Very nice work from Google on deep RL- based optimization for chip layout.
Simulated annealing and its heirs are finally dethroned after 40 years.
This uses graph NN and deConvNets, among other things.
I did not imagined back in the 90s that (de)ConvNets could be used for this.
This is the kind of problems where gradient-free optimization must be applied, because the objectives are not differentiable with respect to the relevant variables. [Continued...]
In this application, RL is used as a particular type of gradient-free optimization to produce a *sequence* of moves.
It uses deep models learn good heuristics as to what action to take in every situation.
This is exactly the type of setting in which RL shines.
UPDATE: the rumor on the street is that the comparison with existing tools from commercial EDA houses is not as favorable as the paper claims.
A story to follow......
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Hey @tegmark, most of us know that super-intelligent machines of the future will have to aligned with human values.
We just don't think it's as difficult as you make it to be.
And we don't think that getting it slightly wrong merely once will spell doom on humanity.
Worrying about superhuman AI alignment today is like worrying turbojet engine safety in 1920.
We do not have a working design for anything that could come close to becoming as smart as a dog, let alone a domestic robot that can clear the dinner table & fill up the dishwasher.
And the idea that intelligent systems will inevitably want to take over, dominate humans, or just destroy humanity through negligence is preposterous.
They would have to be specifically designed to do so.
Whereas we will obviously design them to not do so.
I have claimed that Auto-Regressive LLMs are exponentially diverging diffusion processes.
Here is the argument:
Let e be the probability that any generated token exits the tree of "correct" answers.
Then the probability that an answer of length n is correct is (1-e)^n 1/
Errors accumulate.
The proba of correctness decreases exponentially.
One can mitigate the problem by making e smaller (through training) but one simply cannot eliminate the problem entirely.
A solution would require to make LLMs non auto-regressive while preserving their fluency.
The full slide deck is here.
This was my introductory position statement to the philosophical debate
“Do large language models need sensory grounding for meaning and understanding?”
Which took place at NYU Friday evening.
I must repeat: 1- Auto-Regressive LLMs are useful, particularly as writing aids, particularly for code. 2- they hallucinate too often 3- they have a very primitive understanding of the physical world (hence those puzzles). 4- they have primitive planning abilities
2/
5- they have limited working memory 6- they execute a fixed number of computational steps per generated token 7- hence they are very far from Turing complete 8- Auto-regressive generation is a exponentially-divergent diffusion process, hence not controllable.
3/
My unwavering opinion on current (auto-regressive) LLMs 1. They are useful as writing aids. 2. They are "reactive" & don't plan nor reason. 3. They make stuff up or retrieve stuff approximately. 4. That can be mitigated but not fixed by human feedback. 5. Better systems will come
6. Current LLMs should be used as writing aids, not much more. 7. Marrying them with tools such as search engines is highly non trivial. 8. There *will* be better systems that are factual, non toxic, and controllable. They just won't be auto-regressive LLMs.
I have been consistent while: 9. defending Galactica as a scientific writing aid. 10. Warning folks that AR-LLMs make stuff up and should not be used to get factual advice. 11. Warning that only a small superficial portion of human knowledge can ever be captured by LLMs.
@babgi ChatGPT n'est pas particulièrement innovant.
Il utilise des techniques originellement développées à Google et Meta (FAIR), qui possèdent des systèmes similaires dans leurs labos.
Mais ces entreprises sont moins motivées à déployer des démonstrations publiques qu'OpenAI.
1/
@babgi Les meilleurs experts en France de ces méthodes sont à FAIR-Paris.
FAIR-Paris contribue *énormément* à l'écosystème français de la recherche en AI.
On peut regretter que certaines institutions publiques françaises voient FAIR comme un ennemi et non comme un partenaire.
@babgi Tout cela au nom d'une conception un peu dépassée de la souveraineté.
La souveraineté technologique et la maîtrise locale des nouvelles technologies sont des objectifs désirables et admirables.
3/
By telling scientists they must publish, you get: 1. higher-quality research, more reliable results, less self-delusion 2. better scientists whose reputation will flourish 3. easier external collaborations 4. better research evaluation 5. better internal impact 6. prestige
That's why at FAIR, we not only tell scientists to publish papers and open-source their code, we also use their publications as one component of their periodic evaluation.
To be clear, my original tweet was about scientists in *industry*.
Few companies promote publishing, some tolerate it, many forbid it.
The role of publishing in academia is well established and not in question.