Why and How is Watson a Cognitive System?

Jim Spohrer sends along a link to a talk by Rob High, IBM Chief Technology Officer, Fellow and Vice President,  Does a good job of discussing why Watson is cognitive, unique, and why it differs from previous attempts at delivering expertise.   Non technical.

I ask: So how you convince a complex organization this uniqueness is useful, stable and scalable? Only through clear operational examples in varying contexts.  That appears to be what they are working on.

Abstract

IBM® Watson™ represents a first step into cognitive systems, a new era of computing. Watson builds on the current era of programmatic computing but differs in significant ways. The combination of the following capabilities makes Watson unique:

Natural language processing by helping to understand the complexities of unstructured data, which makes up as much as 80 percent of the data in the world today

Hypothesis generation and evaluation by applying advanced analytics to weigh and evaluate a panel of responses based on only relevant evidence

Dynamic learning by helping to improve learning based on outcomes to get smarter with each iteration and interaction

Although none of these capabilities alone are unique to Watson, the combination delivers a powerful solution:

To move beyond the constraints of programmatic computing
To move from reliance on structured, local data to unlock the world of global, unstructured data
To move from decision tree-driven, deterministic applications to probabilistic systems that co-evolve with their users

To move from keyword-based search that provides a list of locations where an answer might (or might not) be located, to an intuitive, conversational means of discovering a set of confidence-ranked responses

This IBM Redguide™ publication describes how Watson combines natural language processing, dynamic learning, and hypothesis generation and evaluation to give direct, confidence-based responses.

For additional information about how Watson can transform how organizations think, act, and operate in the future, see the IBM Redbooks Point-of-View publication "Transforming the Way Organizations Think with Cognitive Systems", REDP-4961:

Table of contents:
What language is and why it is hard for computers to understand
IBM Watson understands language
Understanding language is just the beginning
Problems come in different shapes
Accuracy is improved through generalization  ... 

Imprecision in Computing

In the CACM:  Decrease precision.  For Saving energy apparently.   How is this making anything better?    The amount of energy involved seems minimal.  Yet understanding precision is important in large and complex problems is important.  Yesterday had a conversation about the wildly increased precision in imaging, since early Defense department experiences.  Perhaps not seen as necessary at the time, but considerable advantages have been derived since.

L'Oreal and Programmatic Digital Advertising

We did early work in this area.  By analytically matching advertising to predicted outcomes based on costs.   Essentially a big data optimization problem.   This worked for about a decade, but eventually the idea was outsourced.   We are now in different times, especially with regard to the amount of data we have.   In Adage:

L'Oreal USA Looks to Bring Programmatic Trading In-House
Beauty Giant Would Follow P&G and Others Amid Industry Concerns

L'Oreal USA looks to be the latest big marketer set to bring programmatic digital trading in house, soon after concluding its North American media review.

The company has posted a job listing for an assistant VP of Precision Advertising Information Technology with experience overseeing data management platforms and demand-side platforms used to build in-house trading desks. The person is to work with the company's chief marketing officer team and brand media leaders to identify "which platforms will be trialed and then scaled for broad use to achieve precision buying and tracking of media buys." .... ' 

See the programmatic ad buying tag below for much more.

Data Science Studying Data Scientists

Nice piece by Bob Hayes.   I like the structural aspects included.    Classification of skills can be imprecise too.  How does that imprecision influence the results.  The factor analysis, eigenvalue based approach used is well worth the read.   " .... Empirically-Based Approach to Understanding the Structure of Data Science by Bob Hayes, PhD    ... Based on a study of 620+ data professionals, we found that data science skills fall into three broad areas: domain expertise (in our case, business), technology/programming and math/statistics. I discuss the implications of study findings for current data scientists, would-be data scientists and the recruiters who try to find them. ... " 

Gershenfeld Discusses Fab Labs

  Via CACM:  Neil Gershenfeld talks about Fab Labs.   We met him in the early establishment of the idea of the Center for Bits and Atoms.

In MIT News

" ... Neil Gershenfeld, director of the Massachusetts Institute of Technology's Center for Bits and Atoms (CBA), launched the Fab Lab a decade ago. Since then, hundreds of Fab Labs have been installed in dozens of countries. In an interview, Gershenfeld discusses the growth and impact of this outreach project.

He notes the CBA runs a digital fabrication research facility, which contains tools worth millions of dollars, with a research roadmap leading up to a Star Trek-style replicator.

The facility was created with support from the U.S. National Science Foundation, and Fab Labs began as a modest project for the agency to expand access to these capabilities.

A Fab Lab today includes three-dimensional scanning and printing, large-format and precision machining, computer-controlled lasers and knives, surface-mount electronics production, embedded programming, and computing tools for design and collaboration. ... " 

Philip Tetlock and Super Forecasting

Recently covered in the BBC with an introductory article and broadcast.    About the gradual improvement of forecasting and the use of tournaments to drive them.  This topic has been covered here for some time.    Precision forecasting is fundamental for strategic and tactical organization planning. " ... Most intelligence agencies rely on the informed hunches of trained specialists. But there are people from ordinary walks of life who routinely do far better at predicting events than the experts. Tara Isabella Burton meets some of them. ... " 

On the Limitations of 3D Printing

In Fast Company.    Prototyping clearly.  But then real world scale and precision become a couple of the problems.  Interesting case study.

 " ... Yet Strauser’s production lines will remain firmly planted in Chinese factories. A 3-D printer wouldn’t be able to handle the volumes of plastic Strauser needs to meet his clients’ volume demands within an acceptable timeframe and with a good quality standard; his distributors ask for a 2,000-item minimum. Mass production, he says, is the only method that can accommodate that volume at a reasonable price point, so he’s staying in China for that, choosing to struggle to keep his business afloat in the sea of factories and companies there. ... " 

Industry Collaborative Robotics: Rethink's Sawyer

In IEEE Spectrum:  How collaborative?  How autonomous?

Rethink Robotics' Sawyer Goes on Sale, Rodney Brooks Says 'There May Be More Robots'   By Erico Guizzo   ....

" .... Rethink introduced Sawyer, the company’s second robot, earlier this year. Like Baxter, its predecessor, Sawyer can safely operate alongside human workers, who can teach the robot to perform tasks simply by moving its arm though a series of motions. But whereas Baxter has two arms and can help with packaging and material handling tasks, the single-arm Sawyer is designed for tasks that require more precision, including machine tending and circuit board testing. .... " 

GE Lighting example below: 

Print v Digital Advertising

Print vs. Digital: Another Emotional Win for Paper - Neuromarketing.   by Roger Dooley.   "... The latest findings come from Temple University consumer neuroscience researchers. In a study conducted for the U.S. Post Office, they found significant differences in the way our brains process the two ad formats. ... " .  I think this is place where neuromarketing can be best used,  to get broad understanding of what works, rather than precision analysis of particular ads. 

Amazon Echo versus Cortana Voice Control

I am continuing to test the Amazon Echo device, now comparing it to Microsoft Cortana.  Amazon has released the developer kit, and am awaiting third party applications.    The precision of the voice control has proved to be impressive, even at low voice levels in noisy environments.

In Engadget:   Amazon lets anyone put Alexa voice control in their devices  ... Amazon released the Alexa Appkit last month in hopes that developers will create cool new features for the voice technology that powers the Echo. This time, the e-commerce giant is offering the Alexa Voice Service (AVS) itself as a developer preview, which both hobbyists and legit hardware manufacturers can integrate into their own connected devices.  ... " 

High Precision 3D Printers

I have looked at several 3D printers now, mostly on the consumer end, and their precision was poor. Now new printer examples are emerging, with higher precision, and cost.   Precision also usually cuts  speed, limiting some applications.

Unilever to Track Online, Offline Retail Gap

Unilever is investing $50,000 in startup Glimr as part of an effort to connect consumers' online and in-store shopping experiences. "We are hoping to bring an actual understating of who you are as an individual based on what stores you visit and help Unilever understand their target audience in a much better way -- so for the first time bridging the offline and online gap," said Robert Hedberg, CEO of Glimr, which helps brands build consumer relationships based on data from beacons  .. " 

" .... Precision. .... Glimr revolutionizes advertisement by allowing brands to truly build 1-1 relationships with their customers by using offline data.  ... "  

Long Range Iris Scanning

We examined iris scanning for security applications.  You had to place your face on the reader, and it thus created a sanitary issue.  Also, some people could not be scanned because of face-eye structure. Now in the Atlantic:  " ...   Long-Range Iris Scanning Is Here ... An engineering professor at Carnegie Mellon says he’s invented technology that can identify someone from across the room with the precision of a fingerprint. ... "

Faster, More Accurate Drones

Have done a little playing around with simple drones.   My impression was that they were hard to control and slow.  Not any more.  Take a look at the videos in this CW article.   Tethered for training.  Swarms of high speed precision flying drones.  More precise and faster flying than a human could do in a plane.   Goal oriented and autonomous.  Smarter too.  Lots of opportunities.  And dangers.

Driverless Platforms

I was pointed to this as a advertisement in IEEE Spectrum, and found it interesting.  Especially dealing with the current and potential autonomy of such systems.   Have worked with a  number of advanced tracking systems in supply chain applications.

" .... KVH Industries is a leading manufacturer of solutions that provide global high-speed Internet, television, and voice services via satellite to mobile users at sea and on land. These solutions include the mini-VSAT Broadband network, TracPhone satellite communications systems, and TracVision satellite television systems. KVH is also a premier manufacturer of high performance sensors and integrated inertial systems for defense and commercial guidance and stabilization applications. ... " 

 " ... The Navigation Technology Behind Driverless Platforms ... Precision sensors are critical components in driverless vehicles. These sensors are integrated with LIDAR, cameras and other systems to enable obstacle avoidance and relative navigation. Learn why KVH sensors are preferred for autonomous platforms, even in GNSS-denied environments.  ... " 

DARPA Promoting New Tracking Methods

What are the broader implications of these new methods commercially?   What comes to mind are technologies that will be more precise and reliable inside buildings,  notably for retail applications and emergency response. .In CWorld:

" ... New types of self-contained instruments are under development that could better track position, time and direction of motion, which are critical aspects of GPS. DARPA is developing high-precision clocks, self-calibrating gyroscopes and accelerometers, and high-precision navigation instruments that can track position for long periods without relying on external sources.

DARPA is also researching new technologies that could make real-time tracking possible through a number of sources. DARPA is developing sensors that "use signals of opportunity" such as television, radio, cell towers, satellites, and even lightning, for real-time tracking. The effort, called ASPN (All Source Positioning and Navigation) alleviates issues related to fixing locations in buildings, deep foliage, underwater or underground, where GPS access can be limited ... "

Design Sketching and Coding

From IEEE: A favorite topic, how do we take the informal sketches of a non graphic artist and convert them into knowledge with precision?    See also post on CogSketch.

How Software Designers Interact with Sketches at the Whiteboard
Nicolas Mangano, Thomas D. LaToza, Marian Petre, and Andre van der Hoek, Member, IEEE

Abstract—Whiteboard sketches play a crucial role in software development, helping to support groups of designers in reasoning about a software design problem at hand. However, little is known about these sketches and how they support design ‘in the moment’ particularly in terms of the relationships among sketches, visual syntactic elements within sketches, and reasoning activities. To address this gap, we analyzed 14 hours of design activity by eight pairs of professional software designers, manually coding over 4000 events capturing the introduction of visual syntactic elements into sketches, focus transitions between sketches, and reasoning activities. 

Our findings indicate that sketches serve as a rich medium for supporting design conversations. Designers often use general-purpose notations. Designers introduce new syntactic elements to record aspects of the design, or re-purpose sketches as the design develops. Designers constantly shift focus between sketches, using groups of sketches together that contain complementary information. 

Finally, sketches play an important role in supporting several types of reasoning activities (mental simulation, review of progress, consideration of alternatives). But these activities often leave no trace and rarely lead to sketch creation. We discuss the implications of these and other findings for the practice of software design at the whiteboard and for the creation of new electronic software design sketching tools. ... "

(Full technical paper at link above)

The New Marketing Golden Age

Long piece, lots of interesting thoughts.  I have been involved in many parts of these efforts, not all successful.  Though the tech is fictionalized, it did happen and continues to evolve:   In McKinsey:

" ... Marketers are boosting their precision, broadening their scope, moving more quickly, and telling better stories. .... Science has permeated marketing for decades. Fans of the television drama Mad Men saw a fictionalized encounter when an IBM System/360 mainframe computer physically displaced the creative department of a late-1960s advertising agency. In reality, though, the 1960s through the early 1990s witnessed a happy marriage of advertising and technology as marketers mastered both the medium of television and the science of Nielsen ratings. These years gave birth to iconic advertising messages in categories ranging from sparkling beverages (“I’d like to buy the world a Coke”) to credit cards (“American Express. Don’t leave home without it”) to air travel (“British Airways: the world’s favourite airline”).  ... "  

Limitations of Geospatial Analytics

In Datamag:   Really some very classic issues about the limitations of long term forecasting in general.  

" ... From a statistical analysis standpoint, modeling natural disasters as black swan risk-mitigation scenarios is better than modeling them as events that can be predicted, much less averted, with any degree of actionable precision. For example, earthquakes remain the classic example of these kinds of catastrophes: they can’t be predicted with confidence, but their impact can be mitigated through various proactive measures. That risk profile also describes the threats that regions face from many meteorological events. In such scenarios, impacted regions are often little more than sitting ducks whose best hope is to mitigate the severity of the inevitable effects. ... " 

Dimensions that Influence Location Value

  Good piece in Cisco blog on the topic. Obvious yes, but the following discussion of why and how are very useful for retail applications.  By Jagdish Girimaji.

" .. Since the earliest maps, location accuracy has always been important. Just one degree off on the high seas and the seafaring navigator could be landing in Cuba instead of Florida. As we progress from GPS to indoor location, the quality of location accuracy is no less critical. To define the quality of location, we really need to understand the variables that influence location. This is especially true with mobile devices in the mix.

Quality of location accuracy has three dimensions:

Location precision
Refresh rate
System latency ...  "