BrokrnAlgorithm

I'm a musician, but am also pretty amused by this anti ai wave.

There was recently a post referencing aphex twin and old school idm and electronic music stuff and i can't help bein reminded how every new tech kit got always demonized until some group of artists came along and made it there own. Even if its just creative prompting, or perhaps custom trained models, someday someone will come along and make a genuine artistic viable piece of work using ai.

I'd pay for some app which allows be to dump all my ableton files into, train some transformer on it, just to synthesize new stuff out of my unfinished body of work. It will happen and all lines will get blurred again, as usual.

I find python's async to be lacking in fine grained control. It may be fine for 95% of simple use cases, but lacks advanced features such as sequential constraining, task queue memory management, task pre-emption etc. The async keword also tends to bubble up through codebases in aweful ways, making it almost impossible to create reasonably decoupled code.

I've been out of the loop for stats for a while, but is there a viable approach for estimating ex ante the number of clusters when creating a GMM? I can think if constructing ex post metrics, i.e using a grid and goodness of fit measurements, but these feel more like brute forcing it

Is there any didactic implementation of the Disruptor / multicast ring available somewhere? I've been curious in working through some practical example to understand the algorithm better.

Is there also a decent c++ implementation of the disruptor out there?

- Location: Switzerland

- Remote: Yes

- Wiling to relocate: Yes

- Technologies: C++, Python, Torch

- CV: https://www.linkedin.com/in/estlan-7217a8aa/

- Email: estebanlanter86 [at] gmail [dot] com

About me: Professional Machine Learning / Quant development background (Python, Torch). Experience with low latency software Engineering in C++. Expertise in time series related engineering topics in both ML and software engineering.

Happy to relocate, no need to stay in my field (curious to see new stuff too!)

This is a known problem in generative workflows for AI vids, but solvable. Midjourney recently introduced a feature that does this for stills, and controlnets available for the comfyui ecosystem also can partially solve this, albeit with some hassle. I'm pretty sure if not OpenAI themselves others will follow with their foundation models.

Coming from finance, I always wonder how and if these large pre-trained models are usable on any financial time series. I see the appeal of pre-trained models in areas where there is clearly a stationary pattern, even if its very hidden (i.e industrial or biological metrics). But given the inherently high signal/noise ratio and how extremely non-stationary or chaotic the financial data processes tend to be, i struggle to see the use of pre-trained foundation models.

I was referring to being stuck with having to create simple / low tri polygonal meshes as opposed to using complex poly meshes such as photogrammetry would provide. The paper specifically addresses clean low poly meshes as opposed to what they call complex iso surfaces created by photogrammetry and other methods

"Lower triangle budget media" is what I wonder if its still a valid problem. Modern game engines coupled with modern hardware can already render insane number of triangles. It feels like the problem is rather in engines not handling LOD correctly (see city skylines 2), although stuff like UE5 nanite seems to have taken the right path here.

I suppose though there is a case for AI models for example doing what nanite does entirely algorithmically and research like this paper may come in handy there.

I'm not a 3D artist, but why are we still, for lack of a better word, "stuck" with having / wanting to use simple meshes? I appreciate the simplicity, but isn't this an unnecessary limitation of mesh generation? It feels like an approach that imitates the constraints of having both limited hardware and artist resources. Shouldn't AI models help us break these boundaries?

Deleted comment, wrong thread

What matters more to me usually in a time series context is how forecasting tools deal with non-stationarity of the underlying data process. I'm not an expert on LLMs, but I assume they wont' be the ideal tool to use in these contexts because even finetuning will be rather expensive...?

I believe Carma Reincarnation used a completely new system, I believe they talked about that in the dev blog. While the system may have been similarly powerful in theory, I distinctly remember it also being less fun than the Carma 2's system, somehow less reactive. Carma 2's approach essentially allowed for the car to be crushed into any polygonal form, best illustrated when you drove the car into the underside edge of an angled bridge or similar. It was rad and so much fun, considering and that was 1998 which still blows my mind.

Agree, Carmageddon 2's physics still feel modern to this day. I can't think of any other game except Beam NG perhaps that achieves the same level. But the impacts in Carma 2 seem to suprass even that one.

I used to do R intensly. What I found, after moving to more "SE"-centric languages such as Python or C++, was that R becomes quite frustrating when you need to build something maintainable. Which is basically a definition that should apply to every package. As soon as S3, R6 and whatnot comes into play at multiple levels, I'm better off moving to C in order to minimize interacting with R's class system and having that done only at the topmost layer.

I'll most likely open source it once it's refactored properly and I have a chance to show it off :)

Right now it's not looking as fancy as it could, i've been focusing more on optimizing the render engine for speed and not much yet on fancy shaders and such.

But I want to move away from OpenGL towards vulkan at which point I'll most likely publish the code.

I'm working on a real-time gpu accelerated 3d visualizer for high frequency market Exchange data.

There's no practical use case for it except that it looks cool. I don't plan on making it useful either, instead I'm gonna implement all kinds of awesome looking but useless eye candy stuff. But it's a good excuse for a quant to get into graphics programming.

I actually agree with the vision, like its clear that media convergence will all kinda lead us into a common tech space, be it AR / VR / 3D whatever.

But I don't think it needs to be called metaverse. We should have another name for it. Not to give "Meta" as a company the same honor as google received for "googling".

It seems to be 0.9% MoM. While still high, a lot of the YoY numbers dominating these headline are driven by statistical base effect from covid anomalies 12 months ago.

Given we're still in recovery mode from Covid, I'd expect more anomalies to show up in econ data over the coming months.

I don't think there's anything persistently big on the horizont w.r.t to inflation.

Jup, that's how it went with me. Read first hyperion, then had to dive into keats because he was such a central figure - in a sci fi book which måde me just more so curious.

I've been wondering about the exact same thing. Basically, a sufficiently calibrated and targeted GPT3 bot swarm could be employed to render at least some parts of these communities into useless echo chambers.

I think that while GPT3 posts are usually identifiable after reading some sentences, I often also find that it's harder to realize when consuming precisely this kind of social media. I often just consume it absent mindedly as opposed to reading a real article.

Good point as well - sometimes its not about dimensionality reduction but more about persistent representation, having this geared towards highly non-stationary environments is nice thing to have.

Still, there are a lot of domains where transfer learning is no the most applicable setting - I'm thinking of highly noisy and non-stationary setting such as finance. In some of these domains, especially time series, lack of data is often not the issue, e.g. high frequency datasets.

Having models constantly re-train as the default setting is essentially what a rolling regression would do - having a rolling regression that doesn't catastrophically forget would be quite valuable.

What about findings w. R. T. Online learning? I find continuous learning quality of algorithms to be a topic which often seems to be more of a side-concern, although it carries a lot of relevance in applied settings.