DARPA Director: Automated Reasoning Within Reach
0 Comments Published by John Dillard on Friday, February 23, 2007 at 1:42 PM.
We've had a great response to the post below on why optimizing decision-making is the most important thing an organization can do. In particular, we've had a number of folks ask about decision automation. I read this interview with Tony Tether, Chief of the Defense Advanced Research Projects Agency (DARPA) on Wired's blog today that touches on advanced research related to decision automation. It's pretty far-out science--more thought-provoking than practical, but nonetheless demonstrates the imperative for organizations to focus on this issue. Here's an excerpt:
TT: Since the '90s to now, our ability to create algorithms that can reason -- can more abstractly reason -- about a problem and come up with answers, and also remember what they did using Bayesian techniques and changing values, has really advanced. I mean, it tremendously advanced in the past -- from the '90s to, say, the early 2000s. At the same time, computers became more powerful. We're on the verge of having computers with densities approaching a monkey's brain, and it won't be long before we'll have a computer with the density of transistors, or equivalent to neurons and almost human. What we're missing is the architecture. So it seemed like it was time. We had great advances in algorithms for reasoning and in algorithms that learned in general. At the same time, the computers, the actual intrinsic hardware, was really approaching the density of a human brain. And so it seemed like it was time to try again. We've had some great success. This cognitive program I told you about is actually showing that it is learning, and it is learning in a very difficult environment. This is the program Stanford Research runs for us.
Interesting stuff. We'll post Part II of our thought on the importance of optimizing decisions soon.
TT: Since the '90s to now, our ability to create algorithms that can reason -- can more abstractly reason -- about a problem and come up with answers, and also remember what they did using Bayesian techniques and changing values, has really advanced. I mean, it tremendously advanced in the past -- from the '90s to, say, the early 2000s. At the same time, computers became more powerful. We're on the verge of having computers with densities approaching a monkey's brain, and it won't be long before we'll have a computer with the density of transistors, or equivalent to neurons and almost human. What we're missing is the architecture. So it seemed like it was time. We had great advances in algorithms for reasoning and in algorithms that learned in general. At the same time, the computers, the actual intrinsic hardware, was really approaching the density of a human brain. And so it seemed like it was time to try again. We've had some great success. This cognitive program I told you about is actually showing that it is learning, and it is learning in a very difficult environment. This is the program Stanford Research runs for us.
NS: Which program is this?
TT: It's PAL [Perceptive Assistant that Learns].Interesting stuff. We'll post Part II of our thought on the importance of optimizing decisions soon.
Why optimizing decisions is the most important thing you can do
4 Comments Published by John Dillard on Saturday, February 17, 2007 at 5:38 PM.
The most important piece of advice we can give to organizations and their leadership for the next 30 years is this:
Optimizing decisions is the single most important factor in long-term organizational success. It's more important than strategy, organization design, quality, customer relationship management, innovation, or any other business model, technique, or practice.
That statement is provocative, but it's the reason why we started a company. It also begs the question, "What has changed to make optimized decisions so important?" This post outlines some of the reasons why; the next post will discuss ideas on what to do about it. The reasons are far too many to list here, but below are my views of the key interdependent factors.
1) Business model innovation. Innovations in business models--the underlying mechanisms that define the way organizations operate to provide goods and services--have been changing at breakneck speed in the last 15 years, and there is no reason to expect a coming period of stabilization. Organizations that are successful don't adopt a model and stick with it; they are hyper-adaptive to new business model opportunities when they emerge. The number of choices in business models and the resulting consequences are rapidly multiplying.
2) Intensifying expectations for regulatory compliance. Companies and governments entered a new era after 9/11 and the Enron scandal marked by a dramatic intensification of oversight by shareholders, regulatory authorities, Congress, OMB, and others. Not only is there pressure to make critical decisions quickly and accurately, but organizations must explain to overseers why the decisions were made. This new emphasis on transparency of decision-making is not supported by 20th century decision-making processes.
3) Compressing decision cycles. As business models shift and information becomes more accessible and available, organizations are faced with compressing decision cycles, particularly in critical capability processes. They have less time to choose options, and more options to choose from. In any decision, data must be aggregated, criteria established, options considered, and decisions made. Organizations have less and less time to pass each gate.
4) Advancing decision automation. Advancements in artificial intelligence compound the severity of the decision making problem. More and more decisions may be automated every year, placing additional pressure on manual decisions to either be expedited or automated themselves. Critical decisions will either be severe time traps in critical processes, or the source of substantial competitive advantage. Ignore this technology at your peril--over the next ten years the ability of software to solve complex, unstructured problems will revolutionize what organizations define as their core capabilities. James Taylor writes the best blog out there on decision automation.
5) Accelerating acceleration. As everyone knows, the innovation in technology, business, and life is accelerating. This is well documented---Moore's Law and studies of technology adoption curves are just two good pieces of evidence. However, what places so much more pressure on decision cycles is that the rate of change is also accelerating. Why? Because enablers of innovation are themselves undergoing rapid, logarithmic change. Ray Kurzweil and Alvin and Heidi Toffler have done some great writing on this phenomenon.
These are just five thoughts on my list. . . it's certainly not exhaustive. Our next post will focus on how we view solutions to the challenge--specifically, decision-centric capability development.
Optimizing decisions is the single most important factor in long-term organizational success. It's more important than strategy, organization design, quality, customer relationship management, innovation, or any other business model, technique, or practice.
That statement is provocative, but it's the reason why we started a company. It also begs the question, "What has changed to make optimized decisions so important?" This post outlines some of the reasons why; the next post will discuss ideas on what to do about it. The reasons are far too many to list here, but below are my views of the key interdependent factors.
1) Business model innovation. Innovations in business models--the underlying mechanisms that define the way organizations operate to provide goods and services--have been changing at breakneck speed in the last 15 years, and there is no reason to expect a coming period of stabilization. Organizations that are successful don't adopt a model and stick with it; they are hyper-adaptive to new business model opportunities when they emerge. The number of choices in business models and the resulting consequences are rapidly multiplying.
2) Intensifying expectations for regulatory compliance. Companies and governments entered a new era after 9/11 and the Enron scandal marked by a dramatic intensification of oversight by shareholders, regulatory authorities, Congress, OMB, and others. Not only is there pressure to make critical decisions quickly and accurately, but organizations must explain to overseers why the decisions were made. This new emphasis on transparency of decision-making is not supported by 20th century decision-making processes.
3) Compressing decision cycles. As business models shift and information becomes more accessible and available, organizations are faced with compressing decision cycles, particularly in critical capability processes. They have less time to choose options, and more options to choose from. In any decision, data must be aggregated, criteria established, options considered, and decisions made. Organizations have less and less time to pass each gate.
4) Advancing decision automation. Advancements in artificial intelligence compound the severity of the decision making problem. More and more decisions may be automated every year, placing additional pressure on manual decisions to either be expedited or automated themselves. Critical decisions will either be severe time traps in critical processes, or the source of substantial competitive advantage. Ignore this technology at your peril--over the next ten years the ability of software to solve complex, unstructured problems will revolutionize what organizations define as their core capabilities. James Taylor writes the best blog out there on decision automation.
5) Accelerating acceleration. As everyone knows, the innovation in technology, business, and life is accelerating. This is well documented---Moore's Law and studies of technology adoption curves are just two good pieces of evidence. However, what places so much more pressure on decision cycles is that the rate of change is also accelerating. Why? Because enablers of innovation are themselves undergoing rapid, logarithmic change. Ray Kurzweil and Alvin and Heidi Toffler have done some great writing on this phenomenon.
These are just five thoughts on my list. . . it's certainly not exhaustive. Our next post will focus on how we view solutions to the challenge--specifically, decision-centric capability development.
Labels: artificial intelligence, automation, decision optimization, decisions
First Steps with Identity Management: White Pages
0 Comments Published by Hanno Ekdahl on Friday, February 02, 2007 at 2:34 PM.
Identity Management solutions are very complex and require collaboration across organizational boundaries, IT systems, and networks. Often times, identity data is assumed to be of sufficient quality (see our Blog entry “The Five Elements of Data Quality”) to provision users automatically, define roles, and develop workflows. Unfortunately, this assumption is often false and has severe consequences on any value that you hoped to get from your IdM solution.
One recommendation that we typically make to organizations that are just starting to build their IdM infrastructure is to release a white pages application as their first step. While the value to the organization is often low, so too is the risk if the application fails. This “low value, low risk” approach does offer significant insight, and value, to a small group of concerned individuals: the project team.
One of the nice things about the white pages application is that you can open it up to your end user community for self-service updates. User attributes like Hiring Manager, Department, phone number(s), and job title can all be opened up for editing. End users can update their own information using self-service tools and publish that information to the directory.
From this simple application we get the following benefits:
One recommendation that we typically make to organizations that are just starting to build their IdM infrastructure is to release a white pages application as their first step. While the value to the organization is often low, so too is the risk if the application fails. This “low value, low risk” approach does offer significant insight, and value, to a small group of concerned individuals: the project team.
One of the nice things about the white pages application is that you can open it up to your end user community for self-service updates. User attributes like Hiring Manager, Department, phone number(s), and job title can all be opened up for editing. End users can update their own information using self-service tools and publish that information to the directory.
From this simple application we get the following benefits:
- Visibility into all data stored in the directory
- Ability to troubleshoot the Identity Management infrastructure (connectors, schema, namespace, self-service tools) in a production environment and develop a data remediation plan
- Centralized tool to collect updates for user data
- Validation of user data against standards and business rules before publication to the directory and synchronization with other systems
- Improved data quality across all connected systems
Labels: data quality, directory, edirectory, Identity Management, white pages
