apnorton

I agree with GEdgar. While some may consider it "silly" (@ScottSeidman, 2024), I find that quoting even a single word helps specify the exact nature of what is being claimed by the source.

There are certainly ways you can work with someone if you have a disagreement, but only sign your name to things that you stand by and think uphold your moral duty as a scientist --- both in terms of both the ethics of the process and the correctness of result.

@ivicts Intentional misrepresentation for personal gain is definitionally fraud.

@ivicts Yes, absolutely it's fraud. It's misrepresenting a difference in the dataset as an advantage/improvement in the model/algorithm, and lying about the difference in the dataset to cover it up. As analogy, imagine someone comparing the fuel economy of two cars, but one they tested while going up a 40% incline and towing a boat, while the other one they tested running downhill. Due to the difference in the testing methodology, any results that are claimed are meaningless --- same thing with the idea of "picking an easier dataset."

@ivicts You're being far too easy-going about this. Only writing papers that meet your ethical standard is not "too purist;" this is called having integrity. You absolutely can publish papers that you're confident meet your integrity standards, even as an undergrad. The idea that "maybe I should do a little bad to do a little good" is a slippery slope to all kinds of fraud. Any amount of fraud in a publication destroys the value of the whole thing, and --- rightfully --- jeopardizes your career. (Also, nobody in their right mind would ever suggest politicians as the model of sound ethics.)

It's really comforting to come back after a long period of time and see at least a few familiar faces... but also sad to see the missing ones :(

Is he gone entirely, or did he just stop attending chat?

And I never got to it in my classes :(

Yeah, that's a set of responsibility that I don't think I'd want anymore.

haha of course

:)

Still trying to narrow down a list of schools, but I think I want to do research in some subfield of cryptography

In fun news, I finally took the GRE and am starting to work on grad school apps!

But yeah, I think eschewing meta is a good idea; I had forgotten how insane the politics gets

(Basic summary post: meta.stackexchange.com/q/333965/201314)

One mod was removed by SE staff for what appears to be a potential to violate a future (but not current) code of conduct, and a lot of other mods didn't take kindly to that

Something like ~20 mods have resigned across the network in the last 24 hours :\

And good grief meta (global, not Math at least yet, fortunately) seems to be quite a mess today

I like to pop in every now and then to see how things are doing

Hey @TedShifrin!

> Chat guidelines: tinyurl.com/hzl2955 | LATEX in chat: tinyurl.com/cfqcvpc

In the info panel on the side there's a bookmarklet

(blargh I'll get the tex right eventually, lol)

Because $\neg P \rightarrow Q \equiv \neg(\neg P) \lor Q \equiv P \lor Q$.

Yep!

$\neg P \vee Q$ is certainly not equivalent to $P \vee Q$

P implies Q

"X" marks the spot, to within +/- Rrrrr

Arrrrrrg(z), me mateys!

:P

The Rrrrr language should walk the plank!

lol I needed something to laugh at... and that completely satisfied :)

I could model this in a simplified way as having a fixed, massive loss on each of those bad end states, but I think that would skew the resulting policy.

I have a process that allows reward to be gained at certain states, but particular end states eliminate the accumulated reward --- e.g. on a series of transitions, you could get +1, +4, +2 rewards, but if I end at a certain state I would then reset my net reward to 0. Ending on other states would allow me to keep all that reward.

Anyone know of some kind of extended markov decision process that allows non-fixed rewards?

hey! :)

:o More people whose names I remember from years ago

That's good to know :)

I've heard the math GRE is hard, while the general GRE is basically just an SAT over again...

I will admit though that, because I'm aiming for grad school in CS, I think I will have a lot easier time than you all who had to take the subject test in math.

I just signed up for the GRE. Now to study for it, haha

Sorry

Wait did I manage to ping the wrong adam?

:P

Accounts don't go inactive upon the user's death

(assuming SE is still around)

@Adam I'll keep getting rep after I die

The differentiating factor of bitcoin is that there is a finite supply. Reputation in inherently inflationary.

Reputation has no cap though.