Talk:Weak supervision

The contents of the Semi-supervised learning page were merged into Weak supervision on 30 October 2022. For the contribution history and old versions of the redirected page, please see its history; for the discussion at that location, see its talk page.

This article has not yet been rated on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Computer science This article is within the scope of WikiProject Computer science, a collaborative effort to improve the coverage of Computer science related articles on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.Computer scienceWikipedia:WikiProject Computer scienceTemplate:WikiProject Computer scienceComputer science articles ??? This article has not yet received a rating on the project's importance scale. Things you can help WikiProject Computer science with: Here are some tasks awaiting attention: Article requests : Requested articles/Applied arts and sciences/Computer science, computing, and Internet Cleanup : Computer science articles needing attention Computer science articles needing expert attention Copyedit : Computing Expand : Computer science Infobox : Computer science articles without infoboxes Maintain : Timeline of computing 2020–present Photo : Find pictures for the biographies of computer scientists (see List of computer scientists) Computing articles needing images Stubs : Computer science stubs Unreferenced : WikiProject Computer science/Unreferenced BLPs Project-related : Tag all relevant articles in Category:Computer science and sub-categories with {{WikiProject Computer science}}

It conflates semi-supervised learning and weak supervision. The two are related but distinct areas of machine learning. Semi-supervised learning DOES NOT use noisy, limited, or imprecise sources. Instead, it propagates information from (small amounts) of labeled data to unlabeled data. Weak supervision on the other hand does use noisy, limited, or imprecise sources about an unobserved target variable. 72.182.48.217 (talk) 16:07, 22 February 2023 (UTC)[reply]

I created a new page for weak supervision and separated the two topics. The new page for weak supervision comments on the difference to semi-supervised learning. Psnby (talk) 17:20, 22 February 2023 (UTC)[reply]

My edit was undone, so now this page is wrong again. I have a PhD in machine learning and I can assure everyone that weak supervision and semi-supervised learning should not be conflated in the same article. Psnby (talk) 18:16, 22 February 2023 (UTC)[reply]

@

reliable source that differentiates the two. (Your PhD, while impressive, is not a reliable source. You'll need to point to actual published sources.) WikiDan61^ChatMe!_ReadMe!! 18:19, 22 February 2023 (UTC)[reply

]

An example would be the article below (cited 24 times at this time), which while a preprint, includes Alex Ratner as an author. Alex Rather is an author whose research articles are cited as primary sources on weak supervision in this wiki entry. The paper below clearly distinguishes semi-supervised learning as a different, complementary approach to weak supervision in Section 7.

Zhang, J., Hsieh, C.Y., Yu, Y., Zhang, C. and Ratner, A., 2022. A survey on programmatic weak supervision. arXiv preprint arXiv:2202.05433. Psnby (talk) 18:30, 22 February 2023 (UTC)[reply]

@Psnby: Very well. Thank you for clarifying. WikiDan61^ChatMe!_ReadMe!! 22:19, 22 February 2023 (UTC)[reply]

I submitted a draft for an article on Weak Supervision which just separates out the weak supervision component of the Semi-supervised learning entry and makes tiny edits such as a clarification regarding the difference to Semi-supervised learning. If there is anything I can do to make things clearer or help the process along I'd appreciate some help as I am still very unfamiliar with editing etc on this platform. Psnby (talk) 22:31, 22 February 2023 (UTC)[reply]

@Psnby: The warning you placed is less than helpful. If you are able to improve the article, do so. (Reference this discussion if you make large changes that might be questioned again.) WikiDan61^ChatMe!_ReadMe!! 12:32, 23 February 2023 (UTC)[reply]

Here is another well-cited survey paper which distinguishes weak supervision from semi-supervision:

Zhou, Zhi-Hua. "A brief introduction to weakly supervised learning." National science review 5.1 (2018): 44-53.

So I think this present article is confusing on two fronts: first semi-supervision is rarely called weak supervision. For instance, both survey articles referenced in this article call the subject of this article "semi-supervision". Almost all of the references in the article refer to this problem as semi-supervision and I cannot see a single instance where it is called weak supervision.

Second, weak supervision usually denotes a problem distinct from the semi-supervision. In addition to the two, papers already mentioned in this thread I'd add the ECCV tutorial: https://hbilen.github.io/wsl-eccv20.github.io/ . I'd also add these survey papers:

Zhang, Man, et al. "A survey of semi-and weakly supervised semantic segmentation of images." Artificial Intelligence Review 53 (2020): 4259-4288.

Zhang, Dingwen, et al. "Weakly supervised object localization and detection: A survey." IEEE transactions on pattern analysis and machine intelligence 44.9 (2021): 5866-5885.

Qu, Linhao, et al. "Towards label-efficient automatic diagnosis and analysis: a comprehensive survey of advanced deep learning-based weakly-supervised, semi-supervised and self-supervised techniques in histopathological image analysis." Physics in Medicine & Biology (2022).

I'm not sure how to demerge these pages by the wiki process. Genusfour (talk) 07:54, 3 June 2023 (UTC)[reply]

I think It would be useful to open a new thread/issue. Clearly, the issue has not been addressed. 136.49.39.161 (talk) 19:14, 12 June 2023 (UTC)[reply]

I'm happy to second @Psnby's opinion. The existing article is very misleading, and most of it is describing semi-supervised learning but not weak supervision. I think the easiest fix is to change the title perhaps WuTheFWasThat (talk) 18:51, 30 November 2023 (UTC)[reply]