NOMad–Networks of Meaning

John Coyne Technology Computer Artificial Intelligence Networks of Meaning

NOMad:  Business Analysis

NOMad is more than an infrastructure for data analysis.  It does what no other competing software system does.  It reveals the “meaning” of communication between people.  It also reveals the impact people have on the “things” they are associated with and the impact of those “things” on them.   Its analysis engines reveal nascent associations between people and things, to reveal potential connections and emergent networks that could have been missed by unaided analysis.  It does this through a unique combination of applied artificial intelligence techniques, institutional best practice knowledge modeled and executable by a computer, and massively-parallel computer power that provides enormous scalability.

It is a single unified environment made up of three major components:

  • REDpillRevealing the information in a meaningful way.  It provides the presentation of the data in multiple formats that allow the user to choose which one they prefer.  
  • NetgraphrSetting up context for roles, responsibilities and relationships.  This uses the system core to set up contexts so that meaning is derived from the data it is presented and inferred.
  • NOMadThe core AI engines and filter algorithms.  There are three main AI engines that are described later together with institutional knowledge modeled to store and maintain the best practices in any given domain.  This is the foundation of the systems ability to find meaning in data and derive insights that are presented to the user through context.


NOMad:  Data Aggregation and Integration

Digital footprints are left everywhere.  NOMad has an open data aggregation and integration facility that makes using the system “frictionless.”  In other words, NOMad formats data in a unified way regardless of its source format.  There is no need to be a database expert to use NOMad.  A simple API allows third parties to associate data and applications to the NOMad environment.  This is especially important for custom filters that are used to analyze particular nuance of data and information.


NOMad:  Man-Machine Collaboration

NOMad works for the human analysts using the system by doing what computers do best:  crunch millions of variables in seconds.  Unlike any alternative, NOMad provides an intelligent collaboration agent that assists the user by presenting alternate views and machine-generated insights.  

Search and discovery using NOMad is based on natural language formats.  Our system interprets and verifies what the user is looking for–and more.  It includes real things and precise or vague concepts.   It may reveal analysis that is associated with the search that the user had not thought to ask.  This is the revelation of nascent connections and emergent networks that enhance the human analysis capability.  This also means that the underlying NOMad capability is usable by novice analysts getting the benefit of the embedded and modeled best practices.  Institutional know-how can be continually enhanced and, to some degree, the system will utilize self-learning based on results feedback.

This works when the system tests a new model of combined algorithms and user feedback is rated.  If good, then it is retained as a new model, if rejected, it is logged as unfit for that particular purpose but can be tested for other contexts or recombined to find a successful new analysis models.


Reasoning reporting

All analysis is associated with a computer-generated reasoning trace that tells the user why it made its value judgements and recommendations.  It atomizes data elements and models of behavior so that we also keep a complete trace of every data element that affected the reasoning engines.  Things like “where did we get this data?”  How reliable is it?  “When did it become relevant and useful?”  “Who is allowed to see it?” Etc.

The NOMad reasoning engines use three types of AI techniques to produce their results:  Deduction, Induction and Abduction.

  • Deductive reasoning basically says, “given these premises to be true, what can you tell me about a set of data?”  For instance:  All people who bomb without warning are terrorists.  These people were caught bombing, therefore we deduce that we can classify them as terrorists.  It is generally used in a closed-world environment and a high degree of certainty is guaranteed.
  • Inductive reasoning is more open to the world–following uncertain paths to find credible conclusions.  This is how nascent connectivity between people, things and emergent networks are found in NOMad.  It basically takes the data, weights or values it and through a given process and says, “here is connection to follow.”  For instance, it may show that a random phone call by someone as yet unidentified was made to one of our aforementioned bombers and creates a nascent link that could be explored and extrapolated.
  • Abductive reasoning is less conclusive, but nonetheless very powerful.  This engine uses its modeled institutional knowledge to provide a best explanation of what it is observing.  It generates theories.   It might say that “Given all the information I am looking at, a terrorist cell is forming in this location and is funded by these people.”  This is a short-cut that allows the computing power to fast-track observations.

These engines can be observing such things as events over time, and the evolution of information and its significance.  They can also track frequency of actions associated with events, and categorize and classify the information to make it useful.  This is true of both Meta data (data about data) and raw data either archived or real-time, or any combination to find effective conclusions or recommendations.


n-array inferences

One of the key differentiators in the use of these engines their ability to mix inference techniques in virtually any array of needed processing predicated on the findings they generate.  The intelligence within the knowledge base of “how to use inferences”–part of NOMad’s inference management function–provides this capability.  It drives the automated collaboration agent’s ability to deliver machine-generated “insights.”


Context and Worldviews

We believe that all truths are “true in a model.”  That is, it may be true in one context, but not another.  We call these “worldviews.”  The ability to establish worldviews allows users to collaborate with one another across contexts that may have differing axioms and rules of analysis.   Importantly, it allows users to collaborate with the inference engines and let the computer use its potential to present information in a variety of ways and contexts.  In some cases NOMad could present a context not previously rendered through its best theory approach.

In people collaboration, we can share worldviews, open dialogues and show users how others are viewing the information.  While computers are powerful, humans can rapidly apply insights that can confound even the best data crunchers.  Unlike other alternatives that rely on human assessment, such insights can be added to the computer’s models to improve its institutional know how and replicate the insights to other users and retain the knowledge in the absence of the human.


Team repositories

Version management.  Change is a constant factor and managing it is a complex task.  NOMad provides consistent archiving of results and factors in even the smallest change to a presented worldview model.  This means that users can view any instance of a worldview.  Importantly, the system itself will monitor and notify the user of changes in the version of the views, the data, the reasoning . . . in fact, all contributing factors.  


Predictive Analysis

Futurecasting is a term we use to provide predictive analysis.  It allows the user to set up a not just a “what if” scenario, but would allow the user to put in change that is known will take place on a certain date and test the impact on the currently known information.  For instance, we know that a certain event is going to take place in six months, the user can test, “what will be the impact on our ability to respond be given these variables.”

Nascent connections and Nodes provide information on emergent networks that may have been overlooked or missed by humans based on the computer’s capability to present new potentials.  This is the underlying true power of NOMad.  It looks for meaning in the information, tests hypotheses that it generates, weights the results and presents its findings.  


Finding degrees of influence  

People and things have differing impacts within a worldview.  NOMad can find those people and things that have the most impact and present them.  This is not only useful for analysis but for planning too.  If you wanted to connect with someone outside of your network, you could look across networks to find the fastest path to an important connection.  Equally, the NOMad system can point out the prospective main influencer on others within the network and report on why.



Many business analysis tools present data in a variety of ways and provide reports for interpretation by analysts.  NOMad finds the underlying meaning, makes suggestions on paths to follow or generates theories to be explored.  That is the NOMad Networks Of Meaning advantage.