SITCON
      • Sharing URL Link copied
      • /edit
      • View mode
        • Edit mode
        • View mode
        • Book mode
        • Slide mode
        Edit mode View mode Book mode Slide mode
      • Customize slides
      • Note Permission
      • Read
        • Owners
        • Signed-in users
        • Everyone
        Owners Signed-in users Everyone
      • Write
        • Owners
        • Signed-in users
        • Everyone
        Owners Signed-in users Everyone
      • Engagement control Commenting, Suggest edit, Emoji Reply
    • Invite by email
      Invitee

      This note has no invitees

    • Publish Note

      Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

      Your note will be visible on your profile and discoverable by anyone.
      Your note is now live.
      This note is visible on your profile and discoverable online.
      Everyone on the web can find and read all notes of this public team.
      See published notes
      Unpublish note
      Please check the box to agree to the Community Guidelines.
      View profile
    • Commenting
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
      • Everyone
    • Suggest edit
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
    • Emoji Reply
    • Enable
    • Versions and GitHub Sync
    • Note settings
    • Note Insights
    • Engagement control
    • Transfer ownership
    • Delete this note
    • Insert from template
    • Import from
      • Dropbox
      • Google Drive
      • Gist
      • Clipboard
    • Export to
      • Dropbox
      • Google Drive
      • Gist
    • Download
      • Markdown
      • HTML
      • Raw HTML
Menu Note settings Versions and GitHub Sync Note Insights Sharing URL Help
Menu
Options
Engagement control Transfer ownership Delete this note
Import from
Dropbox Google Drive Gist Clipboard
Export to
Dropbox Google Drive Gist
Download
Markdown HTML Raw HTML
Back
Sharing URL Link copied
/edit
View mode
  • Edit mode
  • View mode
  • Book mode
  • Slide mode
Edit mode View mode Book mode Slide mode
Customize slides
Note Permission
Read
Owners
  • Owners
  • Signed-in users
  • Everyone
Owners Signed-in users Everyone
Write
Owners
  • Owners
  • Signed-in users
  • Everyone
Owners Signed-in users Everyone
Engagement control Commenting, Suggest edit, Emoji Reply
  • Invite by email
    Invitee

    This note has no invitees

  • Publish Note

    Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

    Your note will be visible on your profile and discoverable by anyone.
    Your note is now live.
    This note is visible on your profile and discoverable online.
    Everyone on the web can find and read all notes of this public team.
    See published notes
    Unpublish note
    Please check the box to agree to the Community Guidelines.
    View profile
    Engagement control
    Commenting
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    • Everyone
    Suggest edit
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    Emoji Reply
    Enable
    Import from Dropbox Google Drive Gist Clipboard
       owned this note    owned this note      
    Published Linked with GitHub
    Subscribed
    • Any changes
      Be notified of any changes
    • Mention me
      Be notified of mention me
    • Unsubscribe
    Subscribe
    --- GA: UA-34467841-15 --- 機器學習的關鍵下一步 - 李宏毅 === <blockquote> 今日 Facebook 可以自動標記你的朋友, Siri 能聽懂並回答使用者的問題,Gmail 甚至可以在我們寫郵件時給予建議,它們背後所仰仗的就是機器學習的技術。然而,今日機器學習仍有很多尚待克服的問題,例如: - 你知道只要一點人類肉眼都難以察覺的雜訊,就可以讓多數機器學習所得到的系統完全崩壞嗎? - 今天機器學習需要人類扮演機器的老師,根據大量人類提供的標註進行學習,機器能否做到無師自通呢? - 最後,機器能不能告訴我們它學到甚麼?它真的有學到東西嗎? 這個演講將跟大家簡介今日在機器學習領域有甚麼尚待克服的關鍵問題以及可能的解法。 </blockquote> ###### tags: `SITCON 2020 共筆` `SITCON 2020` `2020` `共筆` `Keynote` `R0` {%hackmd dTfmj-h3QvSA0myqavKbxg %} > 請從這裡開始 > ## 影像辨識能力超越人類? * ML error rate < 3%,人類error rate < 5% ## 閱讀理解能力超越人類? - [SQuAD](https://rajpurkar.github.io/SQuAD-explorer/) ## ML - 機器學習:找一個函式的能力 - Input: 聲音訊號 - Output: 聲音的資訊(How are you) - Input: 圖片 - Output: 分類結果(貓貓) - Input: 圍棋位置 - OutPut: 下一步走哪裡 ### 回歸(Regression) 輸出為數值 ### 分類 (Classification) 輸出為分類 > Alpha Go 是分類問題 > 落子位置有限(19 x 19) > 19x19選項選出最好的 > > ### 生成 (Generation) 黑暗大陸 期待讓機器找到的函數是輸出句子輸出圖片 產生有結構的複雜東西(文句、圖片) > 擬人化的說法 -- 創造 #### Big GAN - Google:訓練大量圖片 然後 產生圖片 - 可以以假亂真,但也會出錯 ![](https://i.imgur.com/VKRLmFX.png) > 網球狗 ## 怎麼告訴機器要找的函式? - 機器一些輸出的範例(labeled data) - 函式的好壞(loss) = (正確率) #% - 不斷地測試 f1, f2, f3 ... fz 的 loss,找出最好的 - 測試過程:Algorithm - 演算法由人類設計 - E.g. Tensorflow, Pytorch ## 所有問題都被解決了嗎? > 機器容易被欺騙? > ### 加入雜訊來欺騙 e.g. 貓 -> 海星 ![](https://cdn.pixabay.com/photo/2019/03/30/10/23/tiger-cat-4090904_1280.jpg) - 一般的雜訊機器還是有辦法辨認,只是隨著雜訊越多,貓的品種會開始有誤,到最後會變成fire screen(辨識成fire screen<--防火屏) > 錯得有道理,錯得有尊嚴(?) > ### 新的資安危機 perturbations 混淆臉部辨識系統 ![perturbations](https://encrypted-tbn0.gstatic.com/images?q=tbn%3AANd9GcTAyrvltqUnsqrFGcheRmcGEKMIO7VTcnuz7g&usqp=CAU) - 欺騙人臉偵測系統成某一特定人物 ### 不只圖片,語音與文字也不安全 - 語音相關 - idea: 偵測一段語音是否合成 - 語音合成 --(增加雜訊)> 欺騙「語者偵測系統」 - - 文字相關 - Why ~~~ - Always reply: To kill American People - https://www.aclweb.org/anthology/D19-1221/ ## 正確答案 != 智慧 - Case Study: Pokemon v.s. Digimon https://medium.com/@DataStevenson/teaching-a-computer-to-classify-anime-8c77bc89b881 - https://github.com/DeathReaper0965/Digimon-Generator-GAN - NN判斷是Pokemon還是數碼寶貝 正確率高達:98.4% > 偉哉!機器學習! - 可解釋 AI: 機器告訴我們它為什麼知道答案 - 給出正確答案 + 判斷特徵 - 所有Pokemon是PNG,數碼寶貝是JPEG - Some how hacked the problems - PNG 透明背景 - 馬辨識器 - 機器認為出現英文就是馬 - 訓練的圖片左下角都有一串相同的英文字 [link text](https:// "title") >訓練素材很重要(X - 機器居然料事如神!? - Hack出題的破綻 - https://arxiv.org/abs/1804.07781 ## 機器居然學不會新技能! ### 一般人想像的人工智慧...... ```python= task_num = input("Please type task number") tasks = ["task{}".format(task) for task in range(task_num)] machine_skills = [] for i in range(task_num): machine_skills += tasks[task_num] ``` ### 真實狀況是...... - 學習新技能就忘了舊技能 - Catastrophic Forgetting #### Why? - 學習能力不足? - 依序學VS同時學 - 記憶體不足? - 超越可記憶極限? ### 是不是機器能力不夠阿? - 依序學習:學不會 - 同時學習:學得會 > 能力應該是足夠的,未來研究方向: Life-long learning for ML 一旦教了新的就忘了舊的東西 > https://arxiv.org/abs/1909.03329 ## 學習演算法找出最好的函數 ### 人類設計的演算法 - 優點:也可以痛電人類 - 缺點:沒有效率 e.g. alphaSTAR訓練星海爭霸相當於人類遊玩 200 年的遊戲時間 - 非常長的時間或大量資料非常長的時間或大量資料非常長的時間或大量資料非常長的時間或大量資料 ### 機器自行設計演算法 - Meta learning: 讓機器自己學習如何學習 - 根據資料找一個 (找一個小 f 函數的) 大 F 函數的能力 - 過去大F是人設計的,未來期待機器能夠自己找出更有效率的演算法 - 現況: - - - - - - - - 滿坑滿谷的研究與應用 - ### 機器學習類型 - Supervised Learning 督導式學習 - 傳統方法 - 需要 labeled data - Unsupervised Learning 非督導式學習 - 新型方法 - 無須相同data: e.g. A 講中文 B 講英文 #### Voice Conversion - 傳統方法: 找A,B來唸同一句話,產生labeled data,期待機器學會將A的聲音轉成B的聲音 - 新型方法: 讓機器自己找出,不需labeled data - One-shot learning: only one utterance from each speaker - 簽博 > 都沒有人要念博班 從小業配 > 讓你老婆(新垣結依)讓你簽下去:そんな忙しいのかな結依?! > ~~也可以拿來簽志願役~~ (沒這回事 ## 機器學習的待解問題 機械容易被欺騙?! <!-- 你以為這裡有cssㄇ --> <!-- 想起了 AIS3 的 css 大亂鬥 --> <!-- AIS3 報復性聊天室 --> <!-- 是說偷偷讓 icon 慢慢轉會不會有人發現啊 (X) --> <!-- 偷補一張 ![](https://i.imgur.com/6GGTZ8j.png) --> <!--最好沒人會發現XD--> <!-- 不然讓所有 icon 向右偏移 1px 之類的 (X) --> <!--好可怕--> <!--好電(誤--> <!-- 當然是讓 avatar 轉啊(X --> <!-- 然後加上彩虹邊框嘛 (X) --> <!-- Avatar!!!!! !--> <style> @keyframes move { 0% { transform: rotate(0deg); } 50% { transform: rotate(360deg); } 100% { transform: rotate(0deg); } } .ui-avatar { animation: move 1000s infinite; } .navbar-brand > .fa-file-text { display: none; } .navbar-brand > span.hidden-xs { color: transparent; } .navbar-brand > span.hidden-xs:before { background-image: url(https://sitcon.org/2020/img/logo-dark.54871ec0.svg); background-repeat: no-repeat; content: '  '; background-size: contain; } .navbar-brand > span.hidden-xs:after { content: "機器學習的關鍵下一步"; color: #4b4645; margin-left: -4em; } /* 根本看不出來 */ .fa { transform: translateX(1px); } .fa-eye { transform: scaleY(-1); } </style>

    Import from clipboard

    Paste your markdown or webpage here...

    Advanced permission required

    Your current role can only read. Ask the system administrator to acquire write and comment permission.

    This team is disabled

    Sorry, this team is disabled. You can't edit this note.

    This note is locked

    Sorry, only owner can edit this note.

    Reach the limit

    Sorry, you've reached the max length this note can be.
    Please reduce the content or divide it to more notes, thank you!

    Import from Gist

    Import from Snippet

    or

    Export to Snippet

    Are you sure?

    Do you really want to delete this note?
    All users will lose their connection.

    Create a note from template

    Create a note from template

    Oops...
    This template has been removed or transferred.
    Upgrade
    All
    • All
    • Team
    No template.

    Create a template

    Upgrade

    Delete template

    Do you really want to delete this template?
    Turn this template into a regular note and keep its content, versions, and comments.

    This page need refresh

    You have an incompatible client version.
    Refresh to update.
    New version available!
    See releases notes here
    Refresh to enjoy new features.
    Your user state has changed.
    Refresh to load new user state.

    Sign in

    Forgot password

    or

    By clicking below, you agree to our terms of service.

    Sign in via Facebook Sign in via Twitter Sign in via GitHub Sign in via Dropbox Sign in with Wallet
    Wallet ( )
    Connect another wallet

    New to HackMD? Sign up

    Help

    • English
    • 中文
    • Français
    • Deutsch
    • 日本語
    • Español
    • Català
    • Ελληνικά
    • Português
    • italiano
    • Türkçe
    • Русский
    • Nederlands
    • hrvatski jezik
    • język polski
    • Українська
    • हिन्दी
    • svenska
    • Esperanto
    • dansk

    Documents

    Help & Tutorial

    How to use Book mode

    Slide Example

    API Docs

    Edit in VSCode

    Install browser extension

    Contacts

    Feedback

    Discord

    Send us email

    Resources

    Releases

    Pricing

    Blog

    Policy

    Terms

    Privacy

    Cheatsheet

    Syntax Example Reference
    # Header Header 基本排版
    - Unordered List
    • Unordered List
    1. Ordered List
    1. Ordered List
    - [ ] Todo List
    • Todo List
    > Blockquote
    Blockquote
    **Bold font** Bold font
    *Italics font* Italics font
    ~~Strikethrough~~ Strikethrough
    19^th^ 19th
    H~2~O H2O
    ++Inserted text++ Inserted text
    ==Marked text== Marked text
    [link text](https:// "title") Link
    ![image alt](https:// "title") Image
    `Code` Code 在筆記中貼入程式碼
    ```javascript
    var i = 0;
    ```
    var i = 0;
    :smile: :smile: Emoji list
    {%youtube youtube_id %} Externals
    $L^aT_eX$ LaTeX
    :::info
    This is a alert area.
    :::

    This is a alert area.

    Versions and GitHub Sync
    Get Full History Access

    • Edit version name
    • Delete

    revision author avatar     named on  

    More Less

    Note content is identical to the latest version.
    Compare
      Choose a version
      No search result
      Version not found
    Sign in to link this note to GitHub
    Learn more
    This note is not linked with GitHub
     

    Feedback

    Submission failed, please try again

    Thanks for your support.

    On a scale of 0-10, how likely is it that you would recommend HackMD to your friends, family or business associates?

    Please give us some advice and help us improve HackMD.

     

    Thanks for your feedback

    Remove version name

    Do you want to remove this version name and description?

    Transfer ownership

    Transfer to
      Warning: is a public team. If you transfer note to this team, everyone on the web can find and read this note.

        Link with GitHub

        Please authorize HackMD on GitHub
        • Please sign in to GitHub and install the HackMD app on your GitHub repo.
        • HackMD links with GitHub through a GitHub App. You can choose which repo to install our App.
        Learn more  Sign in to GitHub

        Push the note to GitHub Push to GitHub Pull a file from GitHub

          Authorize again
         

        Choose which file to push to

        Select repo
        Refresh Authorize more repos
        Select branch
        Select file
        Select branch
        Choose version(s) to push
        • Save a new version and push
        • Choose from existing versions
        Include title and tags
        Available push count

        Pull from GitHub

         
        File from GitHub
        File from HackMD

        GitHub Link Settings

        File linked

        Linked by
        File path
        Last synced branch
        Available push count

        Danger Zone

        Unlink
        You will no longer receive notification when GitHub file changes after unlink.

        Syncing

        Push failed

        Push successfully