Hello everyone. This is my first post on this forum. I was thrilled to find nLab, as it’s the first time I encountered a group dedicated to “that thing I’ve noticed in math and physics that I can’t quite find the name for”. Ends up the name is higher category theory / cohomology. While I’m just barely caught up on the prerequisites for this field of study, I’m excited to dive in.
But getting to the point: While I’ve been studying math and physics for about 20 years now, my formal education is in computer science. Obviously, machine learning has caught my attention in recent years. I noticed that your forum search method appears to rely upon substring matching. I propose indexing it for semantic search via AI embeddings. For those unfamiliar, embeddings are a way to convert some piece of text into a unit 1600-dimensional vector. These vectors are designed in such a way that the ideal embeddings are ones that produce vectors such that “king” - “man” + “woman” = “queen” (before normalization). One of the key benefits is in the area of search, where you can get the vector embedding for a search query and then return results based on the dot product of the query embedding and each text embedding in the index.
It’s a tool that has been kind of swept under the rug as everyone tends to focus on the generative aspects of NLP. But it seems that it could be very useful in this context for the purpose of building a graph of concepts. These embeddings are not only useful for search, either. They’re also useful for the purpose of clustering / categorization.
My proposal is that, with the help of the administrators, that I could build a search index of embeddings that would enable users to find ideas related to their query (even if those ideas are stated using completely different wording). This is referred to as “semantic search” and I think it might be of great benefit to this effort.
Just an idea I had while diving in here. I look forward to seeing where nLab goes.
]]>I just removed the paragrpah
In addition, the nForum is hosted at NTNU and therefore their guidelines and regulations for what is posted on the nForum apply (this is a special case of Rule 2 since losing the nForum would hinder the work on the nLab!).
from https://ncatlab.org/nlabmeta/show/about+the+nForum, from just under the list in the section “How to behave”. Andrew Stacey pointed it out to me.
If there should be anything replacing that (for instance policies based on who is hosting the physical server), now’s the time to replace it.
]]>Right now the HoTT wiki seems to be de facto an anonymous user’s personal web where they make draft articles before porting them over to the nLab: see i.e. biaction (homotopytypetheory) and biaction (nlab), ETCS with elements (homotopytypetheory) and ETCS with elements (nlab), enriched poset (homotopytypetheory) and enriched poset (nlab), et cetera. Otherwise, there doesn’t seem to be much activity on the HoTT wiki at all. The type theory and homotopy type theory material on the nLab is already much more elaborated than the equivalent articles on the HoTT wiki (see identity type (nlab) and identity type (homotopytypetheory)), and the only user contributing to the HoTT wiki is more interested in their own pet projects in abstract algebra and analysis (i.e. halving group (homotopytypetheory)) than actually expanding on the core homotopy type theory content.
]]>As described at funding of the nLab, the nLab is moving to the cloud! We will be collaborating with the Topos Institute; as again described at funding of the nLab, the Topos Institute will be handling the actual finances, and in particular the handling of donations.
We are now opening for a first round of donations! The goal is to raise enough to support the running of the nLab for a number of years in the cloud. I think it is not unrealistic to hope that we can raise several thousand USD, which would support us for a long time. Many people have indicated previously that they would be willing to donate something.
Please donate something if you can! You can make a donation at the following link:
https://topos.institute/contact
Click ’Donate’, and then select ’nLab Online Wiki Restricted Fund’ in the field ’My donation is for’. As mentioned at funding of the nLab, we will not make donations public (though of course feel free to announce here if you do make a donation, you of course do not need to mention how much!); the contact details are just necessary for the payment to go through.
For now, I think the plan is just to make this announcement here, and to spread it by word of mouth: feel free to spread the word however you wish, e.g. on social media.
I would like too to put a discreet notice at the top of the nLab indicating that we are raising funds, with the option to hide this message; but I have not received clear support for this idea. Please let me know if you think we should do this.
Brendan (Fong) will keep me updated about the progress of the funding; I will post here when we reach notable milestones, or otherwise at intervals.
]]>I would like to request a personal web. The immediate reason is that I want to collect notes for a course on Kant’s logic I am going to supervise. Hopefully, there will be some spin-off for the main nLab.
]]>(Not sure I chose the right category for this discussion.)
There used to be a page which listed nLab authors and the pages they’ve created/edited, and another page that kept count of number of contributions of individual contributors. Do either of those exist anymore?
]]>Dear nLab steering committee,
I’m requesting a personal web. At first I would like it to be password edit-protected, but readable publicly if possible. My initial intention is to use this personal web to compile constructions and theorems related to my current research on categorical gluing, normalization and coherence theorems, internal languages, initiality, etc. Much of the material that starts out here would eventually reach the general nLab at the point when it becomes sufficiently polished.
Thank you, Jon Sterling
]]>Some details regarding personal webs:
I’d like to put a link to Dan Isaksen’s list of journals including at least one algebraic topologist editor (recently posted to the algtop mailing list) somewhere on the nLab, but I’m not sure where. The best extant page I can find is math resources by individuals; but I could imagine also creating a separate page like lists of math journals under math resources. Any thoughts?
]]>For a while, we've been blanking spam pages and putting them in their own category; then people are supposed to use these when they create a new page. This effectively removes the spam page from circulation without having to invoke deletion. Then we found some other junk pages that weren't spam but either mistakes or whatever is going on with those weird slideshow pages. Then we put some of these in category: empty since they weren't spam. But we remarked that they really could all go there.
So now I've put them all there and regularised the names.
Please use these when you make a new page! Only a few people have to remember to do this and they will all go away. (Zoran remembers, but I usually forget.)
]]>Dear all, I have spent few days few weeks ago to create a page called topological state of matter (and few related pages in Lab and in my private lab). This notion, is according to wikipedia, somewhat wider than more narrow notion of topological order. But I am not an expert. A real expert, seemingly Xiao-Gang Wen entered the scene first as Anonymous User and changed the page; his 9 references were the main addition, and in recent edits he (or else) edits under his own name. I like that experts join and help. But the edits were a bit unfair.
He changed the page name to topological order, what I am not the most happy, though the notions are closely related. Topological order was supposed to be a separate page as wikipedia does it with clear argumentation for it (wikipedia – topological order, wikipedia – topological insulator). He put some of his references into the list of references and changed and biased the idea section; some of those are specialized and I would rather put them under quantum Hall effect etc. However, he also ERASED a number of reviews and other references of concurrent authors and did not notify about the changes in Forum. I find this a school example of a selfish spammer behaviour. The most POV article, the one by Freed and Moore has been excluded in Mr Wen’s edit, among other references with this fate!!!!!!!!!! Urs and Toby have done minor edits in the meantime (Urs added physics menu), effectively approving the spammed state; I would call for more attention when overwriting an Anonymous User!
I rolled back the page to version 1 and then repaired some of the damage which this may cause. More careful dealing with vandalism which I found would take too much time. Already this took a precious hour in a day in which I am in a big problem with finishing the things for this week, before travelling away for 2 days.
I now created a separate page topological state of matter, where I moved sentences added to the idea section by Mr Wen. Mr Wen created page for himself, where I moved his 9 references till it is decided how to redistribute better; I also added his review to quantum Hall effect (should finally be in an entry on fractional QHE). Mr Wen also created symmetry protected trivial order and updated Levin-Wen model which I do not plan to touch; I believe those are valuable additions. I retained the reference by Read and Sachdev, which he added to topological state of matter. I also returned the table which Urs has put what has been erased in my rollback.
On the other hand, I added the reference
which is coauthored by Mr Wen (but he did not add it to our Lab) and which is, because of its relation to group cohomology, of prime interest to Lab.
]]>After thinking it over some more, here is a proposal for a convention about page length and sub-pages that I think would make me fairly happy and might also be acceptable to others (perhaps with further tweaking). It’s just an idea; please criticize it — constructively, if possible; but if not, you can do it classically. ;-)
The main organizational units of the nLab would continue to be “main” pages about concepts, generally named after mathematical nouns, as is currently the case. These pages would continue to have sections (## level) such as Idea, Definition, Examples, Related Pages, References, and others that are more flexible in the middle like Properties or Main Theorem. However, I propose two new guidelines regarding such pages:
The material that is excluded by these guidelines goes on sub-pages having a uniform naming scheme, ideally indicating not just the main page that they belong to but the section of it that they are an expansion of. So, for instance, we might have a page called “semigroup/Definition - historical notes
” (as one possibility). Links to these pages are placed in the appropriate section of the main page, along with a brief summary of their content. Moreover, at the top of such sub-pages there should be a “breadcrumb” hatnote linking to the main page that they belong to.
The content of sub-pages is less restricted than that of main pages, and may depend on the purpose of the sub-page. For instance, a hypothetical sub-page “infinity-category/Idea - for novices
” might be written in a very chatty style with lots of pictures and not much organization, while a hypothetical sub-page “(infinity,1)-category/Definitions - web of Quillen equivalences
” might be organized with one section for each Quillen equivalence between a pair of definitions and each section potentially quite lengthy in definition-theorem-proof style. I suppose sub-pages could even have sub-pages, but the naming conventions would get less wieldy.
I’m inclined to discourage linking directly to sub-pages (and, by extension, adding redirects to sub-pages), because such links would break if we decided to reorganize the division of a page into sub-pages. However, I could be convinced otherwise.
There is one further enhancement to this proposal that I think would make it much more reader-friendly, but would require a bit of coding to achieve: transclude the sub-pages into the main page as knowls. I’m imagining a new link syntax for instiki, such as {{page name}}
, that would create a knowl rather than a link — or, even better, a knowl and a link (maybe the page name itself is a knowl, while after it there is a small icon that serves as a link), since otherwise there would be no easy way to edit the sub-page. Moreover, we would probably want some sort of syntax for the sub-pages allowing us to exclude parts of them upon transclusion (such as the title and breadcrumb links), like mediawiki’s noinclude and includeonly.
Thoughts?
]]>Once I wrote this nLab article as an extract from my new general topology theory.
After this I was told that if I keep writing in nLab wiki, I could be considered as a person which writes in foreign lab book and may be banned (at least approximately so).
Now I have released a good draft of my book (I released it as Creative Commons and probably will never publish it “officially”). The draft represents the current research slice. It is dubbed as a draft because it most probably contains some minor errors.
Please somebody review quickly my book and produce allowance for me to write extracts from my book in nLab wiki: I need a media to advertise my free ebook and nLab needs new research.
Note that some concepts from my book apply to abstract category theory (notably my “cross-composition product” is defined for certain partially ordered (I mean with partially ordered Hom-sets) categories, probably for all interesting partially ordered categories in a sense). Also my “category with star morphisms” is an abstract category theory level stuff. These need to be presented in nLab.
]]>While the nLab is down (edit: now back up), maybe I should use the time to finally say the following, since it already came up elsewhere.
As of recently, Andrew Stacey has handed back the tenant-ship of the Lab server to me and is withdrawing from administering the Lab software installation. From now on Adeel Khan Yusufzai is kindly offering help with/is taking care of whatever it takes to administer the server. I hope the workload will be light.
]]>In cleaning up some spam, I may have inadvertently changed the name ot the original ’genral discussion’ which recorded the disccusions on the nCafé about the style etc, of the nLab, before the creation of this forum. I rolled back some but this did not refind the name. (I say more of this on my spam comment.} It is now called general discussion but it is not clear where it should be ‘put’.
(Edit: Later some ’kind’ spammer has attacked the old General Discussion which led me to fix the new page by copying the old page’s source. I hope this works. There may be some ’recent’ edits that may have been lost, but as this is really archive materialrecent edits are not that important, I hope.)
]]>So, apparently twisted forms links to the differential forms page, and is related to the notion of -twisted forms for a line bundle . However, this is also the name of the more general notion of possible descent data on a pulled back sheaf (where I’m being vague about context because it’s a pretty general idea). The notion at differential forms is just one particular example of this. In particular, such things are classified by nonabelian cohomology. Anyway, in an attempt to deal with this, I made a page called formes tordues which is just the French name for it (and the name that Knus and Ojanguren use in one of my favorite references on the subject), though I’ve not written anything there. I figured, however, that I should ask here before going any further, since that’s kind of snarky, just making another version of the page in French…
Thoughts?
-Jon
]]>There is now a brand-new homotopy type theory wiki, hosted as another web on the nlab server: HomePage (homotopytypetheory). There’s not much there yet, though. We’re still discussing its goals and policies and conventions.
]]>Skimming through a number of pages which refer to Lurie’s HTT I noticed that the version which is often linked is the arxiv one (last revised 31 Jul 2008). It is awfully old; wouldn’t it be better to give the reference of the newest one, (last update: February 2012)?
Can this renewal be done with some smart automatic procedure?
]]>A fluke has put some content at generator, but that is a duplicate that won't last. What I want to ask is, should generator redirect to (or just have the content of) generator: disambiguation or separator? (Or generators and relations, but I don't think so.)
]]>Why do we want closed space to redirect to closed subspace?
]]>The fact that publicly editable personal webs still acquire spam has led the steering committee to try to figure out a way to deal with dormant webs. What we decided was to turn them into “private and published”. Following this decision, I wrote to all of the maintainers of the webs that I considered to be dormant telling them that this was what we proposed doing. Some responded, some didn’t. Those who did all agreed that this was a reasonable thing to do. I’ve now carried out this proposal.
Among the dormant labs was the CatLab, which was our main non-nLab attractor of spam. Hopefully this will relieve that.
All of the webs are still available. However, not all links will work as they used to since the URL is now http://ncatlab.org/<web>/published/<page>
instead of show
. Cross-links from other webs will work.
In light of increasing spam with forged signatures, I want to formally propose that we equip the nLab with registered password-protected editor accounts. This requires that we talk to Jacques about modifying Instiki. So first we need to decide how we want him to modify it.
Here are my suggestions:
Here are some more suggestions that I think that Jacques should be freer to change if it helps with the programming:
I'm trying to keep the interface simple, in line with Instiki's philosophy; but this means that users cannot change their passwords later. These things really don't need to be secure, just something that spammers won't guess.
]]>I thought I should drop this note here: I’ll be mostly offline from about right now to an entire three weeks from now.
I’ll try to get online to check for anything of importance once in a while, but otherwise if you think you ought to hear from me but don’t, then it’s because in that chateau in the middle of nowhere in southern France, they won’t have any form of internet connection, probably not even the cell-phone connection that I’d need.
]]>