AI for enterprises and other breakthroughs

Evolution in the natural world occurs on the time scale of millennia, but in the realm of business and enterprise, it occurs within decades.

Fifty years ago, most companies manufactured physical products.

Today, modern firms are integrating hardware with software, and the value propositions of some corporations have transcended physical form altogether, existing solely in the virtual world.

Traditional enterprises that do not acknowledge and adapt to these trends may struggle to survive, said speakers at a series of Enterprise Talks organised by Singapore’s Agency for Science, Technology and Research (A*STAR).

The talks, held on 18 August 2017 at the one-north Festival, featured both A*STAR and industry experts speaking about the importance of embracing technology and innovation, and how this can help companies ride the ongoing wave of change without being submerged.

Enjoy some of the cool highlights from the Enterprise Talk.

Hardware upgrades

As a second-generation entrepreneur leading a business established by his father and uncle in 1964, Mr Eric Lew, Executive Director of local engineering firm Wong Fong Industries, knows how difficult it is to take a traditional business into the 21st century.

“Wong Fong Industries started out as a workshop that repaired vehicles. Today, we are one of the leading land transport engineering companies in Singapore,” said Mr Lew.

“But what concerns me is how we can continue to thrive in the next 50 years, and I believe that innovation is the answer.”

Innovation is required to retain existing customers and capture new markets; hence, existing products must be improved and new products must be developed, Mr Lew commented.

For example, his company has continued to upgrade one of its flagship products — the waste compactor.

“Previously, compactors were purely mechanical products. We have now incorporated technology for the Internet of Things into our modern compactors such that they can send signals to indicate when they are at full capacity,” he said.

Wong Fong Industries is also considering venturing into self-driving utility vehicles, a testament to its willingness to embrace new technologies such as automation, sensors and computer algorithms.

“This is all part of my philosophy to renew, rejuvenate and reinvent a traditional hardware SME,” said Mr Lew.

Robotic software

As automation becomes key for enterprises, new operating systems are needed to command robots and machines to perform complex tasks.

“Software is a key factor for the success of new robotic technologies,” said Ms Chan Min Ling, Programme Manager for the Robot Operating System (ROS)-Industrial Asia Pacific Consortium at the Advanced Remanufacturing Technology Centre (ARTC), A*STAR.

“Hence, ROS-Industrial was set up a few years ago as part of a consortium that caters to industrial robotics applications.”

With ROS, a repository of tools and packages developed by the open source community to control robots, companies can consolidate their hardware and software under a common infrastructure, said Ms Chan. Additionally, if the hardware gets upgraded, the ROS consortium also upgrades the software, thus saving the company time and money.

However, there are concerns over the legal and safety aspects of open-source software: companies want to know who takes credit for innovation and who is liable for safety lapses caused by glitches.

Ms Chan explained that there are a variety of licensing options that allow companies to retain ownership of their projects while tapping on the expertise of the open-source community.

In terms of safety, she noted that ROS can be used in a sandbox, meaning that it is contained by other safety features built into the system.

“ROS will interface with your safety devices, your sensors and your programmable logic controllers to prevent accidents,” she added.

Given the many advantages and possibilities of using ROS, the next generation of automatons may run on community-developed code — a boon for open innovation.

AI = Automation for All?

Although crowdsourcing for software has its benefits, manpower is still required to create, maintain and update operating systems.

But what if an operating system could learn independently, patching its own code to perform specific tasks more competently?

This dream is fast becoming a reality with deep learning and artificial intelligence (AI).

“Writing software is the bottleneck. With AI, we can simply provide data and leave the computer to programme itself,” said Dr Rick Goh, Director of the Computing Science Department of the Institute of High Performance Computing (IHPC), A*STAR.

For example, an AI-based technology developed at A*STAR can now diagnose lung disease with 97 percent accuracy after being trained with chest X-ray images.

If this trend continues, a digital doctor may soon take up residency in clinics and hospitals.

In addition to evaluating human health, AI is also enabling machines to perform self-assessments of their own operational lifespans.

Preventive maintenance can then be performed to drastically reduce operational downtime.

“So instead of questioning ‘why use AI?’, I challenge enterprise owners to think, ‘why not?’ If you are seeking to optimise, enhance or automate part of your business process, you should seriously consider the use of AI,” Dr Goh concluded.