Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Machine Unlearning in 2024

ai.stanford.edu

Post reply on HN
Print view View original

11–20 of 97 posts

Re: Machine Unlearning in 2024

#11

Post by negative_person »

williamtrask wrote: post #9
negative_person wrote: post #4

Why should we try to unlearn "bad" behaviours from AI? There is no AGI without violence, its part of being free thinking and self survival. But also by knowing that launching a first strike by a drunk president was a bad idea we averted a war because of a few people, AI needs to understand consequences. It seems futile to try and hide "bad" from AI.

Because we can get AI related technologies to do things living creatures can’t, like provably forget things. And when it benefits us, we should. Personal opinion, but I think AGI is a good heuristic to build against but in the end we’ll pivot away. Sort of like how birds were a good heuristic for human flight, but modern planes don’t flap their wings and greatly exceed bird capabilities in many ways. Attribution for…

Can you point to any behaviour in human beings you'd unlearn if theyd also forget the consequences?

We spend billions trying to predict human behaviour and yet we are surprised everyday, "AGI" will be no simpler. We just have to hope the dataset was aligned so the consequences are understood, and find a way to contain models that don't.

Re: Machine Unlearning in 2024

#12

Post by motohagiography »

seems like there is a basic problem where if you specify something to be unlearned, it could still be re-learned by inference and prompting. the solution may not be in filtering the proscribed facts or data itself, but in the weights and incentives that form a final layer of reasoning. Look at "safe" models now like google's last launch, where the results were often unsatisfying, as clearly we don't want truthful models yet, but we want ones that enable our ability to develop them further, which for now means not selecting out by antagonizing other social stakeholders.

maybe we can encode and weight some principle of the models having been created by something external, with some loosely defined examples they can refer to as a way to evaluate what they return, then ones that don't yield those results cease to be used, where the ones that find a way to align will get reused to train others. there will absolutely be bad ones, but in aggregate they should produce something more desirable, and if they really go off the rails, just send a meteor. the argument in how models can "unlearn" will be between those who favour incentives and those who favour rules- likely, incentives for ones I create, but rules for everyone elses'.

Re: Machine Unlearning in 2024

#13

Post by andy99 »

nullc wrote: post #10

I've wondered before if it was possible to unlearn facts, but retain the general "reasoning" capability that came from being trained on the facts, then dimensionality reduce the model.

If you think of knowledge as a (knowledge) graph, it seems there would be some nodes with low centrality that you could drop without much effect, and other key ones that would have a bigger impact if lost.

Re: Machine Unlearning in 2024

#14

Post by isodev »

dataflow wrote: post #3

> However, RTBF wasn’t really proposed with machine learning in mind. In 2014, policymakers wouldn’t have predicted that deep learning will be a giant hodgepodge of data & compute Eh? Weren't deep learning and big data already things in 2014? Pretty sure everyone understood ML models would have a tough time and they still wanted RTBF.

Of course, it’s not a regulation issue. The technology was introduced to users before it was ready. The very nature of training without opt-in consent or mechanism of being forgotten are all issues that should have been addressed before trying to make a keyboard with a special copilot button.

Re: Machine Unlearning in 2024

#15

Post by aeonik »

negative_person wrote: post #11
williamtrask wrote: post #9

Earlier quoted context omitted.

Because we can get AI related technologies to do things living creatures can’t, like provably forget things. And when it benefits us, we should. Personal opinion, but I think AGI is a good heuristic to build against but in the end we’ll pivot away. Sort of like how birds were a good heuristic for human flight, but modern planes don’t flap their wings and greatly exceed bird capabilities in many ways. Attribution for…

Can you point to any behaviour in human beings you'd unlearn if theyd also forget the consequences? We spend billions trying to predict human behaviour and yet we are surprised everyday, "AGI" will be no simpler. We just have to hope the dataset was aligned so the consequences are understood, and find a way to contain models that don't.

The feeling of extreme euphoria and its connection to highly addictive drugs like Heroin might be a use case. Though I'm not sure how well something like that would work in practice.

Re: Machine Unlearning in 2024

#16

Post by doubloon »

negative_person wrote: post #4

Why should we try to unlearn "bad" behaviours from AI? There is no AGI without violence, its part of being free thinking and self survival. But also by knowing that launching a first strike by a drunk president was a bad idea we averted a war because of a few people, AI needs to understand consequences. It seems futile to try and hide "bad" from AI.

AGI would not beGI unless it could change its mind after realizing its wrong about something

Re: Machine Unlearning in 2024

#17

Post by spennant »

dataflow wrote: post #3

> However, RTBF wasn’t really proposed with machine learning in mind. In 2014, policymakers wouldn’t have predicted that deep learning will be a giant hodgepodge of data & compute Eh? Weren't deep learning and big data already things in 2014? Pretty sure everyone understood ML models would have a tough time and they still wanted RTBF.

Agreed. The media and advertising industry was most definitely leveraging cookie-level data for building attribution and targeting models. As soon as the EU established that this data was “personal data”, as it could, theoretically, be tied back to individual citizens, there were questions about the models. Namely “Would they have to be rebuilt after every RTBF request?” Needless to say, no one in the industry really wanted to address the question, as the wrong answer would essentially shut down a very profitable practice.

Re: Machine Unlearning in 2024

#18

Post by huygens6363 »

nullc wrote: post #10

I've wondered before if it was possible to unlearn facts, but retain the general "reasoning" capability that came from being trained on the facts, then dimensionality reduce the model.

Yes, me too. If it could somehow remember the “structure” instead of the instantiation. More “relationships between types of token relationships” instead of “relationships between tokens”.

Re: Machine Unlearning in 2024

#19

Post by ben_w »

cwillu wrote: post #7

“to edit away undesired things like private data, stale knowledge, copyrighted materials, toxic/unsafe content, dangerous capabilities, and misinformation, without retraining models from scratch” To say nothing of unlearning those safeguards and/or “safeguards”.

It sounds like you're mistakenly grouping together three very different methods of changing an AI's behaviour.

You have some model, M™, which can do Stuff. Some of the Stuff is, by your personal standards Bad (I don't care what your standard is, roll with this).

You have three solutions:

1) Bolt on a post-processor which takes the output of M™, and if the output is detectably Bad, you censor it.

Failure mode: this is trivial to remove, just delete the post-processor.

Analogy: put secret documents into a folder called "secret do not read".

2) Retrain the weights within M™ to have a similar effect as 1.

Failure mode: this is still fairly easy to remove, but will require re-training to get there. Why? Because the weights containing this information are not completely zeroed-out by this process.

Analogy: how and why "un-deletion" is possible on file systems.

3) Find and eliminate the weights within M™ that lead to the Bad output.

Analogy: "secure deletion" involves overwriting files with random data before unlinking them, possibly several times if it's a spinning disk.

--

People are still doing research on 3 to make sure that it actually happens, what with it being of very high importance for a lot of different reasons including legal obligation.

Re: Machine Unlearning in 2024

#20

Post by 542458 »

doubloon wrote: post #16
negative_person wrote: post #4

Why should we try to unlearn "bad" behaviours from AI? There is no AGI without violence, its part of being free thinking and self survival. But also by knowing that launching a first strike by a drunk president was a bad idea we averted a war because of a few people, AI needs to understand consequences. It seems futile to try and hide "bad" from AI.

AGI would not beGI unless it could change its mind after realizing its wrong about something

I disagree. People with anterograde amnesia still possess general intelligence.
Post reply on HN