Computers are good at recording structured knowledge - who owns that, this merch goes where, things that were recorded on stone tablets since times immemorial. But human experts also have a different type of knowledge, empiric, accumulated through years of experience. Artisan crafters "know" how good some material is by sense of touch, smell, sight, without always being able to say WHY or HOW this wood is better than that wood for this table. This is why apprenticeship with a master was a key part of developing an artisan - you'd transfer SOME of this empirical knowledge from a master that took a lifetime to develop it.

The humanity took a huge step forward by moving a lot of empirical knowledge to structured knowledge through the use of models (mathematical?) and books. Instead of an apprentinceship with a mason, a builder now goes to school and learns how to structurally design a building based on construction codes. This allows huge scaling of knowledge, at the expense of missing subtle details which are not modelled (spherical cows in a void, right?).

With this article it just occured to me that Deep Learning may be the tool to record this type of empirical knowledge. And it's going to scale out - unlike the human version never did - because digital copying is cheap, and the knowledge doesn't die with the artisan that developed it - it only gets more accurate. The models we build about reality will get more and more accurate - the spherical cows will grow legs in the air, not in a void.

Humanity will start doing more and more things without understanding why, but because "the computer said so", and things work out when the computer says they will. Black boxes will explode in usage, and God forgive those that may be on the side where The computer says No.

This paragraph scares me. I'm not convinced that technology getting good enough to where it doesn't need to be understood to be used is a good thing in most cases. Especially when it's a black box.

Banking computer closed your account. Why? We can't tell you why.

Google or Apple rejected your app from their app store. Why, won't tell you why.

Me trying to get my prescription refilled at Walgreens. Walgreens computer shows they never got it from the doctor. Doctor sent it 5 times. Turned out the computer was marking it invalid each time. Pharmacist isn't allowed to tell neither me or my doctor why.

And it doesn't take a deep CNN to become a black box - any sufficient complex algorithm will do. I doubt anybody in the world understands all the details of how the latest CPU actually works, for example.

You would not believe a human is intelligent if he just tells you what to do without being able to explain why, would you?

Will it matter whether the AI is "really" intelligent when companies that implement the black box to spec thrive, while the companies that insist on having a human understand everything fall behind and fail?

Or think of the “uncanny valley” in 3D graphics -- most people can tell when a rendered image looks slightly off somehow, but most people can’t pinpoint the exact problems. “Something about the eyes just isn’t quite right”, etc.

But experience people can. I've worked at a 3D animation studio and we had this 'old guy' who started his career as a classically trained artist and sculptor. He could look at a character model that we all agreed felt a bit 'off' and point out that the problem is how you'd modeled the muscles connecting the shoulders to the arm. Hell sometimes he'd look at a model that we all felt was fine and suggest tweaking the bridge of the nose and that made it much better.

That sounds like a hellish dystopia.

On the one hand, opaque bureaucratic processes already exist today. It's not like the bank that refuses you a loan will tell you "oh, but we would totally have granted it if you were 20% richer and also not a woman, so come back after you've achieved that and we'll give you the loan"!

On the other hand, even opaque models can be studied and systematized. You can't run a simulation of a judge on 100 sample cases, then on the same 100 sample cases with names changed to sound like immigrants, and numerically measure the judge's bias. You can do that with a ML model, and compensate in various ways.

That's because they don't want to tell you, not because they don't know. The bank knows exactly why they refused to give you a loan and exactly which series of policy changes led to that outcome. They also know exactly how to change the policy to achieve a different outcome.

None of those things are true with AIs.

I don't see how machine learning, at least in its current non-AGI state, will be any different. It's just that your average end-user will have no idea how to "open up the machine", but that's also true for compiler technology today.

And you are onto something when you say that "empirical records" can be learned by ML models, and a huge amount of grunt work is required for that. Every tom, harry, and dick can overfit to MNIST today, it used to be hard someday. And note the amount of grunt work it took to build MNIST- thousands of human hours. Building Datasets does not pay you back instantly, but it has benefits for all. I know some people who are paying money out of their own pockets and creating niche Datasets and making them available under MIT Lisence.

Hope that more and more companies and people do that for really niche fields.

And I don't completely agree with not understanding things that DL models do and relying on them. We have started to understand much of it. ML interpretation is a field with mot much success but I am hopeful. We did not know jack about how very deep CNNs work, but that changed with the Zeiler-Fergus paper[0]. Later with things like Grad-CAM[1]. Now we are trying to understand latent representation fully. We (even I) can create GAN generated pictures with different haircolor, different kinds of glasses, etc. from scratch.

I read not more than two days ago that Microsoft and Peking Uni researchers found a way to identify "knowledge neurons" in unsupervized pretrained embeddings in NLP and they can edit facts with that [2].

So, I am optimistic about our "interpretation" future.

[0]: https://cs.nyu.edu/~fergus/papers/zeilerECCV2014.pdf

[1]: https://arxiv.org/pdf/1610.02391.pdf

[2]: https://arxiv.org/pdf/2104.08696.pdf

What happens when this kind of tool will calculate your credit score, social score, or the proceeds of your savings ?

Surely not - more accurately:

New AI tool intuits materials’ stress and strain based on photos

or

New AI tool guesses materials’ stress and strain based on photos

An experienced engineer could probably also guess roughly what shape the stress and strain gradients might take in a shape, but you wouldn't call such a guess 'calculation'.

The bar stretch is also completely wrong, there are no stress concentrations at the top and bottom; it should be uniform stress or concentrated strain in the center depending on what they're attempting to show.

First CGA monitors were bulky, energy hungry and blurry at the same time. 40 years of continuous development, and I am staring at a nice 4K screen that does not strain my eyes and is substantially more energy efficient.

1. materials not in the training set

2. stresses substantially different from those in the training set

So basically, it just memorizes a few patterns and can interpolate between the ones it's seen before. Big deal.

A good rule of thumb for detecting this variety of bullshit is this: given infinite time and resources, would an intelligent enough human be able to perform the task given the input? In this case, the answer is probably no.

This can be useful if you use it to automate review of regular inspection photos - you might be able to get the computer to recognize strain that would be hard to detect without a strain gauge.

Seing the pictures in the article, the model produces very coarse results from very simple inputs.

It's not even really a black box. It's learned a subset of Abaqus's functionality so it's scope is known.

Edit: Further elaboration from the article: product designers, for example, could test the viability of their ideas before passing the project along to an engineering team

This would seem to clearly indicate that the intended application is faster iteration in design, not a replacement for a rigorous engineering process that would include appropriate testing.

Estimating

It might even be good to analyze inspection photos for regular inspections to draw attention to situations where there might be more strain than anticipated and a human could be flagged to inspect more closely.

In my experience building and working with such models in similar domains, they tend to be right in all the easy and uninteresting cases where there are no problems and in all the cases where the problem is super obvious, and wrong in all the hard and subtle corner cases that you actually care about.

The other problem is that traditional engineers hate models they don't understand. If you can't explain, at least roughly, why it gave the answer it gave, then they won't trust it and won't use it.

This is pretty fresh tech, but industry is already using it. We are doing something approximately similar where I work. Instead of running compute intensive finite element/finite difference simulations of physically dependent systems, a neural network (typically something structured like a transformer) is trained to output the calculations up to 6 orders of magnitude faster in our applications.

This changes this allows modeling dependent science and engineering solutions to be iterated over in real time - you can see the results of your edits as you manipulate your models. And the results, at least in our applications, have some ≈98 MSE. It isn't surprising in hindsight - deep neural networks are universal function approximations, and finite modeling is as close to pure mathematics as you can get in an industry setting. It feels like a perfect use case for deep neural nets.