Wikipedia:Articles for deletion/Deeplearning4j


 * The following discussion is an archived debate of the proposed deletion of the article below. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review).  No further edits should be made to this page.

The result was no consensus.  Sandstein  11:10, 18 August 2020 (UTC)

Deeplearning4j

 * – ( View AfD View log  Stats )

What makes this piece of software notable per WP:NSOFT/WP:GNG? My BEFORE did not find any reviews. No awards are listed.There are few hits on Google Books, but I don't see any analysis of the software's importance, all I see are just few pages in manuals about how to use it. There are also some mentions on Google Scholar, but the cited articles have rather low citations (30 is best and comes from a Russian conference, then one paper has 15 citations, and then it gets worse), and the academic papers in question don't discuss Deeplearning4j, they just use it. In other words, it seems to a niche and non-notable tool (software). The article has been tagged as 'promotional' since 2017 (tag added by User:DGG). and given that the (now inactive) article's creator also created Articles for deletion/Chris Nicholson (entrepreneur) which is about to be deleted for similar reasons, I think we should scrutinize all of their creations for similar problems (lack of notability/advertising).PS. I just checked and they have created 15 articles, and as of now, 7 have been deleted already... Piotr Konieczny aka Prokonsul Piotrus&#124; reply here 06:26, 24 July 2020 (UTC)
 * Note: This discussion has been included in the list of Software-related deletion discussions. Piotr Konieczny aka Prokonsul Piotrus&#124; reply here  06:26, 24 July 2020 (UTC)
 * Note: This discussion has been included in the list of Computing-related deletion discussions.  Megan Barris   (Lets talk📧)  06:27, 24 July 2020 (UTC)


 * Delete unless somebody can find better references this one is hopeless.  DGG ( talk ) 09:10, 24 July 2020 (UTC)
 * Keep It's the only deep-learning library written in Java for the JVM. This page Comparison of deep-learning software claims it is notable. Is the proposal to remove it from there as well, and then those looking to perform deep learning on the JVM will not know where to go? Michaelmalak (talk) 14:00, 24 July 2020 (UTC)
 * Being the only whatever in whatever context is not enough to show notability, and a claim of notability sourced to another Wikipedia source is insufficient, unless it has a reliable reference there which can be copied here. Wikipedia cannot be a source for itself. As for "those looking to perform deep learning on the JVM will not know where to go" - WP:NOTMANUAL. That's what Wikibooks are for, or many other sites. --Piotr Konieczny aka Prokonsul Piotrus&#124; reply here 01:43, 25 July 2020 (UTC)
 * Regarding books, the assertion at the top of "just few pages in manuals about how to use it" is false. There are entire print books written on it: Michaelmalak (talk) 15:55, 25 July 2020 (UTC)
 * Java Deep Learning Cookbook: Train neural networks for classification, NLP, and reinforcement learning using Deeplearning4j
 * Java Deep Learning Projects: Implement 10 real-world deep learning applications using Deeplearning4j and open source APIs
 * Hands-On Artificial Intelligence with Java for Beginners: Build intelligent apps using machine learning and deep learning with Deeplearning4j
 * Deep Learning for Search
 * The following books have entire chapters devoted to it: Michaelmalak (talk) 15:55, 25 July 2020 (UTC)
 * Machine Learning in Java: Helpful techniques to design, build, and deploy powerful machine learning applications in Java, 2nd Edition
 * Scala Machine Learning Projects: Build real-world machine learning and deep learning projects with Scala
 * Mastering Java for Data Science: Analytics and more for production-ready applications
 * Java: Data Science Made Easy
 * Deep Learning: Practical Neural Networks with Java
 * I don't see how 4th is dedicated to the subject? It does not mention it in the title or the blurb. All other books are published by Packt which is not described as self-publishing; but I am not really familiar with the publishing house. I Amazon-looked inside one (Java Deep Learning Projects: Implement 10 real-world deep learning applications using Deeplearning4j and open source APIs). It is written by PhD student, which does inspire me (in my field books are usually written by people who got their PhD already but maybe in computer sciences it is different?...). The book also contains a bio of the reviewer, who is an industry professional with a Masters degree. For Java Deep Learning Cookbook: Train neural networks for classification, NLP, and reinforcement learning using Deeplearning4j there is no information about degrees of the author, there are two reviewers only one of which has that information (Masters). And for Hands-On Artificial Intelligence with Java for Beginners: Build intelligent apps using machine learning and deep learning with Deeplearning4j, the author has a PhD, through this book did not provide information about the reviewers. Anyway, they are obviously textbooks which either use the software and/or teach how of use it to a varying extent (from book to book). Thank you for finding the sources. The question is, are they reliable? Reviews on the Internet vary but for example from "Packt books are written by people with a tenuous grasp on the English language and don't seem to be edited at all before being released. The books are VERY low quality and often full of incorrect information.". Now, there are also good reviews online but I am always wary of them as you never know which have been paid for, but, while I acknowledge computer sciences has different habits than my field, I'd be careful of using a textbook written by someone who only did a Masters. --Piotr Konieczny aka Prokonsul Piotrus&#124; reply here  03:40, 9 August 2020 (UTC)
 * Packt books are indeed not held in high esteem; they emphasize quantity and time-to-market over quality. But they are in print and they do sell as people do find them useful. The fourth book is by a publisher held in higher esteem, Manning (disclosure: I have a book published with them), which would be second publisher from the top in the genre (the top being O'Reilly). That fourth book has a "Look Inside" on Amazon, and I encourage you to look at the first half of the first paragraph of section 1.4 on page 12.Michaelmalak (talk) 04:29, 9 August 2020 (UTC)
 * I would suggest you ping the other participants here who voted 'delete' and see what they think. I admit you found better sources than I thought exist, and I am not so strongly convinced this needs to be deleted now as I was before, but still, I find the coverage borderline - just not sure which side to lean on now... --Piotr Konieczny aka Prokonsul Piotrus&#124; reply here 04:43, 9 August 2020 (UTC)
 * Overall, I'm detecting a set of expectations that is unwarranted both in principle and relative to Wikipedia guidelines. DeepLearning4j is a tool for practitioners, not researchers, so it is unreasonable to expect PhD authors to write about it. Would we expect PhD authors for books on, for example, Log4j or Hibernate (framework)? Along the same lines, the consumers of Packt books are looking for hands-on implementation guides, so that is why we see books, as User:Adamant1 points out, that are titled like Using DeepLearning4j for (gasp) Deep Learning, as opposed to waxing philosophical on the framework in isolation of ways to use it. WP:NSOFT contains neither of these proffered provisions: PhD authors or books divorced from all practical use.Michaelmalak (talk) 16:19, 9 August 2020 (UTC)
 * I'm not expecting a book that waxes philosophically about deeplearning4j, just one that is exclusively about it as a tool and doesn't just use it as a "framework" to teach another topic. There's a pretty strong cvase IMO that people reading any of those books are doing so to learn about deep learning and could really give a crap less what framework the books are using to teach it to them. --Adamant1 (talk) 22:25, 9 August 2020 (UTC)
 * Right, good point. A topic that is the subject of a textbook is likely notable, the tool that is used to teach that topic used by the said textbook is much less so. --Piotr Konieczny aka Prokonsul Piotrus&#124; reply here 02:08, 10 August 2020 (UTC)

 Relisted to generate a more thorough discussion and clearer consensus.
 * delete WP:NOTADIRECTORY. Wikipedia is not an annotated product catalog. Graywalls (talk) 23:34, 30 July 2020 (UTC)
 * KEEP Here is an explanation how important Deeplearning4j is to the Java development community. https://www.youtube.com/watch?v=SJldOOs4vB8 --Conrad Kilroy (talk) 16:31, 9 August 2020 (UTC)

Please add new comments below this notice. Thanks, Natg 19 (talk) 01:27, 1 August 2020 (UTC)  Relisted to generate a more thorough discussion and clearer consensus. Relisting comment: This discussion would benefit from further input regarding the sources that have been presented.

Please add new comments below this notice. Thanks, North America1000 05:53, 8 August 2020 (UTC)
 * delete The article just seems to be a trivial listing of features and Wikipedia isn't a directory or software manual. The claimed "entire print books written on it" aren't about it either. Brief passing mentions of it in books about Scala or whatever just don't cut it. So, I see zero about this that warrants an article or passes the notability guidelines. --Adamant1 (talk) 08:14, 8 August 2020 (UTC)
 * User:Adamant1 writes "The claimed 'entire print books written on it' aren't about it either." This is a false statement. There are four entire print books written on it, three of which have DeepLearning4j in the title. The Scala book has a chapter on it, as I clearly wrote, and was not counted in the list of four entire print books. Michaelmalak (talk) 17:03, 8 August 2020 (UTC)
 * The way I see it is that the "four books written on it" are about something else AND Deeplearning4j. Maybe they "use" Deeplearning4j, but it doesn't make them about "Deeplearning4j" anymore then a book that is in English is a book "about" English. Let alone in-depth coverage of the English language as a topic. Or to put it another way, any "how to do X with Y" type guide is more about X then Y. Especially with computer programming books. At least that's how I see it. Adamant1 (talk) 21:31, 8 August 2020 (UTC)
 * It is an odd argument that a book that has "DeepLearning4j" in its title is not about DeepLearning4j. WP:NSOFT makes no such stipulation. One of its four ways to establish notability is "It is the subject of multiple printed third-party manuals, instruction books, or reliable reviews, written by independent authors and published by independent publishers." ...which DeepLearning4j satisfies. Michaelmalak (talk) 22:49, 8 August 2020 (UTC)
 * Personally, I don't think it's an odd argument. Like you say a way to establish notability is if "It is the subject of multiple printed third-party manuals, instruction books, or reliable reviews, written by independent authors and published by independent publishers" and the subject of all those books is deep learning. Maybe you do that by using DeepLearning4j, but people who are buying those books are mainly buying them to learn about deep learning, not to get a deep dive into how to use DeepLearning4j. Otherwise they would be books like "Mastering C++ Game Development" or I'd even say "C++ Game Development Cookbook." More then half the books you cited don't even have Deeplearning4j in the title anyway and none of them are exclusively "how to use Deeplearning4j" books. The good thing about this is that we can disagree though. --Adamant1 (talk) 22:21, 9 August 2020 (UTC)


 * Keep: I find Michaelmalak's sources and arguments to be convincing. Multiple books with titles like Hands-On Artificial Intelligence with Java for Beginners: Build intelligent apps using machine learning and deep learning with Deeplearning4j obviously indicate that Deeplearning4j is worth writing a book about. — Toughpigs (talk) 03:41, 10 August 2020 (UTC)


 * The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.