2017年-亿欧智库_AIIndustrySynthesisReport_64页_2mb

Yiou Intelligence AI Industry Synthesis Report April, 2017 Summary

Core Content

This report provides an overview of the development of artificial intelligence (AI), focusing on its historical evolution, key technologies, and application opportunities in specific industries such as healthcare, finance, and transportation.

Main Viewpoints

• AI Definition: AI is the study of creating machines that can perform tasks requiring human intelligence, such as reasoning, problem-solving, and decision-making. It is a field of knowledge engineering and is closely related to machine learning and deep learning.
• AI Development History: The concept of AI was first proposed in 1956 at the Dartmouth Conference. Early AI development relied on expert systems, which were later replaced by machine learning due to the limitations of computational learning ability.
• Deep Learning: Deep learning is a subfield of machine learning that uses multi-layer neural networks. It has gained prominence due to the availability of big data and high-performance computing, enabling significant breakthroughs in areas like image and speech recognition.
• AI Application Focus: The report suggests that AI will first be applied in healthcare, finance, and transportation due to their high potential for cost reduction and efficiency improvement, as well as their significant impact on daily life and the potential for a "butterfly effect."

Key Technologies in AI

2.1 Overview of AI Technologies

• AI encompasses various technologies such as computer vision, speech recognition, and natural language understanding.
• These technologies are built on a basal layer of big data, computing power, and algorithms.
• The technical layer includes specialized providers and platforms, while the application layer involves real-world use cases.

2.2 Computer Vision Technology

• Definition: Computer vision is the science of enabling machines to "see" and interpret visual information.
• Classification: Includes object recognition, object property recognition, and object behavior identification.
• Recognition Process: Involves training models using sample data and identifying images through signal processing and feature extraction.
• Structure Graph: Shows the integration of computer vision with other technologies like map models and medical imaging.

2.3 Speech Recognition Technology

• Definition: Speech recognition is the process of converting speech signals into text or commands.
• Process: Involves front-end signal processing, feature extraction, and decoding using acoustic and language models.
• Structure Graph: Highlights the three-layer structure (basal, technical, application) of speech recognition systems.

2.4 Natural Language Understanding Technology

• Definition: Natural language understanding involves interpreting and comprehending human language, which is more complex than pattern recognition.
• Applications: Includes search engines and machine translation.
• Structure: Involves corpus processing, model processing, and translation methods, supported by big data and computing power.

Opportunities and Challenges in AI Application

3.1 Why AI Will First Be Applied in Healthcare, Finance, and Transportation

• Healthcare: A field with high demand for cost reduction and efficiency improvement, and it is closely tied to people's lives.
• Finance: A service industry that acts as a lubricant for other industries, with increasing resource allocation and returns.
• Transportation: The lifeblood of cities and a fundamental need for urban life, facing challenges due to urbanization and population growth.

3.2 AI+ Healthcare

• Opportunities: AI can reduce costs and improve efficiency in healthcare, such as through diagnostic tools and patient management systems.
• Challenges: The road to medical AI is still rugged, requiring significant investment and overcoming regulatory and ethical hurdles.

3.3 AI+ Finance

• Opportunities: AI can assist in financial services through robo-advisers, which provide automated investment advice.
• Challenges: Robo-advisers face real-world challenges such as market volatility, regulatory compliance, and user trust.

3.4 AI+ Transportation

• Opportunities: AI can enhance unmanned vehicles, improving safety and efficiency in urban mobility.
• Challenges: Regulatory systems for autonomous vehicles are still under development, and there are concerns about safety, liability, and public acceptance.

Conclusion

The report highlights the transformative potential of AI, especially in the context of big data and computing power advancements. It emphasizes the importance of AI in specific industries, while also acknowledging the challenges in its practical implementation, including lack of interpretability, integration with human knowledge, and regulatory issues. AI is seen as a continuous and progressive technology, capable of driving significant changes in society, but its success depends on overcoming these obstacles.