表紙
市場調查報告書

下一波AI技術(第一波:無監督學習) - 邁向真正的智能

Towards Being Truly Intelligent: Next Wave of AI Technologies (Wave 1 - Unsupervised Learning)

出版商 Frost & Sullivan 商品編碼 940867
出版日期 內容資訊 英文 26 Pages
商品交期: 最快1-2個工作天內
價格
下一波AI技術(第一波:無監督學習) - 邁向真正的智能 Towards Being Truly Intelligent: Next Wave of AI Technologies (Wave 1 - Unsupervised Learning)
出版日期: 2020年05月19日內容資訊: 英文 26 Pages
簡介

隨著全球產業追求整體功能的數位化,為了協助決策和生成洞察,越來越多的數據正在生成與利用。隨著數據的量與複雜度增加,對傳統機器學習(ML)演算法而言理解大量變數也變得困難。為這樣龐大且複雜的數據集添加標籤和註解十分費時費力,也令ML無法擴展。

本報告研究AI技術的無監督學習,分析彙整無監督學習介紹、應用、創新者和創新、成長機會等情報。

第1章 摘要整理

  • 研究範圍
  • 研究方法

第2章 無監督學習- 簡介

  • 無監督學習提供了最小人類干預的真正自動化機器學習框架
  • 無監督學習在無法手動標記的大型數據集中也能準確運作
  • 進行資料分群使數據在分析時更適切且易於理解
  • 根據數據集的類型與目的,可使用基於無監督學習的資料分群方法
  • 降維技術在準備大型數據集分析時扮演重要角色
  • 無監督學習系統中的自主性和最小人為干預,在輸出時創造了模糊性

第3章 創新與企業行動

  • 無監督學習將提升自動駕駛車的自主性
  • 無監督學習能夠透過未知語言或口音,幫助NLP系統更輕鬆快速地學習
  • 無前例的獨特金融詐欺,能夠透過無監督學習方法更準確識別
  • 從數據集中識別離群值是UL系統的關鍵強項,使其適合檢測惡意行動
  • 網路安全成為無監督學習的主要創新領域

第4章 成長機會

  • 為了追求自動駕駛車更高的自主性,推動了無監督學習技術的採用
  • 人工智慧系統的精準度,高度仰賴訓練演算法的培訓數據品質
  • 透過產學界的合作可加速無監督學習商業化速度

第5章 產業聯絡資訊

  • 主要聯絡資訊
  • 免責聲明
目錄
Product Code: D985

An Overview On Emerging Machine Learning/Artificial Intelligence Approach

As industries across the globe pursue digitization across functions, more and more data are being generated and utilized to empower decision making and insight generation. As the volume and complexity of data increases, it is also becoming difficult for traditional machine learning (ML) algorithms to make sense of a large number of variables. The labeling and annotation of such large and complex datasets are highly laborious and time consuming, making ML unscalable.

While most of the current ML-based systems depend largely on supervised ML algorithms, unsupervised learning (UL) systems after years of theoretical and lab research have found applicability in commercial applications and have been at the center of many initiatives in industries such as automotive, finance, and cybersecurity.

In brief, this research service covers the following points:

  • Introduction to Unsupervised Learning
  • Applications of Unsupervised Learning
  • Innovators and Innovations
  • Growth Opportunities

Table of Contents

1.0 Executive Summary

  • 1.1. Research Scope
  • 1.2. Research Methodology

2.0 Unsupervised Learning - Introduction

  • 2.1. Unsupervised Learning Lays the Framework for Truly Automated Machine Learning Where Human Intervention Is Minimal
  • 2.2. Unsupervised Learning Works Accurately with Large Datasets Which Cannot be Labeled Manually
  • 2.3. Clustering of Data into Groups Makes Them More Suitable and Understandable for Further Analysis
  • 2.4. A Variety of Data Clustering Methods Based on Unsupervised Learning can be Used Based on the Type of Dataset and the Objectives
  • 2.5. Dimensionality Reduction Techniques Play a Key Role in Prepping up Large Datasets for Analysis
  • 2.6. Autonomy and Minimal Human Intervention in Unsupervised Learning Systems Create Ambiguity in Output

3.0 Innovations and Companies to Action

  • 3.1. Unsupervised Learning Will Empower a Higher Level of Autonomy Among Self-driving Cars
  • 3.2. Unsupervised Learning can Help NLP Systems Learn More Easily and Rapidly with Unknown Languages and Accents
  • 3.3. Unique Financial Frauds with no Precedent can Be More Accurately Identified with Unsupervised Learning Methods
  • 3.4. Identifying Outliers From Datasets Is a Key Strength of UL Systems, Making Them Fit for Detecting Malicious Behavior
  • 3.5. Cybersecurity is Emerging as a Key Area of Innovation for Unsupervised Learning

4.0 Growth Opportunity

  • 4.1. Pursuit of Greater Degree of Autonomy Among Self-driving Cars Is Facilitating the Adoption of Unsupervised Learning Techniques
  • 4.2. The Accuracy of Artificial Intelligence System Is Highly Dependent on the Quality of Training Data Used to Train Algorithms
  • 4.3. Industry-academia Collaborations can Accelerate the Pace of Commercial Adoption of Unsupervised Learning

5.0 Industry Contacts

  • 5.1. Key Contacts
  • Legal Disclaimer