Edge Rewrite
// HTMLRewriter · presentation

This page was redesigned at the edge.

Cloudflare fetched the original article and streamed it through HTMLRewriter to apply an entirely new visual system without rebuilding the source page.

// request.cf · coarse context

A page that knows where it met you.

Only coarse request metadata is shown. This demo does not display or persist visitor IP addresses.

Country
US
Cloudflare location
CMH
Connection
HTTP/2
Language
Not provided

Ray ID: a21e3135783e284b

Jump to content

Template talk:Differentiable computing

Page contents not supported in other languages.
Add topic
From Wikipedia, the free encyclopedia
Latest comment: 3 years ago by Stellaathena in topic "Implementation" Grouping is Weird

Why Programming languages group?

[edit]

I don't want to remove the group because I didn't contribute anything to the template and removing stuff added by other users seems rude to me. But now using JS or GO for programming gradient descent based algorithm is not that uncommon, does it implies that wikipedia should also add JS/GO to the group. To put it simply, I don't think having a group for popular programming languages used for Differentiable computing helps anyone but rather might push the idea that Python is the language for AI ignoring many other important stuff, for example the C++ core of the python interface. I don't hate python/julia and not a Go/JS fanboy. -- 1e100 (talk) 11:16, 18 December 2021 (UTC)Reply

I agree, any language could in principle be used for differential computing, it's kind of silly to single out specific languages to list here DinoInNameOnly (talk) 23:53, 31 August 2022 (UTC)Reply
+1: This should be removed. Stellaathena (talk) 16:01, 15 March 2023 (UTC)Reply

"Implementation" Grouping is Weird

[edit]

As far as I can tell, the "Implementation" grouping is most similar what AI researchers refer to as "modality." However there are a few notable deviations from that:

1. Audio and Visual modalities are typically treated separately.

2. AlphaFold is really it's own modality ("biomedical structure")

3. It is typical to consider multimodal models such as DALL-E and Stable Diffusion as their own category distinct from unimodal image models

4. "Verbal" is typically referred to as "text," as most NLP research focuses on text data rather than spoken data

Stellaathena (talk) 16:01, 15 March 2023 (UTC)Reply