Simon JOHNSTON leads the Artificial Intelligence and Machine Learning practice at AWS.
Let’s clear up some misconceptions. Artificial intelligence (AI) and machine learning (ML) are often considered beyond the reach of business without data scientists and ML engineers. As explained in Chapter 6 of the AWS eBook Innovate With AI/ML To Transform Your Business, this is no longer the case. AI is being “democratized,” a process that is becoming accessible to people of all levels of experience and skill, creating opportunities for every business.
Here, we discuss two main ways we are seeing the democratization of ML in the real world: accessible education and simplified technology. Next, we look at one of the most effective streams of democratization: reducing the cost of using AI and ML.
AI education for everyone
At a recent SmartCompany/AWS seminar, Simon Johnston, AWS Artificial Intelligence and Machine Learning Practice Lead at ANZ, said: We want to make machine learning accessible to everyone. For AWS and other players, the true benefit of AI is opening doors for everyone, and education is key to achieving that goal.
Education is not years of highly specialized data science training. Rather, it is a foundational study designed to introduce AI and provide foundational skills. As we will see in the next paragraph, AI technology is becoming more accessible, so education need not be as rigorous. One accessible pathway is the AWS Machine Learning University. These courses allow anyone to learn his ML fundamentals, progress at his own pace, and choose the right education for his needs.
At the full entry level, options like Machine Learning Essentials For Business provide guidelines on how ML works within your business, even if you’re just starting to explore ML.Other low barrier tools AWS Some Day Online Conference Builders Online provides some foundation for cloud computing and eases the transition to a more complete knowledge of ML. Wherever you start, education will play an increasingly important role in bringing AI into the mainstream.
Lower barriers to entry
Knowledge of the advanced mathematics, statistics, and programming that underpin ML is still essential, but no longer a prerequisite. Even as education has become more accessible, using AI and ML has become much easier. Chapter 4 of Innovate With AI/ML To Transform Your Business explored use cases powered by ML. It’s capabilities like intelligent document processing and call center automation that let you approach your business without traditional AI expertise. But it goes further.
Tools like Sagemaker Canvas are a no-code ML option. In addition to using the end product (as in the ML-powered use case), you may be able to create custom models and start making ML-driven predictions without writing a single line of code. As Simon Johnston points out, business analysts can complement their deep knowledge with machine learning, inventory planning, churn prediction, revenue optimization, and more. “A business analyst knows his role in the business very well, he knows the data and his use cases, but he doesn’t know machine learning,” he says. “They shouldn’t be prevented from developing these abilities. That’s what Canvas allows.”
Cost effective solution
One aspect of ML democratization that should not be forgotten is its impact on cost. Traditional models in AI and ML rely on data scientists to build and train machine learning models. According to SEEKdata scientists earn an average of about $115,000 per year, making ML an expensive investment, especially when it comes to scaling.
To give an example from the SmartCompany/AWS seminar, Augustinus Nalwan, General Manager of AI, Data Science, and Data Platforms at Carsales, ran into exactly this problem. As businesses focused on his AI to solve problems, the demand for data science skills increased. However, rather than hire Nalwan, he used Amazon Sagemaker to allow other members of the Carsales team to help out at significantly reduced costs. “At Carsales, he can build 70% of his AI models using algorithms from this platform that don’t require a data scientist,” he says. “Anyone with the right practices and guidance from a data scientist can do this job.”
Drive machine learning innovation across industries in Australia and New Zealand, including improving model sophistication, data growth, ML industrialization, ML-powered use cases, responsible AI, and democratizing ML Learn about six major trends.
On-Demand Keynote: look here
Read now: Real-world ways companies can take advantage of ML
