Announcement

SBI's Explorer technology area Knowledge-Based Systems became Artificial Intelligence in September 2009. Knowledge-based systems (KBS) was the first artificially intelligent technology to achieve commercial success, and it is likely to remain a key component of commercial artificial intelligence (AI) for the foreseeable future, with innovation in KBS still taking place. But the broad field of AI is evolving fast, and today companies need to pursue knowledge-based systems alongside synergistic and competing AI technologies such as neural networks and statistical reasoning. To create successful products and services, research and development departments need to leverage the AI technique that works for the problem in hand and have a good awareness of how the myriad AI technologies are evolving.

Given the above complexities, I am therefore delighted to bring you a comprehensive view of commercial opportunities in AI in the Explorer format. Clients will still have access to all of Explorer's Knowledge-Based Systems materials through the Web site, and I hope that you will find the new Artificial Intelligence Technology Map both stimulating and valuable.

— Nick Evans, Director of Explorer

Viewpoints

About This Technology

October 2008

As the first major commercialization of artificial-intelligence research, knowledge-based systems embody a paradigm of computing that has a wide array of useful applications. Relying on the interaction between logical procedures and known facts to conduct searches for solutions, KBS technology offers the capacity to capture and retain scarce human expertise, to aid and inform expert decision making, to improve nonexpert decision making, and to solve specific kinds of complex or time-consuming problems more efficiently. Probabilistic systems, including Bayesian networks, a variant of conventional knowledge-based systems, are becoming increasingly common, thanks to improvements in underlying theory and to dramatically improved hardware, which enables more complex calculations. The addition of machine-learning capabilities has also improved knowledge-based systems dramatically and expanded their use in real-time applications in mobile robots.

The technology has already undergone a boom-and-bust cycle, fueled by initially unrealistic expectations and hype. A number of early adopting companies presided over large-scale KBS debacles, most of which occurred precisely because of the companies' overly ambitious faith in the concept of artificial intelligence rather than in the reality of KBS technology. However, another class of users—the companies that implemented the technology on a smaller scale and treated it as just another tool with its own unique assets and limitations—has seen significant benefits. The technology yields competitive advantage within the right set of parameters.

The future appears to be particularly bright for hybrid systems that derive their "expertise" by combining automated extraction of knowledge from data with human experts in specific knowledge domains. Mixed-initiative systems, which combine knowledge-based systems and artificial intelligence with human-computer interfaces, represent a growing synergy of complimentary technologies. In addition, knowledge-based systems combine synergistically with databases, geographic information systems, and the technology that underlies the World Wide Web, providing great opportunities and benefits, but becoming ever more invisible with integration into larger information systems. These hybrid systems will become increasingly popular as the increasingly digital world gives rise to massive amounts of data that require analysis and as people turn to experts to help them deal with greater complexity and uncertainty.