The Four Phases of Machine Intelligence

Chitta said machine intelligence has evolved through four phases: First, business rules were codified via programming logic into business systems. Second, standardized packages like ERP and BPM let business rules be defined on a higher level to codify business flows and operate business applications. The data generated by these systems allowed the emergence of business analytics. Third, some business processes incorporated machine-based decision systems based on modeling and scoring potential decision outcomes for risk, benefit, or anomaly. Integrating these systems requires an intensive effort by data scientists, domain experts, systems architects, and engineers to build systems tailored to each organization.

“For Machine Intelligence 4.0, we have to remove to reduce the dependency of data scientists creating these learning systems,” Chitta said. “The providers are coming with the concept of automatic machine learning, continuously learning and generating multiple learning models and learning agents, and teach each other in an automated fashion.

“The key is to make the whole process autonomous.”

The Impact of Machine Learning 4.0 on Job Roles

Businesses are looking to employ automation to improve remote selling and improve operational efficiencies. Today, we’re experiencing accelerated digitalization and a move to remote work due to the pandemic, Chitta explained. Because most of the software jobs are now remote, there is a need for modern techniques and technologies such as:

• Agile project management
• DevSecOps
• Internet of Things cyber-physical systems integration
• Full stack machine learning in a similar mode to full stack web development

In Machine Intelligence 4.0, intelligent systems will modify or create models and criteria for decision-making. Data scientists, systems architects, and engineers can build adaptable systems that domain experts can shape for a particular industry or business function. End-users then teach the systems to meet their specific business policies and goals.

Here’s the bright chance to create your dream career in AI and ML! Enroll in the AI ML Course with Purdue University, collaborate with IBM today, and take the first step towards your dream career.

Preparing for a Machine Learning Career

Chitta walked through a self-assessment framework for aspiring machine intelligence workers and cautioned candidates not to choose a career for its monetary value because professionals will generally be in that career for 20–30 years. They need to have a passion for mastering a still at a mature level.

Chitta says having a combination of skills is very important. Speaking of his expertise in statistics, he said that alone wouldn’t help in becoming an expert in machine intelligence. One also needs to know a programming language and domain operational knowledge.

Chitta divided the knowledge needed for a career in machine intelligence into two categories: hard skills and soft skills. Hard skills include statistics, data and systems architecture, solution design, and operations. Soft skills include domain functional knowledge, artistic skills, and storytelling. Data scientists, data engineers, machine learning scientists, machine learning engineers, systems developers, and functional administrators could combine these skills in different ways.

Chitta advises those new to the field to gain statistics, probability, and linear algebra fundamentals. Next, they can tackle machine learning tools like Python, PyTorch, and Tensorflow. He said with that foundation, students can experiment with building solutions in a business or problem domain. Other steps include hands-on learning using competition websites such as Kaggle, Google Colab, and other available public infrastructure. He also suggests using Open Data Sources and government data sources to build and test models.

It’s important for students to take online learning courses to acquire and upgrade skills and regularly assess their abilities and aptitudes.

Experienced professionals pursue a different path into machine intelligence depending on their interests and current roles. Chitta suggests that business analysts learn algorithms and programming to become data scientists. Data engineers can gain experience with tools and platforms to become machine learning engineers. Machine learning engineers can study deep learning and MLOps to become machine intelligence Architects.