www.ar-tiste.xyz
@rrtucci.bsky.social
We love Bayesian Networks and Causal Inference
How Mappa Mundi distinguishes between correlation and causation. causal inference, causal Ai, LLM
January 7, 2025 at 12:37 AM
How Mappa Mundi distinguishes between correlation and causation. causal inference, causal Ai, LLM
You are correct. Either the proofs of do calculus completeness & soundness are subtly wrong (it is possible; they are very complicated) or my method is subtly wrong. Below is the formula given by Cinelli. I think we should prove or disprove by computer simulation.
December 8, 2024 at 1:32 PM
You are correct. Either the proofs of do calculus completeness & soundness are subtly wrong (it is possible; they are very complicated) or my method is subtly wrong. Below is the formula given by Cinelli. I think we should prove or disprove by computer simulation.
Causality
LLM are what are called Transformer Nets (TN). TN are just curve fitters like least squares (math used for fitting a set of points with a line). But in the case of TN, they are fitting the space of all human generated text. That space is causally connected(CC). TN are very good fitting CC
LLM are what are called Transformer Nets (TN). TN are just curve fitters like least squares (math used for fitting a set of points with a line). But in the case of TN, they are fitting the space of all human generated text. That space is causally connected(CC). TN are very good fitting CC
February 24, 2024 at 12:37 AM
Causality
LLM are what are called Transformer Nets (TN). TN are just curve fitters like least squares (math used for fitting a set of points with a line). But in the case of TN, they are fitting the space of all human generated text. That space is causally connected(CC). TN are very good fitting CC
LLM are what are called Transformer Nets (TN). TN are just curve fitters like least squares (math used for fitting a set of points with a line). But in the case of TN, they are fitting the space of all human generated text. That space is causally connected(CC). TN are very good fitting CC
They are very closely related. In Causal Inference, the first and hardest step is deciding on a "causal DAG" (DAG=directed acyclic graph) After you have a DAG, the rest is just coasting downhill from there.
The DAG is the hypothesis part of the scientific method. DAGs can and should be tested
The DAG is the hypothesis part of the scientific method. DAGs can and should be tested
February 23, 2024 at 9:52 PM
They are very closely related. In Causal Inference, the first and hardest step is deciding on a "causal DAG" (DAG=directed acyclic graph) After you have a DAG, the rest is just coasting downhill from there.
The DAG is the hypothesis part of the scientific method. DAGs can and should be tested
The DAG is the hypothesis part of the scientific method. DAGs can and should be tested
This is what is needed
January 27, 2024 at 4:14 AM
This is what is needed
Causal AI, Causal Inference
November 13, 2023 at 5:05 AM
Causal AI, Causal Inference