Curious to know what mathematics you are comfortable with. If you are able to understand the papers you mentioned, you must belong to the 99 percentile.
I was never good at proof writing. I found group theory and algebra interesting, topology and analysis eluded me. It's just been a while since I did any serious math thinking
2) Hackers exfiltrate data from the target (this could be source code, database dumps, employee records, emails, or any combination of the above - basically anything that could be seen that has value to the company staying private.
3) Depending on the model used, the hackers either privately or publicly informs entity they have their data and unless a payment of X if made the data will get leaked or sold to the highest bidder.
I don't understand how anyone would ever pay. There is nothing guaranteeing you the hackers actually destroy their copy of the data on payment, so they could just come back and ask you for another payment every few months.
Or are we really supposed to believe these criminals would follow some sort of made up honor code?
You are completely right, they are criminals there is nothing stopping them from just dumping the data anyway (or launching another attack later down the road).
However the hackers also want to get paid, as soon as they go back on their word no one else will ever pay them.
But there is another "maybe" to consider (OP did ask for a brief explanation so I didn't go into all possibilities), did they encrypt the data? If they did and entity no longer has access to it they then have two options 1) restore the data from backup (if they had them and can restore service in a reasonable amount of time) / write off any data loss 2) pay up for the keys.
With no reputation, you’re presumably less likely to have victims pay up. You want to build reputation so you can get consistent profit from these extortions.
I don’t know if it’s really that interesting; reputation is just a fundamental currency required to facilitate trade when it can’t be guaranteed otherwise — there is in fact an honor amongst thieves.
These arm-chair game theory arguments tend to fall apart instantly as soon as you assume multiple rounds are played.
Aren't they all anonymous, though? So they could just change their name for the next operation. Maybe all these groups are already the same people behind the scenes.
You're missing the incentives. They /could/ change their name each operation, but then, as you note, the target would have reduced motivation to actually pay. By keeping their name, and keeping their word, customers are more likely to pay in the future, because there's a history of good faith transactions. And, of course, a group that is relying on their reputation like this must police their trademark and prevent other groups from abusing it.
Hey, what kind of work are you currently doing in AI? Just asking out of curiosity. I want to have a brief chat with you about these proof checkers and verification systems. I never got the gist of all these. Stuffs like lean theorem prover. I will be glad If you take out some time.
Ashu, i am going to guess that english is not your first language, apologies, if that is not the case, but " Tell me how can I run it locally." is not a great way to ask for info, not here at least. A 'please' would certainly help, but a less direct way might be even better.
Honestly the best way is starting with implementing a Q table for some small grid-world problem. You get a lot of knowledge from doing that. Then a bit more work on understanding various approaches, e.g policy learning, world models. Then, reading text books, blogs tutorials, etc.
But "getting" the idea of Q learning for a small state space is fundamental and surprisingly approachable.
I wrote this extensive tutorial for teaching deep reinforcement learning, with a focus on getting intuition from code. you will find RL theory is heavy on math despite needing math for very little other than abstractly representing some machine goal and intuition, of which code serves a native programmer already very well.
i spent years failing to learn machine learning and RL until i just started reading source code. books of integrals i never ended up needing.
dont be turned away by the joking nature of my tutorials. there is a real depth in there
+1 you beat me to the punch! I think its helpful to start with simple RL and ignore the "deep" part to get the basics. The first several lectures in this series do that well. It helped me build a simple "cat and mouse" RL simulation https://github.com/gtoubassi/SimpleReinforcementLearning and ultimately a reproduction of the DQN atari game playing agent: https://github.com/gtoubassi/dqn-atari.
I enjoyed the book Grokking Deep Reinforcement Learning from Manning. It's written in a very accessible style and explains the mathematical formulas you will see in other RL teaching material.
I'd suggest getting a good book or other teaching resource and solve a few Gymnasium[0] environments. Unlike supervised machine learning, you don't need someone else's data, you generate your own data.
I'm thinkin if your understanding were sound I would have to listen to it, but I rather read about these kinds of things, so whis is why I say unsound is best
on the other hand, putting things down into writing is difficult. So while you diminish this paper's accomplishments, you couldn't have done it yourself in spite of easily understanding it.
reply