Talk:Machine learning/Archive 2

This is an archive of past discussions about Machine learning. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

Archive 2

Material brought here from artificial intelligence

I added a few paragraphs that had been cut from artificial intelligence (which was too long). They now appear in Machine learning § Overview and in Machine learning § Limitations. Feel free to improve the merge if you think that's needed.

Also, if you're an expert, take a look at Machine learning § Other limitations. I'm not an expert, and I don't know if these should have their own header.

One last thing: I don't know whether the contribution in the box below is a valuable contribution or not, and I don't know where you all might want to put it in this article, so please help me find a place for it, or let me know if we should throw it away. ---- CharlesGillingham (talk) 02:47, 13 October 2021 (UTC)

Learners can also work on the basis of "Occam's razor": The simplest theory that explains the data is the likeliest. Therefore, according to Occam's razor principle, a learner must be designed such that it prefers simpler theories to complex theories, except in cases where the complex theory is proven substantially better.^[1]

ML as prediction based on passive observations vs AI as active agent

I don't see how making the following assertion makes any sense:

"The difference between ML and AI is frequently misunderstood. ML learns and predicts based on passive observations, whereas AI implies an agent interacting with the environment to learn and take actions that maximize its chance of successfully achieving its goals.[26]" in section "History and relationships to other fields", subsection "Artifical Intelligence"

Two points: (1) The second part of that sentence describes reinforcement learning, which is one of the three ML paradigms. (2) The idea of intelligent agents only emerged in the 1990s, but the field of AI existed long before. Not all of AI is focused on agents interacting with an environment.

I personally think that this assertion is plain wrong and at the very least extremely confusing to people who are not already familiar with the field.

Both terms are wide-ranging and variable. The statement that you are noting is basically an author putting forth their opinion in that area, and in a way that claims it is the only opinion / answer. IMO it should get deleted or given attribution type wording. North8000 (talk) 19:48, 25 January 2022 (UTC)

Semi-protected edit request on 3 September 2022

This edit request to Machine learning has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.

Examples of Machine Learning The Examples of Machine Learning are:-

Image and Speech Recognition: These are one of the most common uses of ML. Image Recognition is the ability of software to identify objects, places, people, writing and actions in images. Speech Recognition is the ability to translate spoken words into the text. The common goal of image recognition is to classify detected objects into different categories. It is also known as object recognition. Speech recognition focuses on the translation of speech from a verbal format to a text one whereas voice recognition just seeks to identify an individual user’s voice.

Medical Diagnosis: Machine Learning can detect patterns of certain diseases within patient’s electronic healthcare records and inform clinicians of any anomalies. By its developing algorithms it provides information to the machine that can help in imaging and analyze human bodies for abnormalities. Hence, Machine Learning is making healthcare smarter.

Prediction: With the help of Machine Learning, GPS navigation predicts traffic ratio through central traffic managing servers. Businesses use ML in order to recognize patterns and then make predictions about what will appeal to customers and help make a better product.

Finance: Data scientists are always working on training systems to detect flags such as money laundering techniques, which can be prevented by financial monitoring, enhanced with ML. Algorithms of Machine Learning can be used to detect transactional frauds by analyzing millions of data points that humans might miss. Anushka544444 (talk) 07:07, 3 September 2022 (UTC) Introduction to Machine Learning

Objected – direct WP:COPYVIO, low quality, likely even spam. --Zac67 (talk) 07:16, 3 September 2022 (UTC)

India Education Program course assignment

This article was the subject of an educational assignment at College Of Engineering Pune supported by Wikipedia Ambassadors through the India Education Program during the 2011 Q3 term. Further details are available on the course page.

The above message was substituted from {{IEP assignment}} by PrimeBOT (talk) on 19:57, 1 February 2023 (UTC)

Sidebar image should be changed

The current sidebar image is just a 2d illustration of a neural network that has had a projective transformation applied to it to make it look 3d. It is not very informative and the transformation just obscures the clarity of the image. Instead, it should just be the original image, or a different image which is more informative. — Preceding unsigned comment added by Eltyvo (talk • contribs) 11:39, 24 August 2023 (UTC)

Wrong date

The date when Arthur Samuel wrote the first program is 1952 as wikipedia and many other online websites say, however this article says 1959, which should be corrected. — Preceding unsigned comment added by 2601:243:480:285b:d0c1:a859:fb90:cb73 (talk • contribs) 20:55, 28 May 2019 (UTC)

neuromorphic computing

someone should really include neuromorphic a discusion on memristors and thair connection to neural networks would also be nice — Preceding unsigned comment added by RJJ4y7 (talk • contribs) 18:25, 11 June 2020 (UTC)

Approaches

Dimensionality Reduction

It's not clear to me if this is another approach - it is not referenced in the list of three broad category areas at the top of the section and its relationship to the other areas isn't obvious. — Preceding unsigned comment added by 198.36.189.210 (talk) 23:19, 10 April 2023 (UTC)

Thank you Yahyaelmatrwy (talk) 05:48, 17 September 2023 (UTC)

Training models

Should optimization algorithms (in general) be discussed on this Wikipedia page? If so, this is likely the best place to do so.

There is currently no mention of fully decentralized machine learning methods, such as CHOCO-SGD.^[2]^[3] This should definitely be fixed!

References

^ Domingos 2015, Chapter 6, Chapter 7.
^ Kairouz, Peter; McMahan, H. Brendan; et al. (10 December 2019). "Advances and Open Problems in Federated Learning". arXiv preprint. Retrieved 20 November 2020. {{cite journal}}: Explicit use of et al. in: |last3= (help)
^ Koloskova, Anastasia; Stich, Sebastian U.; Jaggi, Martin (1 February 2019). "Decentralized Stochastic Optimization and Gossip Algorithms with Compressed Communication". arXiv preprint. Retrieved 20 November 2020.

Wiki Education assignment: Research Process and Methodology - FA23 - Sect 202 - Thu

This article was the subject of a Wiki Education Foundation-supported course assignment, between 6 September 2023 and 14 December 2023. Further details are available on the course page. Student editor(s): HELLOEXTRACREDIT (article contribs).

— Assignment last updated by HELLOEXTRACREDIT (talk) 19:50, 3 November 2023 (UTC)

[FOOTNOTEDomingos2015Chapter_6,_Chapter_7-1] Domingos 2015, Chapter 6, Chapter 7.

[Kairouz-2] Kairouz, Peter; McMahan, H. Brendan; et al. (10 December 2019). "Advances and Open Problems in Federated Learning". arXiv preprint. Retrieved 20 November 2020. {{cite journal}}: Explicit use of et al. in: |last3= (help)

[choco-sgd-3] Koloskova, Anastasia; Stich, Sebastian U.; Jaggi, Martin (1 February 2019). "Decentralized Stochastic Optimization and Gossip Algorithms with Compressed Communication". arXiv preprint. Retrieved 20 November 2020.

[1]

[2]

[3]