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

Alphafold

github.com

Post reply on HN
Print view View original

91–100 of 170 posts

Re: Alphafold

#91

Post by dekhn »

gnufx wrote: post #83
dekhn wrote: post #35

Earlier quoted context omitted.

many of these resources are available, it's mostly that academic scientists don't have the time, money, or expertise to manage large datasets. However, the community has maintained high quality MSA database for decades and that's exactly the work that DM drafted off.

> academic scientists don't have the time, money, or expertise to manage large datasets I may be cynical about general expertise, as a support person, but large datasets have long been stock in trade of areas I'm more or less familiar with, whether "large" is TBs or PBs like CERN experiments. (When I were a lad, it was what you could push past the tape interface in a few days -- data big in cubic feet...)

Tape is worthless except for archival purposes (and it's not particularly good). it should not be the constraint on the dataset (IE, any important dataset should already be in live serving with replication).

Very few players wrangle petabytes effectively. Many players have petabytes, but they're just piles of disorganized data that couldn't be used for training ML. Moving petabytes is still a huge pain and few folks have proficiency in giving ML algorithms high performance access to the data.

Re: Alphafold

#92

Post by dekhn »

gnufx wrote: post #86
dekhn wrote: post #21

Earlier quoted context omitted.

Yes, all science should be communicated in the form of an academic paper wiht a supporting git repo and quickly downloadable dataset and a fast path to reproducing the work. That would be a huge change from the establishment. It's quite unclear what value this will have to pharma; personally I doubt this has any direct applications (and I'm one of the few people in the world that can say that with deep authority).

Surely not all science. Just as well Dirac wasn't required to communicate that way the equation that fundamentally underlies the phenomenon discussed, and you couldn't put the unique facility my thesis work pioneered into git! I do highly approve of publishing software and data where possible, of course, since before Free Software needed to be coined, and it's much easier now.

If you're just publishing equations, you should have an associated notebook which executes the equations.

I don't know what you mean you can't put your thesis work into git. Is it a physical thing? Too big for git?

Re: Alphafold

#93

Post by gnufx »

dekhn wrote: post #56
miltondts wrote: post #53

Earlier quoted context omitted.

> ... but I don't play that role any more. I was thinking of going into that field. Can you expand a bit on why you left?

Because programming computers is far more lucrative, and I'm better at it. However, if I had an unlimited budget I would return to biology. I spent 15 years trying to be a professor and failed miserably. I was bad at it and didn't like what professors have to do. I then moved to industry to be a random engineer and thrived doing things entirely unrelated to drug discovery. Eventually, I convinced my company to invest…

> much of my time is just telling scientists what I would do if I was in their place anyway.

That sounds familiar. I guess they mostly don't listen, whatever your record -- especially if it was in a different field they could learn from -- but I hope it's not always like that.

Re: Alphafold

#94

Post by dekhn »

HPsquared wrote: post #88
Azrael3000 wrote: post #63

Earlier quoted context omitted.

Incandescent light bulbs are generally very inefficient in producing light, compared to LED for example. They produce a lot of heat and not much light for which they are made. So in this context I suppose that gp implies that these threads don't provide much meaningful discussion but rather lots of hand waving.

Light is also often used in metaphors relating to knowledge, wisdom etc.

"Fiat Lux" not "Fiat Calor"

Re: Alphafold

#95

Post by dekhn »

astro-codes wrote: post #87
dekhn wrote: post #21

Earlier quoted context omitted.

Yes, all science should be communicated in the form of an academic paper wiht a supporting git repo and quickly downloadable dataset and a fast path to reproducing the work. That would be a huge change from the establishment. It's quite unclear what value this will have to pharma; personally I doubt this has any direct applications (and I'm one of the few people in the world that can say that with deep authority).

Why wouldn't this have much value to pharma? Is it because its application is actually really limited in scope?

there are research groups this would be useful for but structures are not on the critical path to drug discovery or approval.

Re: Alphafold

#96

Post by sbierwagen »

nextos wrote: post #16

Alphafold 2 is very very cool, but we need a little dose of reality. It's still a bit away from really solving protein folding as it was marketed. For example, multi-complex proteins are not well predicted yet and these are really important in many biological processes and drug design: https://occamstypewriter.org/scurry/2020/12/02/no-deepmind-h... A disturbing thing is that the architecture is much less novel than I…

>A disturbing thing is that the architecture is much less novel than I originally thought it would be, so this shows perhaps one of the major difficulties was having the resources to try different things on a massive set of multiple alignments.

A similar concern has sparked some worries about "AI overhang" https://www.lesswrong.com/posts/75dnjiD8kv2khe9eQ/measuring-...

Most of the compute in ML research seems to be going into architecture search. Once the architecture is found, training and net finetuning/transfer learning is comparatively cheap, and then inference is cheaper still. This implies we could see 10-100x gains in AI algorithms using today's hardware, or sudden surprising appearance of AI dominance in an unexpected field. (Object grasping in unstructured environments? Art synthesis?) A task could go from totally impossible to trivial in a year. In retrospect, the EfficientNet scaling graph should have alarmed more people than it did: https://learnopencv.com/wp-content/uploads/2019/06/Efficient...

Waymo has been puttering along for years, not announcing much of interest. This may have caused some complacency about self-driving cars, which is a mistake. Algorithms only get better, while humans stay the same. Once Waymo can replace some human drivers some of the time, things will start changing very quickly.

Re: Alphafold

#97

Post by sbierwagen »

nikhilsimha wrote: post #41
rllearneratwork wrote: post #27

Earlier quoted context omitted.

why would it put you out of job? Wouldn't it just become one of the tools you use?

The implicit assumption you are making is that the demand increases in lock step with productivity gains. 100x faster drug discovery, 100x more drugs need to be discovered => same number of people employed. These correlations do hold for technical fields, but logically there should be a point beyond which productivity gains outpace, demand growth / demand could even stop growing. One should either retool to solve a n…

>but logically there should be a point beyond which productivity gains outpace

The limiting factor on drug approval is clinical trials. Once every living person is enrolled in a clinical trial, we will have hit the maximum rate at which humanity can produce new drugs.

That might be more than 10x the current rate, but probably less than 1000x.

Re: Alphafold

#98

Post by mikewarot »

duckerude wrote: post #66

> The AlphaFold parameters are made available for non-commercial use only, under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license. You can find details at: https://creativecommons.org/licenses/by-nc/4.0/legalcode Does CC BY-NC actually do this? As far as I can tell it only really talks about sharing/reproducing, not using. Or is the only thing prohibiting other comm…

If you took their parameters, then trained it for while on a different set of data, it would vary from the original. I wonder how much compute would be required to make the offset far enough to hold up from scrutiny, and in court.

Alternatively, you could manually change the network model, add a few hidden layers, etc... modifying the parameters in step, and result in a new model and new parameters. Some training to vary the parameters, and it's now a new work.

Re: Alphafold

#99

Post by dekhn »

sbierwagen wrote: post #97
nikhilsimha wrote: post #41

Earlier quoted context omitted.

The implicit assumption you are making is that the demand increases in lock step with productivity gains. 100x faster drug discovery, 100x more drugs need to be discovered => same number of people employed. These correlations do hold for technical fields, but logically there should be a point beyond which productivity gains outpace, demand growth / demand could even stop growing. One should either retool to solve a n…

>but logically there should be a point beyond which productivity gains outpace The limiting factor on drug approval is clinical trials. Once every living person is enrolled in a clinical trial, we will have hit the maximum rate at which humanity can produce new drugs. That might be more than 10x the current rate, but probably less than 1000x.

In principle you could put people into multiple trials and gain somewhat additional throughput. Google implemented putting users into multiple different experiments (paper by Tang et al) and that made a huge difference.
Post reply on HN