Closing the Loop

From the Edge Foundation.  A Conversation with Chris Anderson.  I have worked with process control systems that are closed loop, and designed to be ...  We have many in our lives,  but many are hidden, from the lowly thermostat to ambient computers that continually wait for our commands.  In some cases its very obvious when the loop is closed,  but in most computing systems it is not. You should be able to tell if you map the process.

In the Edge: 
Closing the loop is a phrase used in robotics. Open-loop systems are when you take an action and you can't measure the results—there's no feedback. Closed-loop systems are when you take an action, you measure the results, and you change your action accordingly. Systems with closed loops have feedback loops; they self-adjust and quickly stabilize in optimal conditions. Systems with open loops overshoot; they miss it entirely.

Chris Anderson is the CEO of 3D Robotics and founder of DIY Drones. He is the former editor-in-chief of Wired magazine.  ..... "     Podcast and text ... "

Adaptive Machine Learning

Good points in DSC article below on adaptive machine learning.    I often make the case that you should often consider if your modeling should be modeled adaptively.    Data and context can be changing even if we don't expect it.  Time often drives change.  Even if your data is not necessarily streaming.  With an eye towards risk and change management:

Adaptive Machine Learning

Posted by PG Madhavan on May 20, 2016 at 5:30amView Blog
Machine Learning today tends to be “open-loop” – collect tons of data offline, process them in batches and generate insights for eventual action. There is an emerging category of ML business use cases that are called “In-Stream Analytics (ISA)”. Here, the data is processed as soon as it arrives and insights are generated quickly. However, action may be taken offline and the effects of the actions are not immediately incorporated back into the learning process. If we did, it is an example of a “closed-loop” system – we will call this approach “Adaptive Machine Learning” or AML. ISA is a precursor to AML.  .... " 

Talk: Cyber Social Learning Systems

Good talk today, looking at the complexity of delivering learning to health care systems.  The integration of control systems, machine learning and AI with human beings in the loop.   A considerable challenge.

" .... Kevin Sullivan from University of Virginia, presented on "Cyber-Social Learning Systems: Take-Aways from First Community Computing Consortium Workshop on Cyber-Social Learning Systems."  .... " 

Slides
Recording: https://youtu.be/35wGssvVzTo
Further Discussions on Linkedin: https://www.linkedin.com/groups/6729452

Humans are Still at Work

In the HBR.    For now the humans are clearly still in the loop. we can learn from some areas, like autopilots or process control here.    Helping them effectively be in the loop is part of our near term task.    It means integrating a conversation, leading from problem to solution, that may include a number of machines and humans.  We do this already, when an accountant uses a spreadsheet.  Now how will each human or machine most efficiently and credibly contribute in a problem solving process?

The Humans Working Behind the AI Curtain
Mary L. Gray,  Siddharth Suri

There is a human factor at work in tasks promoted as artificial intelligence (AI)-driven, in the form of people paid to respond to queries and requests sent to them via application programming interfaces of crowdwork systems, write Microsoft Research scientists Mary L. Gray and Siddharth Suri. "The creation of human tasks in the wake of technological advancement has been a part of automation's history since the invention of the machine lathe," they note. "We call this ever-moving frontier of AI's development the paradox of automation's last mile: as AI makes progress, it also results in the rapid creation and destruction of temporary labor markets for new types of humans-in-the-loop tasks." 

Gray and Suri predict the enhancement of human services by AI will augment daily productivity, but present new social challenges. "The AI of today can't function without humans in the loop, whether it's delivering the news or a complicated pizza order," the researchers note. Technology and media companies therefore employ people to perform content moderation and curation, while many jobs are outsourced overseas and paid a low, flat rate. "This workforce deserves training, support, and compensation for being at-the-ready and willing to do an important job that many might find tedious or too demanding," according to Gray and Suri.  .... " 

Autonomous Selection of Mars Laser Targets

An example of autonomy for a complex system.  Here the Curiosity Rover on Mars.   At one level this is similar to closed loop process control, but with more complex sensor analysis being done in that loop.  

Assume this increases accuracy, speed in going through analysis goals ... and even decreases targeting labor required to enact,  thus decreasing cost.     So is  closed loop process control we did in manufacturing, though the adjustments here appear to start to arise to the strategic.   Article and image examples:

From the Jet Propulsion Lab: 
  " .... NASA's Mars rover Curiosity is now selecting rock targets for its laser spectrometer -- the first time autonomous target selection is available for an instrument of this kind on any robotic planetary mission.

Using software developed at NASA's Jet Propulsion Laboratory, Pasadena, California, Curiosity is now frequently choosing multiple targets per week for a laser and a telescopic camera that are parts of the rover's Chemistry and Camera (ChemCam) instrument. Most ChemCam targets are still selected by scientists discussing rocks or soil seen in images the rover has sent to Earth, but the autonomous targeting adds a new capability. ... "

Algorithms and Amazon Workers

In the BBC:  Am particularly interested in how algorithms and workers interact.     There has always been an interpretation of the results provided by computing, even close loop systems, like those in process control.  What will this be like when the interaction looks more like a continuous, learning conversation?

Customers are Talking, Companies are Listening

This as news surprises me.  Our company always listened.  In fact pioneered methods for listening and analyzing what we heard.   Hardest part is then, linking that knowledge to some action.  Complete the loop.   From Bain.

Examining Data Preparation Tools

Better tools, likely that includes cognitive and context knowledge capabilities are needed.  Also paying more attention to metadata needs:  In O'Reilly:

Why data preparation frameworks rely on human-in-the-loop systems
The O'Reilly Data Show Podcast: Ihab Ilyas on building data wrangling and data enrichment tools in academia and industry. ... " 

CrowdTruth Collects Gold Standard Data

Nice idea, examining further.   Experiences?  We experimented with similar ideas using Mechanical Turk.   More will follow here.

CrowdTruth  via  #CrowdTruth  by Anca Dumitrache
collecting gold standard data for training and evaluation of cognitive computing systems 

Welcome to the CrowdTruth blog!

The CrowdTruth Framework implements an approach to machine-human computing for collecting annotation data on text, images and videos. The approach is focussed specifically on collecting gold standard data for training and evaluation of cognitive computing systems. The original framework was inspired by the IBM Watson project for providing improved (multi-perspective) gold standard (medical) text annotation data for the training and evaluation of various IBM Watson components, such as Medical Relation Extraction, Medical Factor Extraction and Question-Answer passage alignment.

The CrowdTruth framework supports the composition of CrowdTruth gathering workflows, where a sequence of micro-annotation tasks can be configured and sent out to a number of crowdsourcing platforms (e.g. CrowdFlower and Amazon Mechanical Turk) and applications (e.g. Expert annotation game Dr. Detective). The CrowdTruth framework has a special focus on micro-tasks for knowledge extraction in medical text (e.g. medical documents, from various sources such as Wikipedia articles or patient case reports). The main steps involved in the CrowdTruth workflow are: (1) exploring & processing of input data, (2) collecting of annotation data, and (3) applying disagreement analytics on the results. These steps are realised in an automatic end-to-end workflow, that can support a continuous collection of high quality gold standard data with feedback loop to all steps of the process. Have a look at our presentations and papers for more details on the research.  ... "   

Human in the Loop Planning

Attended excellent CSig talk today by Subbarao Kambhampati and Kartik Talamadupula of ASU and  IBM Watson on Human-in-the-Loop Planning and Decision Support.

Human in the Loop Planning (HILP) is a classic example of studying how people interact with smart systems.  These systems are getting much smarter, but how they interact and cooperate with people is still a problem. You can say that any advisory system has components of planning systems.   In the enterprise we looked at many such systems, and delivered a few.  Yet there are many 'challenges' that still exist in making such systems work, still not resolved after decades of work.

   The slides from the talk give an excellent overview of work still underway.   Recording replay.  Also, they point to their much more detailed tutorial on the subject.    An interesting aspect, an idea we never addressed directly,  they call:  "Planning for Crowdsourcing".  Since we were involved with marketing problems, which dealt with consumer group activities, this could have been of much use to us.

If you have any interest in applied AI for commercial use, this is an area to follow.

Data and the Internet of Things

In O'Reilly: Big Data and IoT.  Data and as we have been seeing, the architecture to support that data.  A four dimensional dilemma:

" ... The Internet of Things (IoT) has a data problem. Well, four data problems. Walking the halls of CES in Las Vegas last week, it’s abundantly clear that the IoT is hot. Everyone is claiming to be the world’s smartest something. But that sprawl of devices, lacking context, with fragmented user groups, is a huge challenge for the burgeoning industry.

What the IoT needs is data. Big data and the IoT are two sides of the same coin. The IoT collects data from myriad sensors; that data is classified, organized, and used to make automated decisions; and the IoT, in turn, acts on it. It’s precisely this ever-accelerating feedback loop that makes the coin as a whole so compelling.

Nowhere are the IoT’s data problems more obvious than with that darling of the connected tomorrow known as the wearable. Yet, few people seem to want to discuss these problems ... "

Internet of Things and Big Data

Is it all about data and leveraging analytics.  And an architecture to support them.

In OReilly: " ... The Internet of Things (IoT) has a data problem. Well, four data problems. Walking the halls of CES in Las Vegas last week, it’s abundantly clear that the IoT is hot. Everyone is claiming to be the world’s smartest something. But that sprawl of devices, lacking context, with fragmented user groups, is a huge challenge for the burgeoning industry.

What the IoT needs is data. Big data and the IoT are two sides of the same coin. The IoT collects data from myriad sensors; that data is classified, organized, and used to make automated decisions; and the IoT, in turn, acts on it. It’s precisely this ever-accelerating feedback loop that makes the coin as a whole so compelling. ... " 

Mining Search Trends for Ad Buys

In Adage: Google mines search trends.    " ....  Two years ago, WPP's Mindshare created a tool for its client Kleenex that used Google search data to see where in the United Kingdom people were searching for things related to the flu. Based on that data, Kleenex was able to shift its TV ad spend to make sure people in those areas saw its spots. ... Now Mindshare has developed the tool into a full-fledged search-trend analyzer with Google. Called Search As Signal, the tool tracks what people are searching on Google, where around the world they're doing those searches and on what device and identifies trends. ... "  

See also Mindshare's Loop Room, which futher looks at the delivery mechanism for this kind of data. Similarities to P&G's Business Sphere.  Get the right data to the right people, in real-time.

Marketing and the Internet of Things

It is about data.  Its about reach.    Anna Russell writes:

" .... Let’s take a step back and look at just what the Internet of Things contains that might excite marketers.  IoT, the result of a growing number of M2M and internet enabled devices, produces vast and diverse amounts of machine created data such as log files, GIS coordinates, thermostat data, and other sensory data. This data is transmitted to a location (or many locations) outside the device, and can trigger actions in other devices that either guide manufacturing efficiency, change environmental factors or overtly assist human users in selecting a course of action, be it workout intensity, room temperature or driving route.

These data driven decisions and suggestions, when well designed, have the power to surprise and delight device users with their perceptive power. And its that feedback loop -the expectation of users that these devices are trustworthy and will advise them of best thing to do - that presents unparalleled albeit slightly big-brother-ish marketing potential. ... "  

Triple Loop Reboots and Resolving to be Different

The triple loop reboot was new to me.  Via colleague Julie Anixter. " ...  Remember, human beings crave stability and permanence.  When change threatens the stability of what we know, we often seek refuge in “what we have always done.” The paradox in this inclination is that change is the most enduring element of our lives. “What we have always done” is change and unless we willfully initiate a simple but systematic way to ensure we keep learning, unlearning, and relearning, we risk going down the same sad path as the French generals in the first half of the twentieth century.

Resolve to be different. Commit to making the Triple-Loop Reboot a natural part of your strategy to navigate change.  Beginning today. ... " 

Analytics for Decision Makers Needed before Autopilots

Ultimately all analytics is about decision making.  From the most simple yes/no, to the most complex management of risk portfolios.  Decision making usually includes the decisions of humans, based on the assistance of computing devices.  True, process control applications do close the loop to create 'autopilots' in every manufacturing environment today, but there are many, many challenges before we close all the loops. One will be to get analytics into the hands of the human decision makers, remarkably rare, even these days.  It can be done, starting with the Cloud.

Collaborative Information Seeking

An abstract, but the challenge is interesting.  " ... Information searches based on expert-seeking technology can prove time-consuming or unsuccessful if search terms do not turn up extrinsic identifiers in profiles and saved documents. In many such cases, knowledge brokers function as "humans in the loop," providing intrinsic enterprise knowledge to mediate between information seekers and expert sources--a fact that future collaborative information-seeking system designs should take into account. .... "   Similar to what the company Zakta has been  attempting at the practical level for years.

Visualization Errors and Design

Some common errors in visualization.  Good to think about, but I am not sure these are the most important errors you could make.    But the final section is right: It is all about good design, and by first focusing that design on key issues, and then provide a means to adapt that design to things you discover as part of the process. Put the decision maker in the loop.  Involve them as early as possible with the design.

Why Innovation Programs Fail

In Innovation Excellence:  The top reason,  and based on my experience, I agree:
"A lack of a feedback loop from decision-makers to participants is the most common contributor to the failure of an innovation program. Ensuring a feedback loop is therefore the best action you can take to sustain the program over months and years ... "

Humans in the Intelligence Loop

Have you noticed?  Real humans are entering the App loop.    ' ... Google Helpouts and Amazon's Mayday are changing everything ... " .  Its long been my contention that people, properly leveraged, are a powerful add in for delivering intelligence.