第1章 開拓今后業(yè)務的機器學習

［人工智能現(xiàn)狀與本書概要］

1 什么是機器學習項目

［AI優(yōu)先］

2 了解AI優(yōu)先的時代背景

［GAFA Microsoft的對策］

3 從企業(yè)來看機器學習的策略

［技術進化的意義］

4 機器學習帶來的沖擊·····························································

［機器寒武紀到來的背景］

5 了解機器學習受到關注的原因

［第四次工業(yè)革命］

6 作為國家成長戰(zhàn)略的機器學習···········································

［日本企業(yè)的對策］

7 日本企業(yè)AI對策實況

［所需的人才供求］

8 AI·機器學習所需人才狀況

［從事機器學習的意義］

9 從事機器學習產(chǎn)生新的價值················································

專欄·如何在信息爆炸中獲取正確的信息···················

第2章 理解機器學習的機制

［機器學習概要］

10 什么是機器學習

［基于規(guī)則］

11 基于規(guī)則和機器學習的區(qū)別·················································

［機器學習的優(yōu)點］

12 從機器學習中能得到什么

［機器學習的全貌］

13 理解機器學習的分類····························································

［模型構建的流程］

14 了解機器學習的模型構建

［數(shù)據(jù)的類型和預處理］

15 理解數(shù)據(jù)和預處理································································

［算法的選擇和調(diào)整］

16 了解算法的選擇·······················································

［深度學習］

17 深度學習的基本機制·····························································

［模型驗證和過擬合］

18 評價模型的精度·······································································

［模型的改善］

19 怎樣改善模型········································································

專欄·現(xiàn)在的AI和過去的AI有什么不同

第3章 了解機器學習所必需的資源
［必需的資源］

20 推進機器學習項目所需資源·················································

［物理資源概要］

21 機器學習所需的軟件和硬件

［編程語言］

22 了解Python的特征

［庫］

23 了解機器學習的庫·································································

［統(tǒng)計分析軟件］

24 幫助機器學習的軟件·····························································

［利用云的硬件資源］

25 機器學習所需的硬件資源·····················································

專欄·AI和中國

第4章 確定項目的目標

［項目的全貌］

26 機器學習項目的階段區(qū)分方式·············································

［構思階段］

27 抓住構思階段的全貌·····························································

［課題的設定］

28 什么是機器學習項目的課題··········································

［課題的設定］

29 理解什么樣的課題可以通過機器學習解決·························

［用于機器學習的數(shù)據(jù)種類］

30 理解對課題可用的數(shù)據(jù)··························································

［機器學習系統(tǒng)］

31 理解機器學習系統(tǒng)化的必要性·····································

［課題方案的探討］

32 考慮機器學習項目的候選課題··············································

［課題的篩選］

33 用期望成果和數(shù)據(jù)利用的可能性縮小范圍························

［設計業(yè)務和系統(tǒng)］

34 設計能夠應用機器學習的業(yè)務和系統(tǒng)······································

［日程的研究與制定］

35 制定機器學習項目的日程······················································

［ 執(zhí)行體制的探討］

36 構建機器學習項目的體制······················································

［ROI的估算］

37 估算ROI（投資回報率）

［方案書的寫法］

38 了解有效的方案書的寫法·····················································

專欄·回答什么問題，解決什么課題

第5章 確立項目的體制

［利用外部的合作伙伴］

39 探討向外部合作伙伴企業(yè)的支援請求

［合作伙伴的選擇標準］

40 確定選擇合作伙伴的標準······················································116

［利用分析服務公司］

41 向分析服務公司請求幫助·····················································

［利用咨詢公司］

42 向咨詢公司請求幫助······························································

［活用公司內(nèi)部人才］

43 確保機器學習項目所需的人才··············································

［與其他公司簽約合作］

44 了解合同形式的特征和注意事項·········································

［費用／成本］

45 什么是機器學習系統(tǒng)的費用預算········································

專欄·10年后工作真的會被AI奪走嗎

第6章 驗證項目

46 實現(xiàn)的可能性

［PoC階段的全貌］

47 了解構成PoC階段的任務

［數(shù)據(jù)評估］

48 如何評價用于機器學習的數(shù)據(jù)··············································

［實體模型的構建］

49 構建用于驗證可行性的模型··················································

［使用已訓練好的模型］

50 利用云服務訓練好的模型······················································

［驗證項目的評估］

51 評估PoC階段的驗證項目

［傳感器的驗證］

52 安裝的傳感器來獲取數(shù)據(jù)··············································

專欄·黃瓜農(nóng)戶與深度學習

第7章 實裝機器

53 學習系統(tǒng)

［實裝階段的全貌］

54 了解構成實裝階段的任務·····················································

［實裝的特異性］

55 機器學習系統(tǒng)與一般系統(tǒng)開發(fā)的區(qū)別·································

［需求定義的推進方法］

56 機器學習系統(tǒng)的需求定義······················································

［設計與開發(fā)的推進方法］

57 機器學習系統(tǒng)的設計與開發(fā)··················································

［測試的推進方法］

58 機器學習系統(tǒng)的測試······························································

專欄·創(chuàng)造超人般的AI

第8章 掌握機器學習系統(tǒng)的使用要點

［應用階段的全貌］

59 機器學習項目特有的應用任務··············································

［KPI的監(jiān)測］

60 應該定義怎樣的KPI

［模型的微調(diào)］

61 修正機器學習模型·································································

［系統(tǒng)的應用］

62 應用機器學習系統(tǒng)的課題······················································

專欄·想制作一個打掃整理機器人

第9章 從成功事例中學習機器學習項目

［案例學習①］

63 根據(jù)顧客的行為作出反饋的推薦系統(tǒng)·································

［案例學習②］

64 從SNS的投稿圖像中分析商品的使用場景

［案例學習③］

65 機器人根據(jù)語音請求作出行動··············································

專欄·AI創(chuàng)造的新工作