DATA IS TRANSFORMING
OUR JOB FUNCTION...
The data explosion: welcome to the era of technology innovation, data analytics and the data revolution. Since the rise of 5G, AI, cloud computing and social media networking, the ICT has opened up infinite possibilities to the information technology frameworks. Such as Internet of things (IoT), application programming interface (API), Artificial Narrow Intelligence (ANI), Artificial General Intelligence (AGI), Machine Learning (ML),Data science (DS) and so on.
TECHOLOGY INNOVATION
MAKE THINGS HAPPEN.
AI DEVELOPMENT
1. COLLECT DATA
2. TRAIN THE MODEL
3. DEPLY MODEL
FOR MACHINE LEARNING /
DEEP LEARNING
DEVLEOPING
ARTIFICIAL NARROW INTELLIGENCE
(ANI)
Aspiration
There is a huge difference between Artificial Narrow Intelligence (ANI) and the Artificial General Intelligence (AGI). Whether you are interested to learn how to build / develop an AI for improving your business efficiency or productivity, or even just using it for your own project, this page might be helpful to gain interesting insights about applying AI and other related technologies into your multiple design projects. This page is for the use of taking notes of my learning progress about AI development and programming in python/C++. The courses are run by Coursera which is an online learning platform authorised by the Council of Higher Education Accreditation (CHEA) and accredited by the leading global universities. Hope you will enjoy reading my articles and feel inspired about the million possibilities in design technology for the future creative industries.
Navigating the rise of AI
There is a huge difference between Artificial Narrow Intelligence (ANI) and the Artificial General Intelligence (AGI). It is essential to start with understanding what AI can do and can not do, and find out which AI task will be valuable for your business, and there are ways to improve your company / organisation ability to use AI.
​
To start with, the ANI is commonly develop for speech recognition, face recognition, web search engines, self-driving vehicles, manufacturing line, and even AI farming. For ANI machine learning, you can train the model to do automated leads sorting in order to optimising a sales funnel; or automated visual inspection to improve the efficiency and quality for manufacturing line manager; or even automated resume screening to save time and improve the recruiting standard. Nowadays, many companies especially many big firms applied ANI into digital marketing for online advertisement automation and strategic data acquisition.
​
Whereas, AI is still less likely to deal with emotional approach. A quick example would be asking AI to define human's face emotion or any double meaning sentences while human could tell in second. You can also say that AI is lack of common sense, and it is very hard to train AI to learn how to read minds or common sense since this will require a lot more technical break through. This is also part of the AGI - Artificial General Intelligence, which is to train an AI to think and react like human being. However, it is not commonly use nowadays and there are many controversy against AGI in terms of AI ethical principles.
​
Brainstorming Framework
What are the main driver of business value? What are the main pain points in your business? When it comes to identifying projects for AI development, it is extremely essential to apply your project into the right type of AI knowledge and domain knowledge. The most effective way is to research into due diligence based on your business value before committing to develop an AI project. Separating Technical diligence to analyse what AI can do for the company; and business diligence to determine what kind of AI project would be valuable and feasible to the business itself.
​
For the technical diligence approach, it is necessary to define whether the AI system meet desired performance. Also calculate how much data will be needed and also define an engineering timeline. Whereas for the business diligence, it is also vital to analyse whether the AI project will increase revenue for the current business. Define whether this AI project build to launch new product or new business. It is also an effective way to build spreadsheet financial models to estimate the value quantitatively such as estimate how many dollars / pounds are currently saved, and what would be a reasonable assumption in terms of entires revenue. As you can also model out the economic associated and reduce costs and budget for your proposal. In addition, you can also think about the ethical diligence before the AI project. For instance, will the AI make humanity and the society better.
​
​
​
Open Resources
Machine Learning (ML) projects can be in-house or outsourced, whereas Data Science projects are more commonly in-house. It could be a much more efficient way to search for resources or invest in resources instead of building the already existed program in-house. Of course it will also be genius to learn the basis of AI programming for your project, but start with the open resource / any useful resources that will help improve time-efficiency on research and cost-efficiency on production. There are thousands of open resources available online for free of use or to purchase depends on the application. I have made a list of research publications and open sources in the next article if you also are interested to learn more about AI development and related researches from other aspects written by the professionals and specialists. Of course remember to get the licence and read the terms of use and condition when applying outsourced resources.
THE SMART CITY
WITH AI & 5G+
The development of 5G really has push forward the progression of AI in machine learning, deep learning and data science. The idea of smart cities, smart home, automation, and even self-driving vehicles has becoming a trend in many cities in recent years , and it is interesting to see that AI application will become much more common in the next few years. Many more industries will apply AI technologies into various businesses. AI technology will continue to grow and develop. It will be an essential studies filed in the next 5-10 years time. Therefore, it will be beneficial for us to learn how AI really functions in order to safely develop / program an AI in order to create the ideal smart home/city that is beneficial for human and for the society.
ANI/AGI
TRANSFORMATION
MACHINE LEARNING
DEEP LEARNING
DATA SCIENCE
​
SUPERVISED
LEARNING
Input A -> Output B (Mapping)
The concept of A/B testing is the basic concept to start with AI development. Most of the AI development project that are developed today are based on A/B mapping. (input A and output B) For example if you want to build a speech recognition project then your input A would be the audio clips and the output B would be text transcript. Give an other quick example, let's say you want to build a self-driving car, then input A would be the captured images of the vehicle in the front and radio information, whereas out put B would be the position of other cars in order to enable self-driving by the AI.
In which, you can also specify your acceptance criteria such as a goal for AI accuracy, in order to provide AI team a dataset on which to measure their performance. Also think about Data Training Set and Test Set for A/B testing and validation test in order to find out a reasonable accuracy for the project.
Certainly, there are many more consideration to take before committing to an AI project. Quick examples such as choosing a CPU/GPU to train for a large neural network, even though it is clear that GPU is playing a big role of the rise of deep learning. It will definitely be much easier to get quick results if you invest in a good quality GPU (Graphic Processing Unit) and cloud by your own choice rather than building your own compute servers. However, the processing speed will be much faster for on-prem deployment and running the service locally in your own company. This is also something to keep in mind when building an AI project.
ONLINE COURSES
Coursera
Coursera is an online educational platform where provides various virtual courses in Business, Management, Marketing, Programming, and AI development. Peri has been learning AI and programming from coursera. Which has opened up so many new ideas for Peri's new project. She also enjoy applying all these useful knowledge into design technology innovation. Bringing imagination to life.
Andrew Ng
Andrew Ng is the main instructor of the course "AI for everyone" and other course that is related to AI and programming in the technology field. Andrew is a professor at Stanford University. He was also a former Vice President and Chief Scientist at Baidu, and one of the scientist who founded and directed the Google Brain Deep Learning Project.
