Book Review: "A Brief History of Artificial Intelligence" by Zhang Junping
Recently, one of the most attention-grabbing news items has been the release of Sora by OpenAI, a system capable of generating one-minute long smooth videos from text inputs. This development underscores the shrinking gap between reality and imagination.
Over the past few years, artificial intelligence (AI) has significantly altered the public's perception with innovations like ChatGPT, GPT Turbo, Midjourney, Runway, and now Sora, all rapidly fueling human curiosity about AI's potential.
While recent AI breakthroughs have primarily focused on content generation, large language models, and pattern recognition, AI's applications are diverse. For instance, facial recognition and fingerprint systems in many residential areas, offices, and institutions are everyday examples of AI technology in use.
Personally, I have been involved in AI research since pursuing my Ph.D. in 2000, at a time when AI was still a niche field with low visibility. During my visit to the University of California, Berkeley in 2009, I had the privilege of studying under Professor Zadeh, the father of fuzzy logic and a world-renowned AI expert. This experience deepened my understanding of AI.
The popular science book "A Brief History of Artificial Intelligence" by Professor Zhang Junping from Fudan University perfectly aligns with my expectations. There's a sense of synchronicity, especially in the story about Zadeh pioneering fuzzy logic, which resonated with me.
The book is written in an accessible and engaging manner, devoid of complex formulas as stated on the cover. Coupled with humorous and witty cartoons, it's a stress-free read. Despite being a professional in the field, I believe the book caters to anyone interested in AI, from high school students to Ph.D. candidates and AI practitioners. Here are some key features I observed:
Firstly, the book offers the most comprehensive coverage of AI history among popular science books I've encountered. It starts from Turing's introduction of automata in 1936, marking the beginning of the AI era, and spans to recent developments like OpenAI's ChatGPT in 2023 and the DALL-E3 drawing software based on large language models.
Each story in the book, starting from ancient speculations about AI, feels relatable and down-to-earth, compelling readers to continue. Despite only featuring 69 stories, it covers all significant milestones in the field.
Secondly, each story is followed by the author's reflections, serving as bridges between topics and stimulating reader contemplation. Notably, in the final section, the author introduces his thoughts on the current state and shortcomings of AI development, adding a forward-looking dimension often lacking in historical books. Interested readers might find potential research directions therein.
According to the author's postscript, the book took two years to write and underwent peer review by AI experts, ensuring its rigor and accuracy.
Thirdly, while the book serves as a primer, understanding its concepts fully requires repeated reading and deeper contemplation of AI's development and future.
The author, a prominent figure on ScienceNet, has authored numerous widely-read articles and books such as "Intelligent Beings Love to Make Mistakes" and "High-Quality Research: Teaching You How to Write Papers and Conduct Research." He has also published a significant number of high-quality papers in the AI field and enjoys considerable popularity in science communication, with nearly a million followers across major platforms. Zhang Junping strikes a fine balance between research and popularization.
After reading this book, I pondered on the future of the next generation of AI. The current reliance on data and computational power has limitations such as excessive resource consumption, poor interpretability, and weak robustness, hindering sustainable development.
I believe current AI technology isn't its final form; humans have a knack for innovation and adaptability. The next generation of AI should reflect human knowledge, experience, and wisdom, with a greater emphasis on humanitarian concerns and Chinese experts taking a leading role.
As the Tao Te Ching says, "The Tao that can be told is not the eternal Tao." The path to true AI may forever remain elusive. "Those who know do not speak; those who speak do not know." The exploration of AI is an ongoing journey, and we should find joy in the pursuit.
As a rare gem in AI popular science literature, this book is highly recommended. Lastly, a small suggestion: producing 69 short videos based on the book's stories would enhance accessibility, making it understandable to everyone, not just the wise.
(The author is a professor at Fujian University of Technology.)
Source: China Science Daily, March 22, 2024, Book Review Section, Page 3
