For the newest instalment in our series of interviews asking leading technology specialists about their achievements in their field, we’ve welcomed Mark Perry - Head of Global Business Development at Cloudpick.
Tell us about the business you represent, what is their vision & goals?
In a modern city where the lifestyle is fast-paced, especially at subway stations and office buildings, we see that the service efficiency of stores is often unsatisfactory. At peak hours, customers are frustrated with the length of the checkout queue.
In the summer of 2017, Cloudpick started the development of a smart store solution in a small apartment. The founding team initially had only eight people, but their vision and mission was very clear from the first day.
As experts in computer vision and machine intelligence, the founders of Cloudpick believe that they can use the power of artificial intelligence to create a store that thoroughly understands customers' shopping behaviours, rapidly identifies in-store product placement drifts as well as efficiently provides customers with a checkout-free experience.
Traditional brick-and-mortars are often troubled by increasing operating costs caused by growing labour costs and difficulty in recruiting cashiers as well as retaining experienced store operators. Cloudpick's technology can perfectly free cashiers from tedious and repetitive routine tasks. So our long-term mission is to create solutions that might possibly change the world and improve the quality of human life.
Can you tell me a little about yourself and how you got into the field of artificial intelligence?
After graduating from CASS Business school, with an economics and finance background, I was at first dedicated to the finance arena. Being versatile and interested in various topics, I was always tech-savvy.
I was intrigued by the emergence of great IT companies and products both domestic and abroad, such as Pinduoduo, TikTok, Yelp, Uber, Tesla, etc. I also was into cryptocurrency and smart contracts; I even created my own Bitcoin nodes.
What are the key differences between computer science, machine learning and AI?
Machine learning and artificial intelligence are both subjects of computer science.
People can always find General AI in movies: friendly, like C-3PO in Star Wars; evil, like SkyNet in Terminator. Top-notch AI technology still only exists in movies and science fiction and it is not difficult to understand why. We have not been able to achieve them, at least not yet.
What we can currently achieve is generally called "Narrow AI". Narrow AI is a technology that can perform specific tasks like humans, or even better than humans. Examples include the image/video classification recognition on TikTok or the face recognition setting on Facebook.
They are Narrow AI in practice. How is narrow AI achieved? This brings us to the next level, machine learning. Machine learning is the core of artificial intelligence. It uses algorithms to parse data, learn from them, and then make decisions and predictions about events in the real world.
Artificial Intelligence contains five layers:
The bottom layer is infrastructure construction, including data and computing. The larger the dataset, the stronger the ability of AI.
The upper layer is algorithms, such as convolutional neural networks, LSTM sequence learning, Q-Learning, deep learning and other algorithms, all of which are machine learning algorithms.
The third layer is for important technical directions and issues, such as computer vision, speech engineering, natural language processing, etc. There are other similar decision-making systems, like reinforcement learning, or statistical systems like big data analysis, which can be generated on machine learning algorithms.
The fourth layer is for specific technologies, such as image recognition, speech recognition, machine translation, and so on.
Industry solutions are the top layer. Just like Cloudpick's AI retail solutions, we use AI to maximize the value of stores and retailers. We are working on something nobody has ever done before.
What are some common misconceptions that you believe people have about AI?
Many people believe that AI is omnipotent and hence it is more powerful than human beings. In fact, this is not the case. News reports on Google's Alpha Go defeating South Korean professional Go player Lee Sedol always tend to simply depict the match as a self-taught AI beating humans at their own game.
In fact, however, AI is still far from general intelligence, and at the present stage, it can only solve problems in specific scenarios. We chose to work on AI-driven retail solutions because retail stores normally have limited space and products, so to some extent, this field is less complicated than the autonomous vehicles field.
What advice would you give to people who want to take a career in artificial intelligence?
- Learn the fundamental subjects of mathematics and computer science; have good programming skills. Taking a career in AI without boasting a solid foundation of knowledge is like building castles in the sky.
- Start from one or two subfields of AI, such as Computer Vision (CV), Natural Language Processing (NLP), and then read recent years' top-notch papers and try to write your own.
Would you like to share any artificial intelligence forecasts or predictions of your own with our readers?
I believe that thanks to AI, human beings will have a better and happier life. Many repetitive tasks will be replaced by AI. For example, taxi drivers will be replaced by autonomous vehicles, and you'll find no drivers in future Uber cars. A large number of customer service representatives will be replaced by chatbots.
In general, AI will improve the happiness of all mankind. AI will bring tremendous changes to mankind.
What is your experience of using AI-backed data analysis or log management tools? What do you think are the benefits of using a log management tool that has machine learning capabilities for an organization?
AI-backed data analysis is like the driving indicator warehouse, an advanced monitoring tool of unmanned stores which keeps an eye on the status and the health of the store at all times.
I believe log management tools can be very helpful for organizations because they can help you easily pinpoint the root cause of any application or software error, within a single query.
Are there any books, blogs, or other resources on the subject of AI that you highly recommend?
There are some popular AI courses on Coursera, Udacity as well as on YouTube. I would like to recommend Stanford University's course CS231 and the famous computer scientist and technology entrepreneur Andrew Ng's course series.
If you enjoyed this article then why not check out our previous guide to software deploy tools?