投稿時間:2021-11-23 05:22:06 RSSフィード2021-11-23 05:00 分まとめ(29件)

カテゴリー等	サイト名等	記事タイトル・トレンドワード等	リンクURL	頻出ワード・要約等/検索ボリューム	登録日
Program	[全てのタグ]の新着質問一覧｜teratail（テラテイル）	jspで誕生日を入力したら年齢を割り出せるプログラムを実装したいです。	https://teratail.com/questions/370571?rss=all	自分の誕生日を入力したら年齢を割り出せるようにしたいのですがうまくいきません。	2021-11-23 04:39:18
Program	[全てのタグ]の新着質問一覧｜teratail（テラテイル）	gnome-terminalをSSH越しにGUIのウィンドウとして起動したい	https://teratail.com/questions/370570?rss=all	gnometerminalをSSH越しにGUIのウィンドウとして起動したいWindowsのWSLからSSH接続でUbuntuに接続しています。	2021-11-23 04:26:36
Docker	dockerタグが付けられた新着投稿 - Qiita	Rails×RSpec×Docker×CircleCI	https://qiita.com/supyolo888/items/14ee51c4e4cac2d39757	dockercomposerunrailsnewforcedatabasemysqlGemfileが更新されたのでコンテナのイメージをビルドし直す。	2021-11-23 04:02:55
GCP	gcpタグが付けられた新着投稿 - Qiita	Google Associate Cloud Engineer 資格取得	https://qiita.com/redpanda/items/a238840e26db0e104d67	AssociateCloudEngineer認定資格  GoogleCloud筆者のスペック昔は主にサーバサイドエンジニアやDBAインフラ屋さんではない最近はデータエンジニアっぽいことをしている年くらい前に「AWSCertifiedSolutionsArchitect」の資格を取得GoogleCloudは、BigQueryとストレージについてはある程度理解し業務で触っていた英語はたいしてできないやったこと参考書を読むCourseraのビデオを観て学習するQwiklabsでハンズオン公式ドキュメントを読む模擬試験それぞれについて、次に詳しく記載します。	2021-11-23 05:00:16
Ruby	Railsタグが付けられた新着投稿 - Qiita	Rails×RSpec×Docker×CircleCI	https://qiita.com/supyolo888/items/14ee51c4e4cac2d39757	dockercomposerunrailsnewforcedatabasemysqlGemfileが更新されたのでコンテナのイメージをビルドし直す。	2021-11-23 04:02:55
海外TECH	MakeUseOf	Should You Buy a Nintendo Switch on Black Friday 2021?	https://www.makeuseof.com/should-you-buy-nintendo-switch-black-friday-2021/	black	2021-11-22 19:40:39
海外TECH	MakeUseOf	9 Useful Tips to Help You Get the Most Out of Google Chat	https://www.makeuseof.com/9-useful-tips-to-help-you-get-the-most-out-of-google-chat/	useful	2021-11-22 19:30:23
海外TECH	MakeUseOf	Find Where You Waste Your Time on Windows With AutoHotKey	https://www.makeuseof.com/log-active-windows-with-autohotkey/	Find Where You Waste Your Time on Windows With AutoHotKeyApps can be serious time vampires but it can be hard to pinpoint what s doing the most damage to your productivity Fortunately AutoHotKey can help	2021-11-22 19:16:12
海外TECH	MakeUseOf	What Is the Zone System? Photography Meets Science	https://www.makeuseof.com/what-is-zone-system-photography/	learn	2021-11-22 19:02:41
海外TECH	MakeUseOf	Early Access: Govee Black Friday Deals Up to 50 Percent Off	https://www.makeuseof.com/early-access-govee-black-friday-deals/	luminous	2021-11-22 19:02:41
海外TECH	DEV Community	How to build a search engine with word embeddings	https://dev.to/mage_ai/how-to-build-a-search-engine-with-word-embeddings-56jd	How to build a search engine with word embeddingsImagine you have a large set of news articles and want to provide a way to allow your users to search through them For example here are random articles from the Economist An investment bonanza is comingWho governs a country s airspace What is a supermoon and how noticeable is it to the naked eye What the evidence says about police body camerasWho controls Syria Let s think about what kind of queries potential users issue and the articles that we would want to return Exact and Substring MatchesA search for “investment should return the st article A search for “supermoon should return the rd article A search for “syria should return the th article This is quite simple to implement with traditional information retrieval techniques since these queries have keyword matching Harder ExamplesA search for “banking should probably return the st article A search for “astrology should return the rd article A search for “middle east should return the th article The challenge with these queries is that none of them are actually a keyword match in the original documents However there isn t any string matching we can perform between the search query and the document What we need is a way to perform searches over text without requiring exact keyword matches between words in the query and words in the documents Word EmbeddingsOne way we can do this is with word embeddings which capture the semantics of words in vector representation If we can find a way to create a vector for the query and a vector for every document in the same vector space then we can just compute the similarity between our query vector and every document to retrieve the document that is most semantically similar to our query Creating Document Embeddings from Word EmbeddingsAssuming we have lookup table of a word to its embedding how do we actually create an embedding representing a string of words We can think of the semantics of a document as just the average of the semantics of individual words and compute a mean word embedding to represent a document Specifically def create mean embedding words return np mean model word for word in words if word in model axis This would capture the average semantic of a document or query Another way we can build a document embedding is by by taking the coordinate wise max of all of the individual word embeddings def create max embedding words model return np amax model word for word in words if word in model axis This would highlight the max of every semantic dimension We can also concatenate the two embedding types to get a better representation of the document def create embedding words return np concatenate create max embedding words create mean embedding words Let s try apply this to the simple example documents we highlighted above We will be using gensim to load our Google News pre trained word vectors Find the code for this here Ranking with embeddingsQuery banking Rank Top Article An investment bonanza is comingRank Top Article Who governs a country s airspace Rank Top Article What is a supermoon and how noticeable is it to the naked eye Rank Top Article Who controls Syria Rank Top Article What the evidence says about police body camerasQuery astrology Rank Top Article What is a supermoon and how noticeable is it to the naked eye Rank Top Article Who governs a country s airspace Rank Top Article Who controls Syria Rank Top Article What the evidence says about police body camerasRank Top Article An investment bonanza is comingQuery middle east Rank Top Article Who controls Syria Rank Top Article Who governs a country s airspace Rank Top Article An investment bonanza is comingRank Top Article What is a supermoon and how noticeable is it to the naked eye Rank Top Article What the evidence says about police body cameras Dual space embeddingSo far we ve used the standard pre trained wordvec embeddings to represent both the query and document embeddings But lets think about how wordvec constructs these vectors to come up with a more clever way to create these embeddings The core idea of wordvec is that words that are used in similar contexts will have similar meanings Focusing on the skip gram variant of the model this idea is formulated as trying to predict the context words from any given word Check out this post if this is confusing to you In a skip gram model a one hot encoded word is mapped to a dense representation which then is mapped back to the context of that word context in this case refers to the words that surround the target word Usually we keep the first transformation matrix W in the diagram below which becomes our word embeddings and discard the matrix that maps to context words W If we think deeply about the behavior of W and W matrix the W embedding s goal is to map a word to an intermediate representation and the W matrix maps from the intermediate representation to a context This naturally results in W capturing the semantics of a word while W captures the context of a word We can frame our search problem in a similar way The query being the “word and the document being the “context In this way if the query is embedded using W while the document is embedded using W comparing the similarity between the two vectors would indicate whether or not the document acts as a good context for the query Lets consider two simple documents that we could potentially return for a query “harvard and decide which one is more relevant Document Harvard alumni receives Document Yale NYU and Harvard As a user you d probably want to see the first document However our current embedding model would assign the second document a higher relevance score Why Because the terms “Pittsburg Steelers “Baltimore Ravens and “Cincinnati Bengals are similar to the term “New England Patriots since they are all football teams Recall that the wordvec model measures similarity as words that have similar contexts Football team names would exists in similar context Therefore the embedding for the query and document will be more similar than the query and document However using W to create the document embedding and W to create the query embedding will allow a similarity measure of “how well does this document represent the context of this query Indeed these are the results published in A Dual Embedding Space Model for Document Ranking This table above shows the closest word embeddings to the words “yale with three different embedding comparisons IN IN The bolded word is embedded with W and the words below are the closest word embeddings in W OUT OUT The bolded word is embedded with W and the words below are the closest word embeddings in W IN OUT The bolded word is embedded with W and the words below are the closest word embeddings in W Observe that while “Yale is more semantically similar to “Harvard if a document about Harvard came up when searching “Yale that wouldn t make a lot of sense With that in mind it seems the IN OUT embedding scheme provides the most relevant words for a search system If someone searched for the word “Yale I would expect documents with the terms “faculty “alumni and “orientation rather than “harvard and “nyu How do we deploy this Too many ML articles stop here but deployment is typically half the challenge Typically these models are deployed as a ranker after a retrieval model The pipeline is typically structured like this The purpose of the retrieval model is to find the most similar documents based on keywords among millions of other documents while the purpose of the ranking model is to find the most semantically relevant document in the set of BM is a common document retrieval method which is implemented in most search systems First lets implement a retrieval model that leverages the rank bm package In a production setting we would probably use something like Elasticsearch for this part from rank bm import BMOkapi as BMclass Retriever object def init self documents self corpus documents self bm BM self corpus def query self tokenized query n scores self bm get scores tokenized query best docs sorted range len scores key lambda i scores i n return best docs scores i for i in best docs Next let s build the ranking model which takes a list of documents and re ranks them according to the similarity between the query embedding and the document embedding And now let s stitch the code together Let s take a look at a real example output of this for the query “harvard hockey against the Microsoft News Dataset However there are no documents in the corpus containing both the terms “harvard and “hockey making this a tricky problem Here are the top results for the query “harvard hockey from the BM model The documents with the highest score from the BM retrieval model is “Admissions scandals unfolding at Harvard university and “How a Harvard class project changed barbecue I suspect many people do not think this is a very relevant article to “harvard hockey But if we re rank the top results from the BM model with the embedding based model here is what we get Now the results are much more semantically relevant to college hockey results SummaryA dual space embedding approach to creating query and document embeddings for embedding based ranking can offer more contextually relevant results than the traditional word embedding approach ReferencesDual Embedding Space Model	2021-11-22 19:43:42
海外TECH	DEV Community	Learn how to use Git and GitHub in a team like a pro (part 2)	https://dev.to/colocodes/learn-how-to-use-git-and-github-in-a-team-like-a-pro-part-2-11j	Learn how to use Git and GitHub in a team like a pro part In the previous part of this tutorial we learned how Harry and Hermione decided to build a SaaS app to allow people to build their own potions online and share them with the rest of the world They named it Potionfy Hermione created a remote repository then an issue to address the task of building a landing page and how Harry worked on that issue locally and created a pull request once he finished working on it You can read the first part of the tutorial here Learn how to use Git and GitHub in a team like a pro Damian Demasi・Nov ・ min read github webdev beginners tutorial Now we are going to see how Hermione reviews Harry s code how the code is merged on the master branch the decision of using a develop branch how the team works in the develop branch and merges changes into main and how the team solves merge conflicts Code review Step Creating a reviewHermione has finished with her marketing and promotion tasks and she now has time to review Harry s code In order to do so she opens the GitHub repository and clicks on the Pull requests tab to find Harry s pull request After clicking on it she then clicks on the Commits tab and finally in Harry s last commit this is just one way of accessing the files modified on the pull request She is not entirely convinced about the lt h gt code so she clicks on the plus icon that appears when she hovers that line of code and writes a comment to Harry Finally she clicks on the Start a review button As she has no other comments about the code she now clicks on the Review changes button to make the review visible to the rest of the team You can find more information about making reviews in this Reviewing proposed changes in a pull request article Step addressing the review and creating a code changeHarry checks his pull request and finds a new conversation there Hermione s review Harry answers Hermione s comment and clicks on the Resolve conversation button Now that the conversation is resolved Hermione can submit the review indicating that there are requested changes so Harry can actually work on them Note this is just one version of the review process that can be achieved with GitHub and it can differ from the actual way your team chooses to handle them Harry checks the pull request again and finds that it has Changes requested Step implementing the changesAs Harry likes to work locally he continues working on the branch he had created in order to implement the code changes git checkout add landing messageOnce he is sure he is working on the correct branch he makes the changes in the index html file Note for simplicity sake we are not creating a separate CSS file here Once Harry finishes tweaking the code he stages the changes commits them making sure to include the id of the issue because he is still working on it and pushes them to GitHub git add git commit m Add colour and remove text git push Step merging the pull requestNow it s Hermione s turn She goes to the pull request and finds a new commit the one Harry did and pushed to GitHub She then clicks on the Files changed tab and finds the ones that she suggested implemented on the index html file As she is satisfied with the changes she proceeds to approve them by clicking on the Review changes button and selecting the Approve option Harry sees that his pull request was approved by Hermione and he proceeds to merge it into the main branch of the project He decided not to delete the branch as he wants to leave it there for future reference although it would be a good idea to delete it As Hermione is satisfied with how the issue was solved she proceeds to close it by going to the Issues tab and clicking on the Close issue button If you want to see a graphical representation of the whole process up to this point you can click on the Insights tab and then on the Network option You will be able to actually see how the branching and merging were performed Using a develop branchWhen working with real projects merging changes into the main branch like you saw up to this point is not recommended Instead of working directly with the main branch often called production you will be working with a develop branch You will be branching issues out of that develop branch and merging them back into the develop branch Once a group of issues have been solved that develop branch will be merged into the main or production branch usually denoting a version change in the app Hermione is aware of this and now that the landing page is live and accessible to customers she decided to preserve that production environment and work on a development branch In order to do so she creates a develop branch out of the main one so she and Harry can work on that branch without impacting the production environment Merge conflictsHermione wants to add a new feature to the landing page a form to capture clients emails In order to do so she creates a new issue Once the issue is created Harry decides to start working on it To do so he branches out from the develop branch by selecting that branch on the GitHub interface a new one called email form including the issue number at the front to make it clear how this branch will relate to the issues He then pulls that branch locally and starts working on it git pull git checkout formHarry decides to include a simple form into the index html file lt form action mailto hermione potionfy com method post enctype text plain gt Name lt br gt lt input type text name name gt lt br gt E mail lt br gt lt input type text name mail gt lt br gt lt input type submit value Send gt lt input type reset value Reset gt lt form gt Note This code is just to exemplify how Harry is working on a file and it s not how this type of form could actually be built Harry stages and commits his changes locally using the Contact form message git add git commit m Contact form git pushBefore Harry could create a new pull request Hermione decides to build a placeholder for the form on the index html file on her own In order to do so she creates a new branch out of develop called email form placeholder To work on the index html file she uses the GitHub online code editor basically a VSCode for the web In order to open it she just presses the key on her keyboard and the GitHub page is transformed into a VSCode interface like magic She then proceeds to add the following code to the file lt form action mailto harry potionfy com method post enctype text plain gt lt form gt After saving the file she commits the changes right there on her browser window by using VSCode graphical interface Once the commit is complete she opens GitHub again and decides to create her own pull request and merge her changes to the develop branch On the other hand Harry also decides to create a pull request to merge his changes into the develop branch At this point GitHub lets him know that his pull request won t be able to merge automatically to the develop branch Harry supposes that his branch is no longer reflecting the state of the develop branch and that the develop branch has changed because someone else merged changes affecting the index html file on which he was working on Nevertheless he proceeds to create a pull request What he sees next is GitHub s way of letting him know that there is a conflict affecting the file he modified He proceeds to click on the Resolve conflicts button He can now see that the index html indeed was modified and the changes made to that file are affecting the lines he himself modified For more information about solving conflicts you can read the Resolving a merge conflict on GitHub article Harry proceeds to modify the file directly on the GitHub site to remove the conflicting changes and then clicks on the Mark as resolved button Once the conflict is marked as resolved he clicks on the Commit merge button Finally his branch was conflict free and he can merge his pull request assuming Hermione reviewed his code and approved it just as she did earlier Conflicts ofter arise when teammates are working on different branches that affect a common file A great way to prevent merge conflicts is to do a pull request on the develop branch merge that updated develop branch into the branch you are working on and then do a push followed by a pull request git branchx my branch This is an example name git checkout develop git pull git checkout x my branch git merge develop You make some changes on the files of the x my branch branch git add git commit m lt a message gt git push Final thoughtsAfter working on their landing page Harry and Hermione managed to get lots of email addresses from potential customers and continued developing their MVP They managed to get funding from a local venture capital firm and they are now in the process of hiring other developers to launch Potionfy to the public I m sure they would love to take a look at your resume to consider you for a position in their company so good luck ️NEWSLETTER If you want to hear about my latest articles and interesting software development content subscribe to my newsletter I will be giving away a Udemy course among my newsletter subscribers if we go over during the month of November TWITTER Follow me on Twitter	2021-11-22 19:03:17
Apple	AppleInsider - Frontpage News	How to enable and use Reachability on an iPhone running iOS 15	https://appleinsider.com/articles/21/11/22/how-to-enable-and-use-reachability-on-an-iphone-running-ios-15?utm_medium=rss	How to enable and use Reachability on an iPhone running iOS If you re having trouble navigating the larger display of your iPhone Apple s Reachability feature can make it significantly easier to use your device one handed Reachability is the name for the iPhone feature that brings the content halfway down your device s screen which makes for more comfortable one handed navigation In older iPhones that featured a home button Reachability could be activated by double tapping on the Home Button Read more	2021-11-22 19:34:32
Apple	AppleInsider - Frontpage News	Black Friday: Buy two books get one free from Amazon	https://appleinsider.com/articles/21/11/22/black-friday-buy-two-books-get-one-free-from-amazon?utm_medium=rss	Black Friday Buy two books get one free from AmazonAmazon is offering a special deal where customers can buy two books from nearly every genre off a select list and get a third for free Amazon has a new book dealIt s not specifically branded as a Black Friday deal but Amazon s new for offer is only for a limited time That period hasn t been confirmed yet so it s likely to continue as long as there is demand and sufficient books Read more	2021-11-22 19:34:03
Apple	AppleInsider - Frontpage News	Black Friday Deals: AirPods 3 dip to $154.99, AirPods Pro with MagSafe drop to $169.99 at Amazon	https://appleinsider.com/articles/21/11/22/black-friday-deals-airpods-3-dip-to-15499-airpods-pro-with-magsafe-drop-to-16999-at-amazon?utm_medium=rss	Black Friday Deals AirPods dip to AirPods Pro with MagSafe drop to at AmazonBlack Friday week has officially started and prices have been slashed on Apple AirPods and AirPods Pro with MagSafe to as low as Save up to instantly Black Friday AirPods Pro sale Read more	2021-11-22 19:32:21
Linux	OMG! Ubuntu!	Cutefish OS 0.6 Beta Released with Lock Screen Controls, More Settings	https://www.omgubuntu.co.uk/2021/11/cutefish-os-0-6-beta-released-adds-lock-screen-controls-and-more-settings	Cutefish OS Beta Released with Lock Screen Controls More SettingsA new beta build of CutefishOS is available to download While things are quite daily driver ready yet the distro provides a quick and easy way to try the latest developments in the Qt based Cutefish desktop environment on top of a familiar apt foundation We looked at an Ubuntu based offshoot a few months back In CutefishOS beta the distro ships with a slew of new settings including Network proxy Sound settings Bluetooth settings hour time option Power management There s some new functionality too like being able to drag and drop tray icons to rearrange them You can also transform This post Cutefish OS Beta Released with Lock Screen Controls More Settings is from OMG Ubuntu Do not reproduce elsewhere without permission	2021-11-22 19:11:37
金融	金融庁ホームページ	非常勤職員（専門調査員）を募集しています。	https://www.fsa.go.jp/common/recruit/r3/souri-09/souri-09.html	専門調査員	2021-11-22 20:00:00
ニュース	BBC News - Home	Wisconsin: Police name victims after car ploughs into Waukesha Christmas parade	https://www.bbc.co.uk/news/world-us-canada-59378571?at_medium=RSS&at_campaign=KARANGA	christmas	2021-11-22 19:47:35
ビジネス	ダイヤモンド・オンライン - 新着記事	価格が上昇した「駅近タワマン」ランキング【兵庫県・神戸】2位プラウドタワー神戸県庁前、1位は？ - DIAMONDランキング＆データ	https://diamond.jp/articles/-/287399		2021-11-23 04:55:00
ビジネス	ダイヤモンド・オンライン - 新着記事	一流リーダーが実践する「冒頭30秒」のスキル、会議への関心度をグッと高める！ - トンデモ人事部が会社を壊す	https://diamond.jp/articles/-/288401	集中	2021-11-23 04:50:00
ビジネス	ダイヤモンド・オンライン - 新着記事	【会津高校】華麗なる卒業生人脈！原子力規制委員会初代委員長、渡部恒三、幸楽苑創業者、サンボマスター山口隆… - 日本を動かす名門高校人脈	https://diamond.jp/articles/-/288034	会津若松市	2021-11-23 04:45:00
ビジネス	ダイヤモンド・オンライン - 新着記事	「安倍晋三が担ぐ高市早苗」が安倍派でものすごく嫌われる理由 - ＤＯＬ特別レポート	https://diamond.jp/articles/-/288390		2021-11-23 04:42:00
ビジネス	ダイヤモンド・オンライン - 新着記事	いすゞのEVトラックが自動車産業と日本経済に与える、侮れないインパクト - 今週のキーワード　真壁昭夫	https://diamond.jp/articles/-/288400		2021-11-23 04:40:00
ビジネス	ダイヤモンド・オンライン - 新着記事	「アラフィフ」の処方箋、なぜ50歳が近づくと自分の生き方に疑問がわくのか？ - イライラ・モヤモヤ職場の改善法　榎本博明	https://diamond.jp/articles/-/286219	働き盛り	2021-11-23 04:35:00
ビジネス	ダイヤモンド・オンライン - 新着記事	年金の受け取り方で絶対に注意すべきこと、変更不可の受給申請で大損も！ - 自分だけは損したくない人のための投資心理学	https://diamond.jp/articles/-/288399	開始	2021-11-23 04:30:00
ビジネス	ダイヤモンド・オンライン - 新着記事	「ジェンダーレスファッション」がコロナ禍で人気の理由とは - ニュース3面鏡	https://diamond.jp/articles/-/285723	関係	2021-11-23 04:25:00
ビジネス	ダイヤモンド・オンライン - 新着記事	「ガチな筋トレ」で脳の認知機能が見事にアップするメカニズム - 「脳力アップ」と運動の知られざる関係	https://diamond.jp/articles/-/287173	認知機能	2021-11-23 04:20:00
ビジネス	ダイヤモンド・オンライン - 新着記事	ひろゆきが断言する「30歳までに持っておきたい貯金額」は、いくら？ - 1％の努力	https://diamond.jp/articles/-/287597	youtube	2021-11-23 04:15:00
ビジネス	東洋経済オンライン	｢第3の歴史決議｣で見えた習近平の権力と脆弱性 ｢中国の特色ある社会主義｣とは共産党一党独裁 | 中国･台湾 | 東洋経済オンライン	https://toyokeizai.net/articles/-/470899?utm_source=rss&utm_medium=http&utm_campaign=link_back	一党独裁	2021-11-23 04:30:00