Gang Fang's Blog

One more question on this statement: “Importantly, when the reward function R represents the product of a prior (over some random variable) times a likelihood (measuring how well that choice of value of the random variable fits some data), the GFlowNet will learn to sample from the corresponding Bayesian posterior.” What does the random variable…

I am reading… when the reward function R represents the product of a prior (over some random variable) times a likelihood (measuring how well that choice of value of the random variable fits some data), the GFlowNet will learn to sample from the corresponding Bayesian posterior https://milayb.notion.site/95434ef0e2d94c24aab90e69b30be9b3 In Bayesian Statistics, why do we go through…

How does GFlowNet represent distributions and sampling over graphs? Related description in GFlowNet tutorial: “Because of the sequential construction of objects x (with an arbitrary number of steps), it is very convenient to generate variable-size structured objects, like sets, graphs, lists, programs or other recursively constructed data structures.” My thought: that means GFlowNet is capable…

Note: a difference definition provided in Yoshua Bengio’s blog post scaling-in-the-service-of-reasoning-model-based-ml: A causal model can be understood as an exponentially large family of distributions indexed by a choice of intervention, … How can this be useful? Compared to Peter Spirtes’ one (i.e., a stat model and a causal graph) My thought: I don’t think I…

Now I have finished studying Peter Spirtes’ Introduction to Causal Inference. I am coming back to these big questions. The bigger question to ask when I study causal inference: what principles I can draw from the field of causal inference to build causal understanding into Yoshua Bengio’s world models? My Thought: To see their relationship,…