One of the Deadliest Sins of Knowledge Management in Practice: Developing a Shared Context

I have selected the fourth deadly sin of knowledge management, “Not Understanding that a Fundamental Intermediate Purpose of Managing Knowledge is to Create Shared Context” (Fahey & Prusak, 1998, p.268). The evolution of Decision Support Systems (DSS) truly illustrates the importance in developing a shared context across the organization. “Making a good decision starts with having or gathering the right information upon which to base a decision” (Remus & Kottemann, 1986, p. 17). An organization depends upon sound decision-making practices in order to compete. Decision Support Systems (DSS) have evolved to serve as a valuable tool in augmenting the decisional process.

“Since the 1970s, Decision Support Systems (DSS) have been conceived to fill the gap by helping decision makers in solving basically unstructured or semi-structured problems. It has only partially fulfilled its goal because decision-making is basically done through human cognitive process. Pure computer automation processing is inadequate to meet this dynamic process which involves both computer automation processing and cognitive processing” (Lee, 1989, p. 123). The development of business intelligence (BI) and subsequent knowledge management (KM) are not replacements for human cognition; they are available resources for the decision maker in order to generate a more “informed” decision. Given the dominate role of human cognition within BI and KM development, it is absolutely imperative that organizations understand how competing decisional context can lead to poor decision-making. “In the absence of shared context, individuals’ differing perspectives, beliefs, assumptions, and views of the future are most likely to collide and thus immobilize decision making” (Fahey & Prusak, 1998, p. 268).

One solution in attempting to create a universally shared decisional context might be found through the evolution of neural networks. As Stephen Haag relates, there are essentially two types of neural networks, self-organizing and back propagation (Haag, 2004). The primary difference between the two is in how they “learn”. The self-organizing neural network searches for patterns and possible correlations in vast pools of data. “Self-organizing neural networks often form part of data-mining tools for data warehouses” (Haag, 2004, p. 202). Back propagation neural networks are trained, much like a child, where several examples and scenarios are inputted. The key issue lies in, “As the neural network is learning to differentiate between good and bad, the weights (of various strengths) change. The flow of information to the output layer also changes. After you have fed the system enough examples, the weights stabilize, and the neural network then consistently classifies portfolios correctly” (Haag, 2004, p. 202).

A neural network’s reliability can be enhanced via training algorithms that will serve as mechanisms of evolutionary calibration. “A training algorithm, based on statistical sampling theory, is presented, which guarantees that the trained networks provide a designer-specified degree of accuracy in mapping the functional relationship. Within each iteration of this statistical-based algorithm, a sequential design algorithm is used for the design and training of the feedforward network to provide rapid convergence to the network goals. Here, at each sequence a new network is trained to minimize the error of previous network. The proposed method should work for applications wherein an arbitrary large source of training data can be generated” (Maghami, 2000). Essentially a developer would utilize a training algorithm, like the one above, in conjunction with the input of multiple constraints, parameters, and variables. Over time, a neural network’s validity would increase raising the prominence of a shared decisional context.

References

Fahey, L., & Prusak, L. (1998, Spring). The Eleven Deadliest Sins of Knowledge Management. California Management Review, 40(3), 265-276.

Haag, S. Cummings, M. & McCubbery, D. (2004). Management Information Systems for the Information Age (Fourth Edition ed.). New York, NY: McGraw-Hill/Irwin.

Lee, D. T. (1989, September). An Overview of Intelligent Decision Systems. Journal of Information Technology, 4(3), 123-135.

Remus, W. E., & Kottemann, J. E. (1986, December). Toward Intelligent Decision Support Systems: An Artificially Intelligent Statistician. MIS Quarterly, 10(4), 402.