Adversarial images targeting an image classifier have been shown to be transferable to separately trained models (i.e., models with different weights or different architectures relative to the model for which the adversarial image was constructed to target).
I'm curious if the adversarial CA reprogramming techniques are similarly transferable. That is, do the adversarial CA and/or the adversarial perturbation matrix transfer to separate CAs (trained on the same task) with different weights or architectures than the original CA that was targeted?
Three factors I view as appealing, versus navigating to each site separately, include 1) the aggregation itself has value (e.g., to save time), 2) the primarily textual interface and 3) the search functionality.
Kind of negates the purpose of even using wine. If your running a windows machine to run a Linux VM then you might as well install straight to Windows and not deal with the underpowered VM
Also, last time I tried, the virtual graphics cards in the vm were much less powerful than a fully drivered up raw graphics card even on Linux, unless I was doing something wrong?
This has actually changed. A friend of mine plays all modern games with sometimes even better FPS using the PCI passthrough in QEMU[1], I have to say it's a pretty impressive setup and I've played Witcher 3 with Arch Linux + QEMU + Windows 7, solid 60fps with ultra graphics.
I was asking because the parent post seemed to be concerned about untrusted Windows binaries. A VM gives you the ultimate sandbox (even if it's not perfect).
Per your comment, I wrote a Chrome extension that creates links to other HN submissions pointing to the same URL. Hovering over the links provides additional details.
This provides similar functionality as the 'past' link in a submission's subtitle, but shows the duplicates automatically.
> "It is not ad hominem to bring someone's motivation and prior record into consideration."
I don't speak latin and don't use that expression often. Are you saying it's not ad hominem because the motive is considered as opposed to attacked? I ask because I had thought that it is ad hominem "to bring someone's motivation and prior record into consideration".
Here's the definition on Wikipedia: "a logical fallacy in which an argument is rebutted by attacking the character, motive, or other attribute of the person making the argument, or persons associated with the argument, rather than attacking the substance of the argument itself".
https://en.wikipedia.org/wiki/Ad_hominem
I didn't read the link of this post, or the link in your comment, so hopefully I didn't miss something. I was just curious since I thought "ad hominem" describes what you were saying is not ad hominem.
An Ad Hominem attack is misdirection. It's saying the war in Iraq was good because some reporter who said it was bad is a dick. Ad Hominem attacks are Non Sequiturs. It is not an Ad Hominem attack to suggest a reporter might not a credible source because of their history of being a dick.
Bringing motivations or character into consideration is relevant to gauging the reliability of a cited witness. The consideration of their reputation is not an attack on the subject matter, it is bringing into question whether they should be considered a reliable witness.
Well, you are wrong. Ad Hominem means, "X is wrong because Y making that claim is a jerk".
I am saying "X is a claim in a highly technical field that most of us lack expertise in. Y making that claim has a made large factual errors in prior publications, and his claims run contrary to the vast majority of his peers."
Sorry if you miss the nuance in my point. People also have referred to this concept as "taking with a grain of salt" or "heuristics".
Now, outside of arguments over semantics, the link article is lousy and the interpretation over what the lousy article claims by the great-grandparent post are incorrect.
The TL;DR of the linked article is: "They should have used data from another sensor, too. Based on 5 months of overlap, I'm going to assume I can just adjust the other sensor's data and use them in a single trend. Even if you accept that, I'm going to use a linear regression in this new cherry picked range, after making a big point about the nonlinearity of the underlying data. Even minimizing it the most I can, it still shows a 40 year trend of 3 mm annual sea level increases."
Are you saying Wikipedia's definition of ad hominem is wrong? It clearly says attacking a person's character or other attribute rather than the argument itself. A person's track record of mistakes is surely part of their character. A person's history agreement with others is too.
These really aren't compelling arguments. It's easy to call a crank a crank and dismiss him, but it's not easy to show why he's wrong. skeptics won't be satisfied with the personal attack that ignores the substance of the argument.
I interpreted your original comment similarly to your description in your reply, so I don't believe I missed nuance (unless I missed it twice). My comment and question were intended to clarify the meaning and usage of the expression "ad hominem", independent of this thread's topic and nuance in your comment. Like I mentioned, I don't use that expression often.
I'm curious if the adversarial CA reprogramming techniques are similarly transferable. That is, do the adversarial CA and/or the adversarial perturbation matrix transfer to separate CAs (trained on the same task) with different weights or architectures than the original CA that was targeted?
Is this something you've explored or plan to?