search for: counterfactual

...nt to >> do, where these "other transforms" pretend certain abnormal cases do >> not exist. Poison is a "stand-in" for these transforms, which are >> sometimes non-local. > > I'm saying that you still get reasonable behavior from reasoning about > counterfactuals for these things, and I haven't been convinced that you > lose optimization power except in code that actually exhibits undefined > behavior. I'm not sure I understood your concern fully, but you could have cases like: for (...) { if (condition){ i32 prod = x *nsw y i6...

Hi John, John McCall wrote: >>> Well, we could say non-nsw add and mul are actually "bitwise" operations, so "mul i32 poison, 2" is guaranteed to have its bottom bit zero (but "mul nsw i32 poison, 2" is poison). Of course, there's a limit to how far we can go in that direction, or we just end up with the old notion of undef. Off the top of my head,

...package management infrastructure as we can. To this end, I humbly submit a few small patches to that infrastructure. The first of the two patches below is the more important one; It adds to 'getDependencies' an 'installed' option, defaulting to NULL. This permits us to specify a counterfactual set of "installed packages". With this option in place, we can ask getDependencies "What would someone need, to install this package, if they only had -thus- installed". In practice, -thus-, for us, attempts to be "just the base packages". The second of the two patc...

Hello, I am running a model with four categories and want predicted probabilities in each category. Now for this example I wont give a counterfactual just the training data is fine but is there anyway to get a confidence interval around the predicted probabilities in each group? I have tried but it gives me probabilities and I have used interval="confidence", level=.095 and then interval = "prediction". Below is a sample of m...

...ed values), and the difference in the predicted values are averaged across all observations. So here's my questions: 1. (R-part:) How can I calculate AME_dummy in R? While predict() gets the fitted values for the actual observations, I'm not quite sure how to get predictions for the "counterfactual" part in AME_dummy, i.e. setting the dummy to one and then zero for each observation, while keeping the rest of the variables in the model at their observed values for each observation. 2. (Econometric part: ) Would it be ok to use the formula for the AME_cont even though the variable in ques...

...package management infrastructure as we can. To this end, I humbly submit a few small patches to that infrastructure. The first of the two patches below is the more important one; It adds to 'getDependencies' an 'installed' option, defaulting to NULL. This permits us to specify a counterfactual set of "installed packages". With this option in place, we can ask getDependencies "What would someone need, to install this package, if they only had -thus- installed". In practice, -thus-, for us, attempts to be "just the base packages". The second of the two patc...

...can come up with an example that does not depend on signed overflow being “undefined” ? Peter Lawrence. > > >> >> >> Hi John, On 11/11/16, 1:58 PM, Sanjoy Das wrote: >> I'm saying that you still get reasonable behavior from reasoning about >> counterfactuals for these things, and I haven't been convinced that you >> lose optimization power except in code that actually exhibits undefined >> behavior. > > I'm not sure I understood your concern fully, but you could have cases like: > > for (...) { > if (condition...

We're being ping-flooded by the Nachi worm, which probes subnets for systems to attack by sending 92-byte ping packets. Unfortunately, IPFW doesn't seem to have the ability to filter packets by length. Assuming that I stick with IPFW, what's the best way to stem the tide? --Brett Glass