Scan Monthly No. 023January 2005 |
|
<< Other Issues of Scan Monthly |
Download this Scan™ Monthly (PDF) |
New Perspectives on Intelligence | View full summary |
D05-2498 | Download this Insight |
For the past 20 years, while establishing his reputation as a leading designer of consumer computer and telecommunication devices, Jeff Hawkins has closely tracked developments in neuroscience research and initiatives in artificial intelligence (AI). He notes that the lack of a conceptual framework for how the brain operates has hindered progress in neuroscience. Not having a conceptual framework for how human brains work has also resulted in frustration on the part of artificial-intelligence researchers. Hawkins claims that the AI community never managed to agree on what intelligence is or just what the process we call understanding entails. So AI researchers have pursued a variety of options as diverse as expert systems, parallel-processing computing architectures, and neural nets. These options have resulted in only limited success because they replicate or model only small parts of the brain processes active in intelligence. Hawkins claims that a more comprehensive framework will enable the integration of these partial solutions and models into more effective implementations of the technologies already available for creating intelligence. Hawkins believes that his memory-prediction model of intelligence provides both the definition of intelligence and the framework of understanding necessary to advance neuroscience and AI. Hawkins hopes that researchers working with his new memory-prediction framework of human intelligence will be able within ten years to build machine-intelligence capabilities that will far outstrip today's AI implementations and possibly even human capabilities. Hawkins's rethinking of brain processes is ambitious, but it is generating extensive discussion among neuroscientists, psychologists, and researchers working on AI. Whether the memory-prediction framework will serve as a turning point in machine intelligence remains a question, but the framework is generating controversy and original thinking in all three fields. The chances are good that out of such turbulence will come innovative approaches that will reshape how society and businesses approach fairly fundamental business and social processes involving intelligence. Hawkins published in 2004 a book that proposes a new conceptual framework for understanding intelligence. On Intelligence describes Hawkins's framework and explains the reasons why Hawkins believes that researchers' understanding of brain processes has reached a turning point that will rapidly enable machine intelligence in a variety of formats after years of frustration with traditional AI techniques. This study provides an overview of Hawkins's framework and speculates on some of the implications of the theory. Author: Kermit M. Patton. 9 pages. |
|
Learning in Context of Business Processes and Workflows | View full summary |
D05-2499 | Download this Insight |
In recent years, executives and managers responsible for corporate learning and training have come under increasing pressure to demonstrate their business value in terms of business performance. We expect that this change will continue and help build support for changing learning and training operations toward the new model in which learning aligns more closely and integrates with and even embeds within business processes and workflows. The resulting higher degree of proximity of learning to work and workflows holds promise of bringing us closer to a situation in which we can more confidently understand the correlation and causal relationships between learning and business performance. This report, developed by SRI Consulting Business Intelligence's Learning-on-Demand program, examines key characteristics of workflow learning—the term that Jay Cross, a learning guru leading the Workflow Institute, coined and one that we use to represent learning that connects closely and aligns with work tasks specific to job roles and relates to business objectives, processes, and workflows. The report also presents case studies that illustrate how organizations are starting to design and deploy workflow learning systems. The report includes recommendations and action steps for enterprise adopters and vendors as they plan for the emergence of workflow learning. The case-study research indicates that the key benefits that early adopters are experiencing from workflow learning include:
|