From f40b5e9941ba1f98ab37a172df732cbbe6b5a207 Mon Sep 17 00:00:00 2001 From: howard-haowen Date: Sun, 8 Dec 2024 15:02:23 +0800 Subject: [PATCH] Add the nfu2024 talk --- docs/nfu2024.md | 637 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 637 insertions(+) create mode 100644 docs/nfu2024.md diff --git a/docs/nfu2024.md b/docs/nfu2024.md new file mode 100644 index 0000000..75b5061 --- /dev/null +++ b/docs/nfu2024.md @@ -0,0 +1,637 @@ +--- +marp: true +lang: zh-TW +title: 人工智慧與自然語言 +theme: graph_paper +transition: fade +paginate: true +style: | + section { + font-size: 40px; + } +header: | + 2024-12-09@NFU +footer: | + *語文和工程專業之跨域生涯* +_header: "" +_footer: "" +--- + + + + + +## 人工智慧與自然語言：
語文和工程專業之跨域生涯 + +- 講者：[江豪文](https://bit.ly/4fL1hX0)博士 +- 日期：2024-12-09 + +![bg left](https://lifestylemanagment.com/wp-content/uploads/2023/11/92career.jpg) + + + +--- + +## 大綱 + +- 教育背景 +- 工作經歷 +- 轉職歷程 +- 問答時間 + +![bg right](https://c4.wallpaperflare.com/wallpaper/30/814/497/dark-road-asphalt-wallpaper-preview.jpg) + +--- + + + + + + +# 教育背景 + +![bg](https://manybackgrounds.com/images/hd/education-open-book-blackboard-24neqyjtkzc2rc21.jpg) + +--- + +![](https://github.com/howard-haowen/NLP-demos/raw/main/img/education.png) + +--- + +#### 外文系
做什麼 + +![bg right:70% fit](https://64.media.tumblr.com/tumblr_lzrm2mijKS1qhe5udo1_1280.jpg) + +--- + +#### 語言學
學什麼 + +![bg right:70% fit](https://i.pinimg.com/originals/71/05/f2/7105f2619966da7357f49fd6d82ca925.png) + +--- + +#### 原來語言學
不是學語言 + +![bg right:70% fit](https://miro.medium.com/v2/resize:fit:1024/0*e-Ywdg69I79tfX2r.jpg) + +--- + +#### 我以為我會研究的語言 + +![bg right:70% fit](https://i0.wp.com/starkeycomics.com/wp-content/uploads/2019/11/IE-2-Colours-Corrected.jpg?ssl=1) + +--- + +#### 但我實際上研究的語言 + +![bg right:70% fit](https://preview.redd.it/hand-in-austronesian-languages-v0-n3f8vemugcuc1.png?width=1080&crop=smart&auto=webp&s=15b53cab3e832db8d6c3fb034f1be3c8f673615d) + +--- + + + + + +#### 南島語系的範圍 + +![bg](https://research.sinica.edu.tw/wp-content/uploads/2023/07/chang-yung-li-08-scaled.jpg) + +--- + +#### [台灣南島語](https://zh.wikipedia.org/wiki/%E5%8F%B0%E7%81%A3%E5%8D%97%E5%B3%B6%E8%AA%9E) +Formosan Languages + +![bg right:70% fit](https://eclass2.nttu.edu.tw/sysdata/course/11193/7341.jpg) + +--- + +#### 學位論文 + +- 碩士: [噶瑪蘭語空間認知之研究](http://ntur.lib.ntu.edu.tw/handle/246246/59387) +- 博士: [台灣南島語的名物化與領屬結構](https://repository.rice.edu/items/954266f5-f096-4de5-a00e-44a6b0dfa81a) + +![bg right:65% fit](https://www.memecreator.org/static/images/memes/5469549.jpg) + +--- + + + + + +#### 語言類型學 + +![bg fit](https://www.wolframcloud.com/obj/resourcesystem/images/7ee/7ee344c1-361f-4f1e-8d67-d9bae308f204/66558806191ced22.png) + +--- + +- 名物化論文 1/2 + +Jiang, Haowen. 2021a. Argument nominalization in Formosan languages: A functional-typological approach. Osaka: Osaka University Press. + +![bg right](https://m.media-amazon.com/images/I/31+TciAecsS.jpg) + +[Amazon >>](https://www.amazon.com/%E4%BD%93%E8%A8%80%E5%8C%96%E7%90%86%E8%AB%96%E3%81%A8%E8%A8%80%E8%AA%9E%E5%88%86%E6%9E%90-Nominalization-Theory-Linguistic-Analysis-Japanese-ebook/dp/B094ZSCJRT) + +--- + +- 名物化論文 2/2 + +Jiang, Haowen. 2023b. 16. Nominalization in Formosan Languages. Leiden: Brill + +![bg right fit](https://im1.book.com.tw/image/getImage?i=https://www.books.com.tw/img/F01/a55/16/F01a551648.jpg&v=659bf5ffk&w=375&h=375) + +[Brill >>](https://referenceworks.brill.com/display/db/hflo) + +--- + +- 布農語[線上詞典 >>](https://e-dictionary.ilrdf.org.tw/) + +![h:450 center](https://www.tipp.org.tw/website/thumbs/4487142022.jpg) + +--- + + + + +- 標記式語言[XML >>](https://zh.wikipedia.org/wiki/XML) + +```xml + + + Я тебя люблю. + + Я + I + I + + + тебя + you + you + + + люблю + love + love + + I love you. + +``` + +--- + + + + +![bg fit](https://i.imgur.com/sjwH2Zy.png) +[來源 >>](https://e-dictionary.ilrdf.org.tw/bnn/terms/598210.htm) + +--- + +- 通過布農語認證考試 + [台灣立報 >>](https://www.tipp.org.tw/news_article.asp?F_ID=21367&PageSize=15&Page=3421&startTime=&endTime=&FT_No=&NSubject_No=&SelectSubject=&Subject_No=&SubSubject_No=&TA_No=&Orderby=&KeyWords=&Order=&IsSelect=) + +![bg right:70% fit](https://i.imgur.com/pJHj1Wm.png) + +--- + + + + + + +# 工作經歷 + +![bg](https://a.storyblok.com/f/252228/07dfc8d70a/adult-advice-american-job-interview-work-experience-1024x638.jpg) + +--- + +#### :teacher:英文講師 + +臺北科技大學 +應用英文系 + +![bg right:65% saturate:3.0 90%](https://4.bp.blogspot.com/-lvpwx0NvAHE/WfkiITI5OEI/AAAAAAAAAFE/98Wtc8AuUkAL41mK9U4t-MVW8rSUTSBPwCLcBGAs/s1600/g1377083943860215164.jpg) + +--- + +#### :scientist:博雅博士後研究員 + +北京大學 +中文系 + +![bg right:65% fit](https://i.imgur.com/AzQHqWr.png) + +--- + + + + + + +#### 職涯轉淚點 + +![bg](https://healthmatters.nyp.org/wp-content/uploads/2021/01/new-covid-19-variants-thumbnail.jpg) + +--- + +- 心態 + ![center](https://pbs.twimg.com/media/E_ckPl2VEAAVJ71.jpg) + +--- + +- 自學 + +![center](https://i.pinimg.com/736x/76/ec/68/76ec68eacf540f07ab8339413301f273.jpg) + +--- + +#### :technologist: AI 工程師 + +哈瑪星科技 +創新研發中心 + +- 業界初體驗 + ![bg right:65% fit](https://www.tw-fastener.com/tfsc/images/mm/com/VND_deb0f85a-305e-404e-a354-ec9328f769f2/logo.jpg) + +--- + +#### :technologist: AI 工程師 + +香港商慧科訊業 +AI Lab + +- 跨區合作 + ![bg right:65% fit](https://www.zhangyutong.net/uploads/img/20191219/1576724567695112.png) + +--- + +#### :technologist: 數據分析
襄理 + +新光人壽 +數位資訊部 + +- 資安法規 + ![bg right:65% fit](https://osaas.commerce.nccu.edu.tw/uploads/media/15136543963.jpg) + +--- + +#### :technologist: AI 暨資料解決
方案技術工程師 + +[IBM](https://www.ibm.com/us-en) +科技事業部 + +- 美商文化 + +![bg right:60% 70%](https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2F489ea0af-4ef1-4af3-919e-be486597e278_1080x1080.jpeg) + +--- + +## **watsonx**:
IBM的生成式AI平臺 +[source >>](https://itpromag.com/2023/05/11/ibm-watsonx/) +![bg right:60% fit](https://itpromag.com/wp-content/uploads/2023/05/IBM-Watsonx.jpg) + +--- + + + + + + +# 轉職歷程 + +![bg](https://theblue.ai/wp-content/uploads/2023/06/GenAI-models-Generative-AI-Hamburg.png) + +--- + + + +# 學習進程 + +![bg vertical](https://fakeimg.pl/800x600/9fc5e8/fff/?text=01Python程式語言&font=noto&font_size=60) +![bg](https://fakeimg.pl/800x600/67b8e3/fff/?text=02網路爬蟲&font=noto&font_size=60) +![bg](https://fakeimg.pl/800x600/0288d1/fff/?text=03自然語言處理&font=noto&font_size=60) +![bg](https://fakeimg.pl/800x600/02669d/fff/?text=04機器學習&font=noto&font_size=60) + +--- + +#### Python 是蟒蛇也是語言 + +![](https://fakeimg.pl/800x300/9fc5e8/fff/?text=01Python程式語言&font=noto&font_size=90) + +![bg right fit](https://images-na.ssl-images-amazon.com/images/I/517we1JrQoL._SX331_BO1,204,203,200_.jpg) + +--- + +#### 程式語言相比自然語言的優勢 + +- 沒有人是母語者 +- 不需練聽說，只需讀寫 + +![bg right fit](https://github.com/howard-haowen/blog.ai/raw/master/images/learn-python-from-natives.jpg) + +--- + +#### Python 相比其他程式語言的優勢 + +- 職缺多 + ![bg right:65% fit](https://images.ctfassets.net/aq13lwl6616q/26wpQmcyB0f81krnGj1urY/62291de400aa8e5bcb671ac110e78a80/1.png?w=655&fm=webp) + [source >>](https://zerotomastery.io/blog/best-programming-languages-to-learn/) + +--- + +#### Python 相比其他程式語言的優勢 + +- 待遇高 + ![bg right:65% fit](https://images.ctfassets.net/aq13lwl6616q/afdYSZ1UOsogAbd28RLFN/0a807ec771df0cda3b67f25662a911f3/2024_PROGRAMMING_LANGUAGE_SALARY_BREAKDOWN.png?w=655&fm=webp) + [source >>](https://zerotomastery.io/blog/best-programming-languages-to-learn/) + +--- + +#### Python在今年首次成為最熱門的語言 + +![bg right:65% fit](https://regmedia.co.uk/2024/11/04/github-octoverse-2024-top-languages.jpg) +[source >>](https://www.theregister.com/2024/11/05/python_dethrones_javascript_github/) + +--- + +#### 連小朋友都學得會!!! + +## 真心不騙！！！ + +![bg left fit](https://cdn.kobo.com/book-images/82012983-1b89-46f7-9986-00404214f15f/1200/1200/False/python-for-kids-for-dummies.jpg) + +--- + + + + +![bg 80%](https://www.englishradar.com/wp-content/uploads/2017/03/how-long-learn-English-4-300x378.png) +![bg fit](https://i.ytimg.com/vi/5GYeia8IRbg/sddefault.jpg) + +--- + +#### 軟體工程師的自學比例 + +![bg fit right:70%](https://blog.hyperiondev.com/wp-content/uploads/2017/09/90percent.jpg) +[source >>](https://blog.hyperiondev.com/post/professional-programmer/) + +--- + +#### 高達 80%的人透過線上資源自學 + +![bg fit right:70%](https://cdn.codegym.cc/images/article/84351083-29a8-4761-a5d3-78e4a5ad74ea/800.jpeg) +[source >>](https://codegym.cc/groups/posts/18452-is-becoming-a-successful-self-taught-programmer-realistic-nowadays-yes-weve-decoded-the-formul) + +--- + +#### 爬蟲能快速取得資料 + +![](https://fakeimg.pl/800x300/67b8e3/fff/?text=02網路爬蟲&font=noto&font_size=90) + +![bg right fit](https://blog.ehackify.com/media/2021/07/web.jpeg) + +--- + +#### 網路爬蟲 - 用 Python 解析 HTML + +![h:450 center](https://learn.microsoft.com/en-us/microsoft-edge/devtools-guide-chromium/css/inspect-images/highlighted-styles.png) + +--- + +#### 契機 + +![](https://web.klokah.tw/image/logo.png) +[族語 E 樂園 >>](https://web.klokah.tw/) + +![bg right fit](https://web.klokah.tw/vocabulary/img/02.png) +![bg right fit](https://web.klokah.tw/vocabulary/img/03.png) +![bg right fit](https://web.klokah.tw/vocabulary/img/24.png) + +--- + +#### 爬蟲效果 + +![bg vertical fit](https://i.imgur.com/jsdfwXG.png) +![bg vertical fit](https://i.imgur.com/V0bmAJ8.png) + +--- + +#### 台灣南島語爬蟲成果 + +[ 台灣南島語-華語句庫資料集 APP >>](https://howard-haowen-formosan-languages.streamlit.app/) + +![](https://raw.githubusercontent.com/howard-haowen/Formosan-languages/main/sample-dataframe.png) + +--- + +#### 自然語言處理能透過程式處理大量文字 + +![](https://fakeimg.pl/800x300/0288d1/fff/?text=03自然語言處理&font=noto&font_size=90) + +![bg right fit](https://www.revuze.it/blog/wp-content/uploads/sites/2/2023/01/unnamed-7.png) + +--- + +#### 自然語言處理的定義 + +> Natural language processing (NLP) is a subfield of **linguistics**, computer science, and artificial intelligence concerned with the interactions between computers and human language, in particular how to **program computers to process and analyze large amounts of natural language data**. + +[Wikipedia >>](https://en.wikipedia.org/wiki/Natural_language_processing) + +--- + +#### NLP 的應用 + +![bg right:70% fit](https://www.xenonstack.com/hubfs/nlp-applications.png) + +--- + +#### NLP 的市值 + +![bg right:70% fit](https://artsmart.ai/wp-content/uploads/2024/09/Natural-Language-Processing-Market-1024x621-1.jpg) +[source >>](https://artsmart.ai/blog/natural-language-processing-nlp-statistics-2024/) + +--- + +#### 生成式 AI 的市值 + +![bg right:70% fit](https://market.us/wp-content/uploads/2023/10/Global-Generative-AI-Market.jpg) +[source >>](https://market.us/report/generative-ai-market/) + +--- + +#### 機器學習能透過大量資料學習到模式，並做出決策 + +![](https://fakeimg.pl/800x300/02669d/fff/?text=04機器學習&font=noto&font_size=90) + +![bg right fit](https://d1eipm3vz40hy0.cloudfront.net/images/AMER/deepvsmachineblog/dlvsml4.png) + +--- + +#### 兩類預測模型 + +- 回歸：
預測數值 +- 分類：
預測類別 + +![bg right:65% fit](https://media.licdn.com/dms/image/D5612AQHleCueKC_lww/article-cover_image-shrink_600_2000/0/1677785069046?e=2147483647&v=beta&t=C6GRtT_VWW1c-WkYggBLllLx6Zxor1sSrM9lMW9FGdA) + +--- + +#### 兩類 AI + +- 分辨類：
預測數值或類別 +- 生成類：
產生內容 + +![bg right:65% fit](https://dce0qyjkutl4h.cloudfront.net/wp-content/uploads/2023/06/generative-ai-benefits.png) + +--- + +#### 傳統AI與生成式AI + +![](https://dataplatform.cloud.ibm.com/docs/api/content/wsj/analyze-data/images/fm-overview-diagram.svg?context=wx&locale=en) +[>> source](https://dataplatform.cloud.ibm.com/docs/content/wsj/analyze-data/fm-overview.html?context=wx) + +--- + +#### 生成式 AI 對不同職業的衝擊 + +![bg right:75% 90% fit](https://www.constellationr.com/system/files/uploads/u25646/coursera%20generative%201.png) + +--- + + + +# 學習 4B + +![bg vertical](https://fakeimg.pl/800x600/9fc5e8/fff/?text=01Basics&font=noto&font_size=60) +![bg](https://fakeimg.pl/800x600/67b8e3/fff/?text=02Badges&font=noto&font_size=60) +![bg](https://fakeimg.pl/800x600/0288d1/fff/?text=03Bootstrapping&font=noto&font_size=60) +![bg](https://fakeimg.pl/800x600/02669d/fff/?text=04Blogging&font=noto&font_size=60) + +--- + +#### 利用[geeksforgeeks](https://www.geeksforgeeks.org/python-programming-language-tutorial/)
一步一步練基本功 + +![](https://fakeimg.pl/800x300/9fc5e8/fff/?text=01Basics&font=noto&font_size=100) + +![bg right vertical 90%](https://static.startuptalky.com/2021/06/GeeksforGeeks-StartupTalky.jpg) +![bg right vertical fit](https://media.geeksforgeeks.org/wp-content/uploads/20191023173512/Python-data-structure.jpg) + +--- + +#### 利用[cognitiveclass.ai](https://cognitiveclass.ai/courses/python-for-data-science)
或IBM[SkillsBuild](https://skillsbuild.org/)獲取技術證照 + +![](https://fakeimg.pl/800x300/67b8e3/fff/?text=02Badges&font=noto&font_size=100) + +![bg right vertical 90%](https://sn-portals-cognitiveclass.s3.us-south.cloud-object-storage.appdomain.cloud/an9mrjt7jllzv4u346kr134d19u3) +![bg right vertical fit](https://howard-haowen.github.io/images/cognitive-class-python101-for-data-science.png) + +--- + +#### 利用[GitHub](https://github.com)
借力使力開發應用程式 + +![](https://fakeimg.pl/800x300/0288d1/fff/?text=03Bootstrapping&font=noto&font_size=100) + +- [AI 模型輔助語言學習 APP >>](https://howard-haowen-ailanguageguru.streamlit.app/) + ![bg right vertical 90%](https://miro.medium.com/v2/resize:fit:1125/1*wotzQboYWAfaj-7bvGNIkQ.png) + ![bg right vertical fit](https://docs.streamlit.io/images/streamlit-community-cloud/app-menu.png) + +--- + +#### 利用[GitHub Pages](https://pages.github.com/)
撰寫部落格文章分享所學 + +![](https://fakeimg.pl/800x300/02669d/fff/?text=04Blogging&font=noto&font_size=100) + +![bg right vertical 90%](https://team-coder.com/images/posts/2020-06-14-github-pages-and-jekyll/title-image.jpg) +![bg right vertical fit](https://pages.github.com/images/slideshow/bootstrap.png) + +--- + +#### 學位的
投資報酬率(ROI) + +![bg fit right:70%](https://www.visualcapitalist.com/wp-content/uploads/2024/10/Which_Degrees_Are_Worth_the_Most_SITE.jpg) +[source >>](https://www.visualcapitalist.com/which-college-degrees-have-the-greatest-return-on-investment/) + +--- + +#### 2024 年 10 大AI 職位 + +![bg fit right:70%](https://datasciencedojo.com/wp-content/uploads/10-Highest-Paying-AI-Jobs-in-2024.png) +[source >>](https://datasciencedojo.com/blog/highest-paying-ai-jobs-in-2024/) + +--- + +### 提詞工程 +Prompt Engineering + +[source >>](https://cdn.prod.website-files.com/64412ffddd39557ab2db1cc6/64c8d7f3f93d12246ca2f904_TUeLJwUwp9l4iIWUt1gSRZekzMo-IV3kZ_0s6RPQtSUInN_c9B8CUYqSJhjWoSQAP7pnMaxod5ff32YERglIcYJSz9nnyaOSfZKryUi8H4ZRS1dWtCGlhDDRhGsyfXgIJuhusgLCBWhEKFStPAEV1RA.jpeg) +![bg right:70% fit](https://cdn.prod.website-files.com/64412ffddd39557ab2db1cc6/64c8d7f3f93d12246ca2f904_TUeLJwUwp9l4iIWUt1gSRZekzMo-IV3kZ_0s6RPQtSUInN_c9B8CUYqSJhjWoSQAP7pnMaxod5ff32YERglIcYJSz9nnyaOSfZKryUi8H4ZRS1dWtCGlhDDRhGsyfXgIJuhusgLCBWhEKFStPAEV1RA.jpeg) + +--- + +#### 其實你已經掌握了最強的程式語言 + +![bg fit right:70%](https://i.imgur.com/DuUmNoP.png) + +[source >>](https://santiagof.medium.com/english-is-the-most-powerful-programing-language-even-for-data-science-introduction-to-prompt-998406a499be) + +--- + +#### 9 歲小孩可以
你也可以！ + +![bg right:65% fit](https://cdn.ftvnews.com.tw/engnews/images/2024/cfd1b651-48d8-4e6f-8dfb-49da3f76cc13.jpg) +[source >>](https://english.ftvnews.com.tw/news/2024919W01EA) + +--- + +#### 借力LLM + +- 常見的大語言模型[source >>](https://www.codeandchats.com/2024/04/27/notable-llm-makers.html) + ![bg right:70% fit](https://www.codeandchats.com/assets/images/posts/notable-llm-provides-chart.jpg) + +--- + +### Demo Time! + +- 點擊[這裡](https://colab.research.google.com/)進入
Google Colab + ![bg right:60% fit](https://i.ytimg.com/vi/DjVsnv62i3M/maxresdefault.jpg) + +--- + +## 結語：有產出的學習更有說服力 + +- 獲取技術[證照](https://howard-haowen.github.io/certifications/) +- 撰寫[部落格](https://howard-haowen.github.io/blog.ai/)文章 +- 開發[應用程式](https://howard-haowen.github.io/projects/) + ![bg right:55% fit](https://d2ds8yldqp7gxv.cloudfront.net/Blog+Explanatory+Images/project+deliverables+1.webp) + +--- + + + + + +![bg](https://quotefancy.com/media/wallpaper/3840x2160/860347-David-Brin-Quote-The-best-time-to-act-on-this-was-decades-ago-The.jpg) + +--- + + + + + +# 問答時間:question: + +![bg](https://t3.ftcdn.net/jpg/04/34/94/90/360_F_434949006_MtUycXdKs8P4Qg6ElGkuP9UdsEX012YE.jpg) +![bg fit](https://img-9gag-fun.9cache.com/photo/aV7jvXy_460s.jpg) + +--- + +## 聯絡方式 + +- Email: + `howard.haowen@gmail.com` +- 網站: + [https://howard-haowen.github.io](https://bit.ly/4fL1hX0) + +![bg left:30% fit](https://raw.githubusercontent.com/howard-haowen/blog.ai/master/images/profile-removebg.png) \ No newline at end of file