuanced consideration of temporal constraints.#Bayesian #modelling #statistics #causal #DAG #DirectedAcylicGraph #diagram #interpretation
-
I have come to appreciate the food from In the Wood in Berkeley!
Here is my full blog post:
https://bayareaexploreer.blogspot.com/2025/09/in-wood.html
#americanfood #burgers #pizzas #yummy #bayarea #berkeley #drinks #causal #food
-
I have come to appreciate the food from In the Wood in Berkeley!
Here is my full blog post:
https://bayareaexploreer.blogspot.com/2025/09/in-wood.html
#americanfood #burgers #pizzas #yummy #bayarea #berkeley #drinks #causal #food
-
Collecting All Causal Knowledge
#HackerNews #Collecting #Causal #Knowledge #Causality #Research #Knowledge #Graphs #Data #Science
-
Collecting All Causal Knowledge
#HackerNews #Collecting #Causal #Knowledge #Causality #Research #Knowledge #Graphs #Data #Science
-
Collecting All Causal Knowledge
#HackerNews #Collecting #Causal #Knowledge #Causality #Research #Knowledge #Graphs #Data #Science
-
Collecting All Causal Knowledge
#HackerNews #Collecting #Causal #Knowledge #Causality #Research #Knowledge #Graphs #Data #Science
-
Collecting All Causal Knowledge
#HackerNews #Collecting #Causal #Knowledge #Causality #Research #Knowledge #Graphs #Data #Science
-
Determinants and causal effects of admission to selective private colleges [pdf]
https://www.nber.org/system/files/working_papers/w31492/w31492.pdf
#HackerNews #Determinants #causal #effects #selective #private #colleges #education #research
-
Determinants and causal effects of admission to selective private colleges [pdf]
https://www.nber.org/system/files/working_papers/w31492/w31492.pdf
#HackerNews #Determinants #causal #effects #selective #private #colleges #education #research
-
Determinants and causal effects of admission to selective private colleges [pdf]
https://www.nber.org/system/files/working_papers/w31492/w31492.pdf
#HackerNews #Determinants #causal #effects #selective #private #colleges #education #research
-
Determinants and causal effects of admission to selective private colleges [pdf]
https://www.nber.org/system/files/working_papers/w31492/w31492.pdf
#HackerNews #Determinants #causal #effects #selective #private #colleges #education #research
-
Determinants and causal effects of admission to selective private colleges [pdf]
https://www.nber.org/system/files/working_papers/w31492/w31492.pdf
#HackerNews #Determinants #causal #effects #selective #private #colleges #education #research
-
'Learning causal graphs via nonlinear sufficient dimension reduction', by Eftychia Solea, Bing Li, Kyongwon Kim.
http://jmlr.org/papers/v26/24-0048.html
#causal #nonparametric #observational -
'Learning causal graphs via nonlinear sufficient dimension reduction', by Eftychia Solea, Bing Li, Kyongwon Kim.
http://jmlr.org/papers/v26/24-0048.html
#causal #nonparametric #observational -
'Learning causal graphs via nonlinear sufficient dimension reduction', by Eftychia Solea, Bing Li, Kyongwon Kim.
http://jmlr.org/papers/v26/24-0048.html
#causal #nonparametric #observational -
'Learning causal graphs via nonlinear sufficient dimension reduction', by Eftychia Solea, Bing Li, Kyongwon Kim.
http://jmlr.org/papers/v26/24-0048.html
#causal #nonparametric #observational -
'Learning causal graphs via nonlinear sufficient dimension reduction', by Eftychia Solea, Bing Li, Kyongwon Kim.
http://jmlr.org/papers/v26/24-0048.html
#causal #nonparametric #observational -
While the URI points to a Wired article entitled to be about the #Metaverse the article is actually about #industrial, #manufacturing and #robotics use of #DigitalTwins — that term is used frequently throughout the article. While LLM/GPT style #generative technology is mentioned, imagine, as you read the article, #causal methods for multi-modal, multiple model composable, causal digital twins
https://www.wired.com/story/the-metaverse-is-here-and-its-industrial/
-
While the URI points to a Wired article entitled to be about the #Metaverse the article is actually about #industrial, #manufacturing and #robotics use of #DigitalTwins — that term is used frequently throughout the article. While LLM/GPT style #generative technology is mentioned, imagine, as you read the article, #causal methods for multi-modal, multiple model composable, causal digital twins
https://www.wired.com/story/the-metaverse-is-here-and-its-industrial/
-
While the URI points to a Wired article entitled to be about the #Metaverse the article is actually about #industrial, #manufacturing and #robotics use of #DigitalTwins — that term is used frequently throughout the article. While LLM/GPT style #generative technology is mentioned, imagine, as you read the article, #causal methods for multi-modal, multiple model composable, causal digital twins
https://www.wired.com/story/the-metaverse-is-here-and-its-industrial/
-
While the URI points to a Wired article entitled to be about the #Metaverse the article is actually about #industrial, #manufacturing and #robotics use of #DigitalTwins — that term is used frequently throughout the article. While LLM/GPT style #generative technology is mentioned, imagine, as you read the article, #causal methods for multi-modal, multiple model composable, causal digital twins
https://www.wired.com/story/the-metaverse-is-here-and-its-industrial/
-
While the URI points to a Wired article entitled to be about the #Metaverse the article is actually about #industrial, #manufacturing and #robotics use of #DigitalTwins — that term is used frequently throughout the article. While LLM/GPT style #generative technology is mentioned, imagine, as you read the article, #causal methods for multi-modal, multiple model composable, causal digital twins
https://www.wired.com/story/the-metaverse-is-here-and-its-industrial/
-
'Recursive Causal Discovery', by Ehsan Mokhtarian, Sepehr Elahi, Sina Akbari, Negar Kiyavash.
http://jmlr.org/papers/v26/24-0384.html
#causal #discovery #observational -
'Recursive Causal Discovery', by Ehsan Mokhtarian, Sepehr Elahi, Sina Akbari, Negar Kiyavash.
http://jmlr.org/papers/v26/24-0384.html
#causal #discovery #observational -
'Recursive Causal Discovery', by Ehsan Mokhtarian, Sepehr Elahi, Sina Akbari, Negar Kiyavash.
http://jmlr.org/papers/v26/24-0384.html
#causal #discovery #observational
