The AI Index Report

The AI Index is a starting point for informed conversations about the state of artificial intelligence (AI). The report aggregates a diverse set of metrics, and makes the underlying data easily accessible to the general public. 

The 2019 edition tracks three times as many data sets as the 2018 edition. To help navigate the data, we've produced two tools. The Global AI Vibrancy Tool compares 28 countries’ global activities across 34 indicators, including both a cross-country perspective, as well as a country-specific drill down. The AI Index arXiv Monitor helps people conduct their own research into current technological progress in AI. 

The AI Index has worked hard to avoid measurement and evaluation bias. As part of this effort, we convened over 150 industry and academic experts to discuss the many pressing issues that arise from AI data measurement. Workshop Proceedings from the Stanford HAI-AI Index Workshop on Measurement in AI Policy: Opportunities and Challenges will be available shortly. 

The AI Index Report is comprised of nine chapters:

Research and Development

Examines bibliometrics data, including volume of journal, conference and patent publications and their citation impacts by world regions. We also present Github Stars for key AI software libraries, and gender diversity of AI researchers based on arXiv.

Conferences 

Outlines data from a variety of sources on AI conferences. Specifically, we dive into event attendance, summaries of conference topics, and policy milestones achieved. 

Technical Performance 

Tracks technical progress in tasks across Computer Vision (Images, Videos, and Image+Language), Natural Language, potential limitations (Omniglot Challenge), and trends in computational capabilities.

The Economy 

Covers three specific topics: jobs, investment, and corporate activity. We present both global and US-specific data relating to AI jobs, hiring, and skill levels. We also analyze startup investment trends for the world, by country, and by sector. The final section includes data on adoption of AI capabilities in industry and presents global trends in robot installations across countries. 

Education 

Investigates trends in education and AI. This includes analyzing global data in machine learning (ML) and AI training and digging into trends in gender and international diversity for AI PhD’s. We also examine efforts to integrate ethics into computer science curricula and look at global trends in undergraduate enrollment in introductory ML and AI courses. 

Autonomous Systems

Analyzes data around autonomous vehicles (AV’s) and autonomous weapons (AW’s). We highlight countries and cities testing AV’s and present known types of autonomous weapon deployments. 

Public Perception 

Covers public perception of central banks, global governments, and the corporate world. We analyze data on how central banks communicate around AI, investigate AI mentions within the US Congress and Canadian and UK Parliaments, and examine AI-related terms mentioned on US earnings calls. We also dive into US web search data for AI-relevant phrases. 

Societal Considerations 

Examines ethical challenges, global news on AI ethics, and AI applications for sustainable development. We present ethical challenge data by looking across ethical AI guidelines and also examine news coverage around AI’s ethical use. This section also maps AI use cases to the UN’s Sustainable Development Goals. 

National Strategies and Global AI Vibrancy 

The National Strategies metrics looks at official strategy documents issued by countries. The Global AI Vibrancy Tool, as mentioned above, covers 28 countries across 34 metrics grouped into three high-level dimensions of AI starting in 2015: research and development, economy, and inclusion.

轉貼至 : https://reurl.cc/Rd31rx