MathYouF

"Bachelor's degrees have become essential for well-paid jobs in the US."

The lies we continue to allow ourselves to tell as a society.

The real moment of truth will be if any models start to assist massively in research in the hard sciences.

Based on the quality of outputs I get when asking for help with somewhat complex AI research problems, I think it'll likely help accelerate the pace of other research as well, and discovery will be limited by people's speed of running the tests it suggests and feeding it back the results.

The main value of using SDF shapes for 3D modeling workflows is you don't have to worry about topology (the vertex, edge, face graph structure which has to be formed over the surface of all 3D models) which makes a lot of modifiers (like boolean combinations of intersecting objects) vastly less tedious (Womp calls this feature "goop").

Right now Blender work still involves a lot of tedium, mostly related to topology. A lot of upcoming 3D ML applications also work considerably better when using SDF instead of mesh representations. I wouldn't be surprised to see this form of 3D modeling take off to a significant degree because of those two factors.

The cost of materials and transportation and time using the expensive CNC machine will be the major costs of sculpture. Generating the same quality 3D models is at the very furthest 18 months away. And animating and rigging the models and giving them auto-generated RL policies will surely come very quickly next.

I don't think AI will be able to replace human creativity for discovering new paradigms as fast as it will replace human application of existing paradigms. And by doing the latter really well with AI, we're killing our ability to do the former. We'll end up with a sterile art trajectory.

This may actually end up making the few artists creative enough to create bold new art styles even more valuable, if they can basically not release their art and hide it behind a model.

Though I guess anyone with access to that model's output could then just generate a few samples and train on those, so maybe not.

With your expertise and the fact that they're so early, you should be able to build a competing business with your superior product right?

Then you can keep all the rewards (and find out all they had to do to be in the position to be able to hire you).

Sounds like a win-win, you should be talking to startup accelerators not startups hiring.

Numeracy is just as valuable as literacy.

I actually think this is likely not true.

If you were to take an 18 year old who can't read, and another one who can't do addition, which do you think is less employable?

There's a lot of jobs one can do without any math. Almost none one can do without any reading.

There's more to it than all this of course, but I think literacy is the clear winner compared to numeracy.

The tone of this betrays a possibly more argumentative than collaborative conversation style than that which I may want to engage with further (as seems common I've noticed amongst anti-connectionists), but I did find one point intersting for discussion.

Parameters are just the database of the system.

Would any equations parameters be considered just the database then? C in E=MC^2, 2 in a^2+b^2=c^2?

I suppose those numbers are basically a database, but the relationships (connections) they have to the other variables (inputs) represent a demonstrable truth about the universe.

To some degree every parameter in a nn is also representing some truth about the universe. How general and compact that representation is currently is not known (likely less than we'd like of both traits).

If greater parameterization leads to memorization rather than generalisation it's likely a failure in our current architectures and loss formulations rather than an inherent benefit of "fewer parameters" improving generalizaiton. Other animals do not generalize better than humans despite having fewer neurons (or their generalizaitons betray a misunderstanding of the number and depth of subcategories there are for things, like when a dog barks at everything that passes by the window).

The team I work on is building tooling with exactly that in mind, making this a part of an artists workflow rather than any sort of replacement.

I know as much about how to get the best image outputs from text inputs as the person who designed an airport knows the best place to eat in it. The emergent properties of the system are a result of the data put into it, so I can only discuss the system itself, not what it ended up doing with the data in that system.

The models are a product of their datasets, specifically the relationship of the images and prompts via CLIP. CLIP puts both images and text into coordinate space, imagine just a 2D graph. It tries to assure that for any real image and its caption, they will each be each others closest neighbor in that coordinate space.

So if you want a certain image, you have to ask "what caption would be most likely and most uniquely given to the image I'm imagining".

I'm sure this advice is way less helpful than what you find in prompt engineering discord channels and guides I've seen.

Pixel values are discrete (length x width x r256 x g256 x b256) and vertex values are continuous, so that is one major difference.

Secondly, there's vastly more labeled image data in the world than 3D data, so creating a CLMP (contrastive language and mesh pairing) model is harder.

It's very late but I may be able to give a much better answer on more of the nuances of 3D generation tomorrow.

Like anyone deeply in a field I know maybe several thousand people who could probably give a better answer, but I figure I'll give an effort to provide one since I don't see any good ones posted yet.

The moment everyone knew this was going to be big was in 2019 when StyleGAN came out. They used a lot of tricks like aligning face features (like eyes) and had all their pictures of a single domain (the most famous being faces) but none the less, that was the moment everyone in the AI field knew this was going to be big, and so three years ago a lot of big people shifted to this line of research.

The four main innovations since then have been:

1. Transformers

Generalized computation kernels which allow for images to consider non-localised relationships between pixels of an image. Released in 2017, and originally used for language.

2. Pixel Patch Encodings

Different resolution semantic and geometric image information encodings which allow for better representations of relationships between image areas than pixels are able to achieve given the same compute. Allows using Transformers on high resolution images.

3. CLIP

Contrastive Language and Image Pairing. Before, the only way we knew to classify an image was as a "face" or "cat" or "ramen". When the genius idea of labeling images as semantically meaningful vectors rather than one hot encoded classes was revealed, it changed everything in computer vision very quickly, and problems that used to be hard became trivial. Released in 2021

4. Diffusion Models

GANs penalise you for making an image which does not seem to be part of an existing dataset. This encourages one to make the worst quality image that looks like a member of that dataset. Diffusion learns to denoise an image, removing noise is perceptually similar to increasing resolution, people like images that look that way. There may be more people with better intuition about diffusion models may be able to add on why they're superior. I've read all the papers leading up to the latest unCLIP (Dalle2) but it's complicated. Released in 2020, with major improvements to the training process continuously being made since then.

Hope this was helpful. All of the above were only implemented for images in any real way in the last three years. Putting them all together is something many people only just this year did, resulting in DallE, Stable Diffusion, and Imagen.

I'm working on doing this for 3D and later for use cases in AR. 3D generation still hasn't been cracked the same way image has but the above will likely contribute to the solution to that as well. Anyone who's intersted in working on that feel free to message me.

After an excellent all day workshop at CVPR this year explaining diffusion in more detail, it seemed pretty clear that any noise function could be used. I'm not sure if it should have been obvious, but I felt this paper coming from a mile away after seeing that.

I wonder to what degree certain parts of diffusion dictate using certain noise functions, and how much this paper truly challenges how we understand them. Cool to see it was researched.

Next idea: it seems like a lot of steps could be skipped by using things like momentum during the inference time. I'm sure OpenAI has already implemented several clever tricks like that in production for DallE.

I'm working on (various, non-diffusion) methods for 2D drawing to 3D output right now.

What kind of things are easier to debug with lots of RAM and how would you do it?

I'd love a chance to talk to this guy (maybe at a rock-star after party at NeurIPS this year) because my view is:

1. "Either everything is magic or nothing is", and magical statistical tools are currently some of the coolest magic we have harnessed.

2. Making realistic pictures of "Corgi's with sushi" is cool.

3. Papers describing architectures which can make ever more realistic or interesting pictures of corgis and sushi are deserving of academic recognition, even if they can't precisely describe how, just as renderers which can do so would be.

I've found a lot of papers which primarily study the theory of ML coming out of the UK, A yet very few of them end up being of much value to the advancement of the field in terms of applications.

Conversely, the people focused on making real applications ("tools") seem to also be innovating on the science.

Tesla and Edison contributed massively to the progress of the study and application of electricity and both spent nearly no time in academia and instead focused on practical applications in industry. Edison's methods of investigation were said by Tesla (somewhat admiringly) to be entirely empirical. I wonder if the UK's strict academic approach to ML may be holding them back from making bigger contributions to the field. I'm glad some are trying to be strictly scientific about it, but I'm also glad it's not everyone, because it doesn't seem to be what's delivering the "magic" we all want.

My philosophy is if you think they're wrong and you don't need those people, then they have poor judgement, and you're better off not working with them in that case, and in fact could beat them by pursuing similar goals while not using that strategy.

And if they're right, well then what's there to be upset about?

I think the popularity of my post suggests it better represents the general sentiment of people who viewed the page than not, but I do accept the tone reflects my negative feelings towards the perceived credentialsim:substance ratio more than it does an impartial review of the content. Genuinely, feel free to remove the entirely of my thread if you think it is unproductive, I'm unable to edit it.

I am indeed bitter (def: anger and disappointment at being treated unfairly; resentment) about people using their credentials rather than real merits of their work to attempt to get a leg up when advertising their work. I covered my dissatisfaction about that in my post.

Distill.pub for example, despite being written by people who worked at OpenAI and (also funded by) Google Brain, makes nearly no mention of either. They've managed to successfully accomplish their goal (creating resources to help people improve their intuition about machine learning) without resorting to those tactics.

I invited more positive and productive discussion on the advertised topic:

"I'd enjoy hearing some concrete discussion of what unique vision you have for improving the intuition of people trying to do data science, and how this helps humanity, and your journey beyond A/B testing marketing techniques.

So far i've not received it from any replies (including yours) and haven't seen it happen in any other comment threads so far (including ones which could be made by yourself).

The lack of thoughtful discussion makes me believe this post is meant to be just a marketing/sales pitch (something I covered as a potential issue in my comment) rather than a post for having deep, thoughtful, expansive discussions on this topic.

Like I said, I think you are perhaps the one who isn't in sync with what kind of conversations are valued.

You can find the number of times each is mentioned (3) by using a page search (ctrl+f).

From my admittedly limited knowledge of how self driving policy models are implemented, I believe the main black box AI is determining the part

`if personInFrontOfCar:`

(using object detection, masking, 3D pose estimation, et/or cetera)

And actually defined policies on how to handle certain situations given that inference are possible, in the case specifically of self driving vehicles.

My assumptions come from when I'd asked someone at NeurIPS in 2019 about what reinforcement learning methods they use as they said "I don't think any self driving companies are using reinforcement learning, it is too risky". I don't mean to imply this is hard to find information either, reading a few papers would be all it takes to clear up the degree to which most self driving policies are "controllable" in the way you describe, or at least to what degree.

My main point is that I think it is the inferring of the environment (current state) rather than the chosen policy at each time step which is more of an error prone black box.

I was watching this video recently from Weights and Biases on Linear Algebra intuition for ML.

A mention was made in it that the real reason matrices rather than code or formal logic are used in ML is because of their ability to be changed in small ways with a relevant directional change in the operability of that matrix. Changing a single character of a code block doesn't have the have smooth changes.

So your paper seems like a cool move in the right direction, and supportive of the work people are doing to try to make ML more interpetable by turning it into formal logic.

This makes me wonder what laws and tedium of administration merchants of the 1320's would complain about. I'm sure they still had a few depending on the region, and maybe more severe possible outcomes (highway robbery, unlawful arrest because of the influence of a rival merchant, arbitrary taxation and tariffs, etc.).

We seem obviously to be the smartest living things in the universe we know of.

We're also approximately the dumbest possible things that could construct the society and technology we have (if we weren't, we would've done it sooner).

I agree that some humility of our own intellect as primates would do us a lot of good in the coming age of AI.

I thought this would be about best practices for deployment of models like this.

I once asked in the LAION discord what that was and they said basically flask/fastapi server + load the model + gunicorn + ec2 instances.

I'm curious if any HN readers have any further insight or thoughts on ML model deployment for projects like DALL-E Mini.

Well, we could at least start with having people not use profanity, or tell each other to kill themselves.

I admit though I can't think of a good way to legislate the nuances of someone expressing an opinion that one ought not to be able to compete in sports of a gender they believe they belong to, without making them feel they are being dehumanized.

----------

On the subject of what led to this law, my girlfriend and I were avid watchers of the show Terrace House as the events that led to this person's death unfolded.

The girl had been quite rude to a foreigner on the show.

This was considered a huge disgrace to the nation of Japan, to have one of their own act so meanly to another, especially on as international of a platform as Netflix, especially to someone from what is considered a partner country, like the U. S..

This led to social ostracization where many Japanese were (civilly) telling her that she was no longer respected in general society, and questioning the extent of her ability to behave with civility.

I believe this (civilly said) cultural censure for her imopliteness likely contributed more to her feelings of worthlessness than any truly hateful and ignorant things said to her. Or maybe because she grew up in Japan she was just not emotionally prepared for hearing some of the hateful things people can say on the Internet (this is the public narrative).

Either way, it's a tricky situation, and many in Japan felt she was actually the one who began the degregation of civility with her behavior on the show, but none the less the resulting way people spoke to her was also not considered acceptable.

How can one balance expressing the idea that she brought shame to the nation while also not making her feel dehumainzed? I don't know. But I'm quite sure in the minds of the Japanese who wrote this law, the main thing they want to avoid is people telling others to kill themselves directly or cursing at them, rather than avoiding expressing judgemental ideas on one's behavior, which is common enough in Japanese society, to be subtly hinting at one disapproving of the behavior of others, while remaining civil.

I've lived in both countries, living with citizens of both countries, seeing popular media in both countries, long enough to know there's a massive difference in the degree of hateful and impolite behavior the average person is willing to engage in for any given topic of conversation, and that citizens of both countries seem to enjoy the same degree of freedom of expression of political views which diverge from the interests of those in power. The difference is how they choose to say things, and that has downstream effects of their ability to continue to view those they disagree with as humans worthy of at least some degree of compassion and respect.

I think the fact that Americans have trouble understanding how one could express the full range of political ideology while still remaining civil and not bullying each other is part of the problem.

One can express any proposal for legislative or judicial course of action without demeaning the inherent humanity of another and resorting to profanity, lewdness,and hatred. Learning to express opinions in this way can help contribute to us recognizing ways to solve our mutual problems as a society,ans avoid people feeling so dehumanised that they harm themselves or others.

I do appreciate your concern though for preserving our ability to disagree and express any political belief without censure.

Maybe people shouldn't say evil and hateful things about anyone, no matter how bad they are, because it makes the entire culture of public discourse more toxic. Americans seem to have very little reverence for the quality and civility of public discourse in general anymore, and often say "the other side is worse" or "I'm intolerant of intolerance" as their justification for personally contributing to the normalization of uncivil ways of speaking.

Japan doesn't seem to have these problems with people culturally defecting from civility even when they disagree, and I believe their willingness to not tolerate it to any degree is what keeps the rest of society defecting to low quality behavior patterns.

There are many morally self righteous individuals who think they are doing the right thing and don't realise their counterparts they spew (self righteously) hatred towards are the exact same as them.

I think both groups should receive censure for valuing their personal crusades over the quality of stability and civility of society the rest of us value; who have more nuanced and humble opinions of our own self righteousness. Really glad there's another country that agrees with me!