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

Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

nobelprize.org

Post reply on HN
Print view View original

781–790 of 876 posts

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#781

Post by Ma8ee »

eigenket wrote: post #40
Ma8ee wrote: post #15

I think this is the Royal Academy of Sciences way to admit that Physics as a research subject has ground to a halt. String theory suffocated theoretical high energy physics for nearly half a century with nothing to show for it, and a lot of other areas of fundamental physics are kind of done.

I think this is (very) inaccurate. It feels more like them trying to jump on a "hot topic" bandwagon (machine learning/AI hype is huge). Physics as a discipline hasn't really stalled at all. Fundamental physics arguably has, because no one really has any idea how to get close to making experimental tests that would distinguish the competing ideas. But even in fundamental physics there are cool developments like the s…

I just briefly looked into what Jonathan Oppenheim is working on, and I’d say he’s part of the problem. More speculative work that might or might not be testable in a distant future.

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#782

Post by aulin »

radioactivist wrote: post #769
nabla9 wrote: post #79

Most commenters here don't know that Boltzmann machines and associative memories existed in condensed matter physics long before they were used in cognitive science or AI. The Sherrington–Kirkpatrick model of spin glass is a Hopfield network with random initialization. Boltzmann machine is Sherrington–Kirkpatrick model with external field. This is price in physics given to novel use of stochastic spin-glass modelling…

I'm a condensed matter/statistical physicist and am very aware of the connections to statistical physics I still think that the committee has completely lost it with this choice. There is a sharp line for me between things that are inspired by physics and those that are physics (and I really don't buy that physics is anything physicists do) -- and this clearly falls on the "inspired by" side.

I know plenty of physicists that would be very pissed off by all this drama.

There is so much more than fundamental physics and there is much more in physics than breakthrough discoveries. Medical physics just to name a fun field everybody always forgets about has been studying and using neural networks since about forty years. Applied physics, biophysics, atmospheric physics. Even particle physics is mostly data science these days.

This idea that physics should only be about fundamental theories and discoveries is really detrimental to the field and leads to the false idea of stagnation that permeates this whole thread.

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#783

Post by sashank_1509 »

A lot of commentators here saying that while fundamental physics has possibly stalled, a lot of applied physics is still bustling with activity. Even buying that premise, is there anything in the applied physics fields that is comparable in impact to the neural network?

It feels like the Nobel committee’s decision is an indictment on the lack of impact of modern physics. They had to stretch definitions and go into AI, to get something that they found impactful enough for receiving a Nobel prize. This is impact outside of physics, impact on a broader societal sense that physics in the past had in spades and AI has now.

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#784

Post by fooker »

bunderbunder wrote: post #708
fooker wrote: post #703

Earlier quoted context omitted.

> because physics is currently stalled. Even if it's not completely true, maybe some introspection is required? I understand developing new theories is important and rewarding, but most physics for the last three decades seems to fall within two buckets. (1) Smash particles and analyze the data. (2) Mathematical models that are not falsifiable. We can be pretty sure that the next 'new physics' discovery that gives us…

Those two buckets only contain the work in physics that have a sustained presence in popular media. But take gravitational wave astronomy as a counterexample. It doesn't make it into the news much, but I'm pretty sure the entire field is less than ten years old.

Yes, good example. They already got a Nobel prize, one of the quickest in history.

This is the kind of physics we might need more of.

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#785

Post by fooker »

codethief wrote: post #710
fooker wrote: post #703

Earlier quoted context omitted.

> because physics is currently stalled. Even if it's not completely true, maybe some introspection is required? I understand developing new theories is important and rewarding, but most physics for the last three decades seems to fall within two buckets. (1) Smash particles and analyze the data. (2) Mathematical models that are not falsifiable. We can be pretty sure that the next 'new physics' discovery that gives us…

> most physics That's an interesting definition of "most physics". I mean, I find high-energy physics as fascinating as the next guy but there are other fields, too, you know, like astrophysics & cosmology, condensed-matter physics, (quantum) optics, environmental physics, biophysics, medical physics, …

I wonder if your list approaches AI at some point!

The Nobel committee seems to think so

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#786

Post by fooker »

ImaCake wrote: post #738
codethief wrote: post #710

Earlier quoted context omitted.

> most physics That's an interesting definition of "most physics". I mean, I find high-energy physics as fascinating as the next guy but there are other fields, too, you know, like astrophysics & cosmology, condensed-matter physics, (quantum) optics, environmental physics, biophysics, medical physics, …

The corners of physics I have some contact with (climate/weather modelling and astrophysics) seem pretty dynamic to me. Each generation of CMIP models seems to be significantly better than the previous.

What's the core technology making these models significantly better?

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#787

Post by eigenket »

Ma8ee wrote: post #781
eigenket wrote: post #40

Earlier quoted context omitted.

I think this is (very) inaccurate. It feels more like them trying to jump on a "hot topic" bandwagon (machine learning/AI hype is huge). Physics as a discipline hasn't really stalled at all. Fundamental physics arguably has, because no one really has any idea how to get close to making experimental tests that would distinguish the competing ideas. But even in fundamental physics there are cool developments like the s…

I just briefly looked into what Jonathan Oppenheim is working on, and I’d say he’s part of the problem. More speculative work that might or might not be testable in a distant future.

I would say that's an overly simplistic view. The only way we ever obtain testable things is by working on things which are not currently testable.

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#788

Post by dguest »

moelf wrote: post #230
Separo wrote: post #210

Well maybe in time ML will help break through the high energy physics roadblocks.

it already had, bottom-quark tagging has improved O(10)x in efficiency in the last decade without any new "physics" understanding, just from training with more low-level data and better ML arch (now using Transformers) but we haven't found new physics with or without ML, making this prize a little weird.

We used ML to discover the Higgs.

Whether or not the original Higgs discovery decay channels used ML, confirming that it was in fact the Higgs required measuring the decay to b-quarks, which has used ML since the LHC started taking data.

Over the lifetime of the LHC, he backgrounds got around 10x smaller for the same "efficiency" (fraction of true b-quarks tagged) if you want to be pedantic about the definitions. We've used NNs in b-tagging for decades now, so it was always possible to dial in a threshold for tagging that was e.g. 70% efficient.

Transformers gave us a factor of a few smaller backgrounds in the last few years though [1].

[1]: https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PLOTS/FTAG-20...

Re: Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]

#789

Post by varjag »

DiogenesKynikos wrote: post #476
varjag wrote: post #323

Earlier quoted context omitted.

Human death toll in Libya and Syria differ by almost 60x. Half a million Syrians could have lived, the refugee crisis and the rise of far right in the West could be avoided had Assad been droned in 2013. Putin would also not have dared the 2014 annexation either.

The US had hundreds of thousands of troops on the ground in Iraq for over a decade. More than half a million Iraqis died. There was intense violence between different religious groups and political factions. But you come here and say that everything would have magically gotten better with more US involvement in Syria. A direct American intervention in Syria probably would have made things even worse. Droning Assad, a…

Notice how I specifically talked about Syria and Libya. I (along with a lot of other people) opposed Iraq war as well and it took you to pull it here for lack of consistent argument.

Don't see the point arguing with you further. Some day both Putin and Assad are going to be dead and I hope they suffer in their last minutes. I will be cheering while you will be mourning your tyrants.

Post reply on HN