Essentially every learning algorithm / inductive bias is applicable to this. E.g., think of learning a Bayes net where some of the variables are your actions. (How well it works is another question, and here causal theories are presumably very useful, but not exhaustive.)
@pmddomingos
-
Physics Laws as Master Algorithm: Intelligence Emergence
By
–
The laws of physics are the master algorithm: run them long enough and intelligence emerges.
-
Inductive Bias, Complete Data, and Causal Theories Recap
By
–
To recap:
– We always need prior assumptions/inductive biases, unless we have complete data (no free lunch theorem).
– If we have complete data, we don't need anything else for any rung.
– Causal theories are a type of inductive bias.
I guess that's a No answer to your question. -
Complete Data Includes Mental States and Counterfactual Questions
By
–
That's because the data is incomplete. Complete data includes the state of the actors' minds, and allows you to correctly answer counterfactual questions (which just correspond to different states of the world).
-
Incomplete Models: Action’s Role in Learning and Decision-Making
By
–
Those models are incomplete, because they don't include your actions. Convergence aside, and tautologically, if you know everything there's nothing left to learn. You can then make optimal decisions without causal assumptions.
-
Priors and inductive biases are equivalent concepts
By
–
Yes. (Again, priors and inductive biases are really the same thing.)
-
Network Effects: Three Types Driving Trillion-Dollar Businesses
By
–
Types of network effect:
– Benefits to you of having many other users (e.g., Microsoft)
– More data to learn from -> Better product (e.g., Google)
– Cost of rebuilding your network (e.g., Twitter)
Likely many more to be discovered, along with new trillion-dollar businesses. -
Need for Comprehensive Inductive Bias Zoo Reference
By
–
I cover the highest levels in "The Master Algorithm". Below that there are subfield-specific ones. But someone should really do a serious Inductive Bias Zoo, a la Scott Aaronson's Complexity Zoo.
-
Causal Theories as Inductive Bias in Machine Learning
By
–
On the contrary: it's by viewing causal theories as a type of inductive bias that ML types can become comfortable with using them. (As I said before, there's a taxonomy of inductive biases; by placing a bias in the appropriate node you don't confuse, you clarify.)
-
Contrived Datasets and Knowledge Transfer in Causal Analysis
By
–
That's a very contrived data set, which makes it very unpersuasive. You then have to transfer knowledge to this problem from similar ones with better data, which is what the "causal bias" is really doing.