Decision Intellig...
Follow
3.6K views | +1 today
Scooped by Lorien Pratt
onto Decision Intelligence
Scoop.it!

AI / Machine Learning / Neural Networks: what's missing? Lorien Pratt speaks at CMU Silicon Valley

Lorien Pratt's insight:

I've spent the last 18 years since I left my career as a college professor answering a single question: "Why is this incredibly powerful AI/ML/neural network technology not more widely used?"  It's been restricted to just a few use cases, and we're still not using data and objective evidence in thousands of situations across countries and companies around the world.


This talk is the result.  My life's work.  I hope you enjoy it.


(Hint: think complexity)

more...
No comment yet.
Decision Intelligence
Decision intelligence is a framework to apply an interdisciplinary set of technologies to solve the world's most complex problems in an increasingly nonlinear and rapidly changing world. Technology includes predictive intelligence, visual decision modeling, complex systems modeling, big data, predictive analytics, machine learning. UX design, statistical analysis, business intelligence, business process management, causal reasoning, evidence-based analysis, and more. For an overview, see the webinar at http://youtu.be/XRTJt3bVCaEand more videos at http://www.youtube.com/quantellia. Many of these topics are vigorously discussed in the LinkedIn group Effective Decision Making in the Midst of Complexity: http://www.linkedin.com/groups?gid=205078.  Also see http://www.quantellia.com
Curated by Lorien Pratt
Your new post is loading...
Scooped by Lorien Pratt
Scoop.it!

How the Exec Team Gets Better at Decision Making

Join TDI3 for an informative discussion about how the work done behind the conference room door is evolving, how it needs to evolve, and how “unintended consequences’ can be contained.

Lorien Pratt's insight:

We've substantially improved many aspects of corporate life, except decision making at the executive level, which is done the same as it was 100 years ago.  Seems there might be room for improvement.  Especially as we seem to see an increasing disease of unintended consequences.

more...
Lorien Pratt's curator insight, July 17, 1:31 PM

Better executive decisions => reduced unintended consequences.

Rescooped by Lorien Pratt from Unintended Consequences
Scoop.it!

Change one thing. Change everything.

Change one thing.  Change everything. | Decision Intelligence | Scoop.it

As human beings, we cannot change the past. We have control only of our individual destiny and the decisions that we make today. The tagline for the movie The Butterfly Effect is, “Change one thing, Change everything.”

Lorien Pratt's insight:

An analysis through the lens of unintended consequences of the situation with the U.S. border children.

more...
Lorien Pratt's curator insight, July 22, 12:54 PM

An analysis of the situation with the U.S. border children through the lens of unintended consequences.

Scooped by Lorien Pratt
Scoop.it!

Mental Model: Complex Adaptive Systems

Mental Model: Complex Adaptive Systems | Decision Intelligence | Scoop.it
Let's explore the concept of complex adaptive systems.
Lorien Pratt's insight:

Probably the clearest explanation I've read of the distinction between complicated, complex, and adaptive.  Well worth a read

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

The wicked problems of the world share something in common

The wicked problems of the world share something in common | Decision Intelligence | Scoop.it

As many of my colleagues and online acquaintances may know, my professional work has largely been grounded in the application of systems theory, to analyze and solve problems calling for a systems approach"

Lorien Pratt's insight:

Barry's discovered what I call a "systems pattern":


"As I assay it, the ten most intractable plagues of western civilization are conflict, violence, oppression, injustice, corruption, poverty, ignorance, alienation, suffering, and terrorism.


All ten of these hellish problems have something in common. Like cancer, they tend to reseed themselves, round-robin, from one instance to the next, in a never-ending cycle of recursion.


Systemic problems call for a systems approach to problem-solving. But that’s not going to happen until we elevate our collective problem-solving skills to near-genius levels."

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Patterns for Systemic Change: Commons Abundance Network

Patterns for Systemic Change: Commons Abundance Network | Decision Intelligence | Scoop.it
Details of: Exploring how to embed into the system the code for its renewal. Fostering factors of opportunity and renewal as commons. Leveraging agency and empowering change agents in their own contexts across the board.
Lorien Pratt's insight:

This looks pretty interesting.  The idea of a community-based approach to implementing systematic change is relatively new to me; this group looks pretty informed on some of the complex systems dynamics.  I'll join, and read,and see what it's about.

 

Some bits that are particularly tantalizing:

"Our System is stuck in self-reinforcing feedback loops...the interactions between the multiple human and natural systems that make up our world, and the flows and accumulations they generate."

 

"We end up with resources attracted, captured and extracted by a few 'winners', cumulating advantage (opportunity, success, power, credit, capital) in a reinforcing feedback loop, dedicated to reinforce the mechanisms that maintain the status quo..."

 

"We need to embed the reproduction of distributed factors of opportunity & renewal in the code of the system at all levels and scales for the system to become generative rather than extractive."

 

I bet World Modeler could help.  I'd appreciate an introduction if anyone here knows folks in this group.

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Ignorance Is Bliss: Study Shows People Avoid Information About Complex Social Issues

Ignorance Is Bliss: Study Shows People Avoid Information About Complex Social Issues | Decision Intelligence | Scoop.it
Lorien Pratt's insight:

I've long talked about the "complexity ceiling" in my work, and now here's a study that illustrates it.  The actionable item here, is that when communicating complex social issues, we need to break them down into understandable pieces.  (Yet, if we stop there, we'll miss key dynamics: there must be a reduction/synthesis cycle).


From the article:

The participants who received the complex description indicated higher levels of perceived helplessness in getting through the economic downturn, more dependence on and trust in the government to manage the economy, and less desire to learn more about the issue.

 

“This is despite the fact that, all else equal, one should have less trust in someone to effectively manage something that is more complex”

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Autocatalysis: the mother of all tipping points

Lorien Pratt's insight:

The best analysis I've seen on why things are changing so much faster.  Well, put, Reese, though I had to look up Autocatalysis (http://en.wikipedia.org/wiki/Autocatalysis).  You're right.

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Red Lobster: Get out your bibs, it's getting messy

Red Lobster: Get out your bibs, it's getting messy | Decision Intelligence | Scoop.it

This just screams "DECISION MODEL NEEDED" at me - how about you?

Lorien Pratt's insight:

How decision modeling can be used to clarify different assumptions and systems understanding in a conflict.  I think that the analysis shown here can apply to many negotiation situations as well.

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Unintended consequences: Research reveals concerns over EU ban on discarding fish

Unintended consequences: Research reveals concerns over EU ban on discarding fish | Decision Intelligence | Scoop.it
New EU rules that ban fishermen from throwing away unwanted fish they have caught could actually harm wildlife and fail to improve fish stocks, according to a new report.
Lorien Pratt's insight:

Decision: Ban fisherman from throwing away unwanted fish

Intended consequence: Improve fish stocks

Unintended consequence: Decrease fish stocks

Analysis: A key link in the chain was not understood: discarded waste from humans is valuable in the ecosystem

more...
Lorien Pratt's curator insight, May 14, 8:49 PM

Decision: Ban fisherman from throwing away unwanted fish

Intended consequence: Improve fish stocks

Unintended consequence: Decrease fish stocks

Analysis: A key link in the chain was not understood: discarded waste from humans is valuable in the ecosystem

Scooped by Lorien Pratt
Scoop.it!

Murders and Mozarella: Things that correlate

Murders and Mozarella: Things that correlate | Decision Intelligence | Scoop.it

"Why do these things correlate? These 20 correlations will blow your mind."

Lorien Pratt's insight:

To construct an effective decision model, which shows how actions lead to outcomes, we need to be sure to get our lines of causation right.  And in a world of Big Data, there are an awful lot of things that can appear to cause one another, but do not.  Such as per capita consumption of mozzarella cheese / industrial manufacturing doctorates.  Divorce rate in Mississippi / murders by bodily force.


Now, these may appear to be obviously ridiculous causal links, but there are many situations where we don't all share a common systems model that would rule these out.  So precipitation in Indiana may indeed cause more people to drown after falling out of a canoe or kayak.  Or maybe not. 


The graphs on this site help to illustrate the critical importance of a) making our causal assumptions explicit so that b) we can QA them (either through a systems modeling exercise or through the right randomized controlled trial (RCT) design).  Otherwise we may end up reducing California slip/tripping deaths by awarding more mechanical engineering doctorates.  Probably not a good idea :-)


Be sure to try the "discover a new correlation" link at the top of the page.





more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Decision Intelligence for Credit Risk Management

Decision Intelligence for Credit Risk Management | Decision Intelligence | Scoop.it
Lorien Pratt's insight:

A case study for creating risk models and rapidly deploying them into production.

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Transcendence and Analytics: What If the End of Mankind as We Know It Were a Good Thing?

Transcendence and Analytics: What If the End of Mankind as We Know It Were a Good Thing? | Decision Intelligence | Scoop.it
I saw the movie Transcendence last week, and it was an impressive telling of what might happen were we to achieve singularity. Given this is likely where AI research is taking us and that analytics is a major component to getting there, I not only recommend the movie as entertainment but also as a means to wrapping your head around what we must correct before we take that next big step.
Lorien Pratt's insight:

Yes, analytics can save the world.  Working on that.

more...
No comment yet.
Rescooped by Lorien Pratt from Unintended Consequences
Scoop.it!

Do We Need Data Scientists, or Do We Need Decision Scientists?

Do We Need Data Scientists, or Do We Need Decision Scientists? | Decision Intelligence | Scoop.it

If you're thinking something along these lines: "We have all this data, and we need data scientists to tell us what we can do with it," you're at the wrong end of the trail - and you're walking backwards. Blindfolded.

Lorien Pratt's insight:

In my experience, often when people say that they need a "data scientist," they really need a decision scientist.  This article contains the six key steps that decision scientists follow.

more...
Lorien Pratt's curator insight, May 1, 4:19 PM

In my experience, often when people say that they need a "data scientist", they really need a decision scientist . This article contains the six key steps that decision scientists follow

Rescooped by Lorien Pratt from Non-Equilibrium Social Science
Scoop.it!

Decision Making in a Complex and Uncertain World

Decision Making in a Complex and Uncertain World | Decision Intelligence | Scoop.it
This course will teach you the first principles of complexity, uncertainty and how to make decisions in a complex world.

Via Jorge Louçã, NESS
Lorien Pratt's insight:

This course looks awesome!  I've signed up. See you there?

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Tree roots form a sort of "neural network" with one another, in an antifragile complex adaptive system

Tree roots form a sort of "neural network" with one another, in an antifragile complex adaptive system | Decision Intelligence | Scoop.it

"In this real-life model of forest resilience and regeneration, Professor Suzanne Simard shows that all trees in a forest ecosystem are interconnected, with the largest, oldest, "mother trees" serving as hubs. The underground exchange of nutrients increases the survival of younger trees linked into the network of old trees. Amazingly, we find that in a forest, 1+1 equals more than 2"

Lorien Pratt's insight:

This blows me away; I had no idea about these interconnected, synapse-like links from one tree to the next. 

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Collaborate.org: Global platform for GIS, collaboration, videos, and more

Collaborate.org: Global platform for GIS, collaboration, videos, and more | Decision Intelligence | Scoop.it
Lorien Pratt's insight:

This looks like a good platform on which to build Decision Intelligence: combining collaboration, GIS, videos, and more.

more...
No comment yet.
Suggested by Julien Rio
Scoop.it!

Size doesn't matter: can SMEs conquer Big Data?

Size doesn't matter: can SMEs conquer Big Data? | Decision Intelligence | Scoop.it
Big Data involves collecting a large volume of information. As SMEs don't have access to the same amount of data as larger corporations, does it mean SMEs can't make use of Big Data? Here is a case study with Ovolo Hotels.
Lorien Pratt's insight:

Yes!  Or maybe I should say "duh" :-)  Good job Julien.  "Ovolo Hotels takes a different path: only questions with a purpose are used. If the answer to a question cannot impact the business directly, the question shall not be asked."

more...
No comment yet.
Rescooped by Lorien Pratt from Non-Equilibrium Social Science
Scoop.it!

Complex systems more popular in books than Artificial Intelligence

Complex systems more popular in books than Artificial Intelligence | Decision Intelligence | Scoop.it

A search in the Google corpus of books shows that the ngram Complex Systems surpassed Artificial Intelligence in 2006. Check the Google Books Ngram Viewer.


Via Jorge Louçã, NESS
Lorien Pratt's insight:

Not a surprise, but good to see the data. This only goes through 2006, though.  Plotting the same curve on Google Trends (http://www.google.com/trends/explore#q=%22artificial%20intelligence%22%2C%20complexity%2C%20%22unintended%20consequences%22%2C%20%22machine%20learning%22&cmpt=q) shows that "complex systems" has maintained its dominance over "artificial intelligence" since this crossover, except for one short period in early 2012, around the time that Judea Pearl won the Turning Award.


Interestingly, in recent years, "machine learning" has been on the rise, and is on a trend to cross "artificial intelligence" in the near future.

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

When you connect information to action, it becomes valuable

Will Hearst in conversation with futurist Paul Saffo.
Lorien Pratt's insight:

Well worth a watch.  Some of my favorite decision intelligence-related quotes:


"When you connect information to action, it becomes valuable."


"One person's externality is another's value...Google succeeds by creating profit deserts: they'll find something that someone else is paying money for, and they'll offer it for free.  Are we unwittingly paying a whole lot more than we think?"

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

Teaching Computers to Think Like Decision Makers: Mark Zangari

Now that computers can do a good job with processing and storing data, what's next?
Lorien Pratt's insight:

If you watch one video this year on decision intelligence, this should be it: 


Now that computers can do a good job with processing and storing data, what's next? Today, we create beautiful data visualizations so that we can understand the data, to think through systems and make decisions. Tomorrow, that systems thinking and decision making will benefit from computer support, as well.


A breakthrough talk, by Quantellia CEO Mark Zangari at the University of San Francisco.

more...
No comment yet.
Rescooped by Lorien Pratt from Unintended Consequences
Scoop.it!

The U.S. Intelligence Community's Kodak Moment

The U.S. Intelligence Community's Kodak Moment | Decision Intelligence | Scoop.it
The game is changing rapidly. Can Washington's intelligence community keep up?
more...
Lorien Pratt's curator insight, May 16, 6:57 PM

Another article by Josh Kerbel.  Like the last one I shared, I'll say if you read anything this month, this should be it.  Kerbel is one of our generation's great thinkers, and this article reflects his insightful prescription for how one complex, data-rich organization - the Intelligence Community (IC) in the US -  must evolve to understand context, interdependencies, systems, nonlinear effects, sensemaking, and more.  This is the essence of Decision Intelligence (no wonder, since Kerbel was very influential on our early work).


Please share.

Scooped by Lorien Pratt
Scoop.it!

Now that we're better at strategy, the challenge has shifted to the "execution imperative": translating strategy into effective action.

Now that we're better at strategy, the challenge has shifted to the "execution imperative": translating strategy into effective action. | Decision Intelligence | Scoop.it

In the book Execution: The Discipline of Getting Things Done, co-authors Larry Bossidy and Ram Charan essentially wrote that, “It’s all in the execution." Perhaps the most important step in executing well is to build a bridge that connects the vision and the strategy to the execution plan. Readers will learn what those bridges are and how they can build them to last.

Lorien Pratt's insight:

We're hearing more and more about the execution imperative lately in a Decision Intelligence context.  This article explains that strategy, once king, has given way to a new bottleneck now that it's conquered: and that's mapping that strategy into effective execution. The key: building effective bridges between strategic decisions an their implementation. 


This article describes many of those bridges.  A particularly interesting one: "Use an objective, systematic process so that you won’t let emotion or bias cloud the issues or simply default to the kinds of decisions you’ve made in the past. This will also force you to incorporate risk assessment in your decision-making."

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

When Machine Learning fails

When Machine Learning fails | Decision Intelligence | Scoop.it
Machine Learning has been used successfully in so many apps that you'd
think you can use it to predict just about anything you want.
Unfortunately, that's not the case.
Lorien Pratt's insight:

An excellent article even if you don't know anything about machine learning, to understand the basics of why it fails. 


The other reason: as you can see here, our ability to build the classifier depends on a) lots of data; and b) that data not being obsolete. 


In my experience, it turns out that ML fails to meet our needs in situations where data is not plentiful and/or the situation changes such that yesterday's data/classifier are not relevant to today. This article will help you to visualize what I mean by that.  For this reason, we tend to "look under the lamp post" for ML use cases.   There are a lot more places we'd like to use data for which these "lamp post" conditions aren't true.  And so ML doesn't just fail, it doesn't even try. 

more...
No comment yet.
Scooped by Lorien Pratt
Scoop.it!

The capacity to decide: one of humanity's 15 greatest challenges

The capacity to decide: one of humanity's 15 greatest challenges | Decision Intelligence | Scoop.it
Lorien Pratt's insight:

One of these things is not like the other: we must understand how a decision, once made, will ripple through interdependencies - whether across the earth or across our business (or both) to impact the future. We must decide how this cause-and-effect chain warps as external circumstances shift.  We must decide, and re-decide, in a complex dance.

more...
Lorien Pratt's curator insight, May 5, 11:21 PM

Good interdependency analysis allows us to avoid unintended consequences and to capitalize on synergies, in many domains. And a decision sets the chain in motion.

Jacqueline Chan's curator insight, May 28, 3:39 AM

We have the knowledge, resources, and capacity to come up with sustainable and effective solutions to these 15 global challenges. Collaboration, collective intelligence, and collective wisdom are key.

Rescooped by Lorien Pratt from Unintended Consequences
Scoop.it!

Economics students call for shakeup of the way their subject is taught

Economics students call for shakeup of the way their subject is taught | Decision Intelligence | Scoop.it
Students from 19 countries argue economics courses failing wider society by ignoring need to address 21st-century issues
Lorien Pratt's insight:

Economic modeling is a field that stands to benefit tremendously from decision intelligence.  This article is a data point confirming that.  Specifically:

- We need models that include links to other social science disciplines: how does climate impact the economy?  Medicine?  More. 

- We need to embrace the fact that the future is not like the past, and take accept the black swan nature of the economic events.  A forward systems model is the only way to do that: induction from past data is not enough

- Explicitly tracking the impact of varying assumptions (e.g. Greenspan's versus others') is essential to effectively working together.

more...
Lorien Pratt's curator insight, May 4, 8:54 PM

Economic modeling is a field that stands to benefit tremendously from decision intelligence.  This article is a data point confirming that.  Specifically:

- We need models that include links to other social science disciplines: how does climate impact the economy?  Medicine?  More. 

- We need to embrace the fact that the future is not like the past, and take accept the black swan nature of the economic events.  A forward systems model is the only way to do that: induction from past data is not enough

- Explicitly tracking the impact of varying assumptions (e.g. Greenspan's versus others') is essential to effectively working together.