Archive for the ‘Smart Systems’ Category

Ontologies as Productive Assets

January 22, 2018


An often overlooked benefit of artificial intelligence has been a renewed interest in seminal philosophical and cognitive topics; ontologies coming top of the list.

Ontological Questioning (The Thinker Monkey, Breviary of Mary of Savoy)

Yet that interest has often been led astray by misguided perspectives, in particular:

  • Universality: one-fits-all approaches are pointless if not self-defeating considering that ontologies are meant to target specific domains of concerns.
  • Implementation: the focus is usually put on representation schemes (commonly known as Resource Description Frameworks, or RDFs), instead of the nature of targeted knowledge and the associated cognitive capabilities.

Those misconceptions, often combined, may explain the limited practical inroads of ontologies. Conversely, they also point to ontologies’ wherewithal for enterprises immersed into boundless and fluctuating knowledge-driven business environments.

Ontologies as Assets

Whatever the name of the matter (data, information or knowledge), there isn’t much argument about its primacy for business competitiveness; insofar as enterprises are concerned knowledge is recognized as a key asset, as valuable if not more than financial ones, and should be managed accordingly. Pushing the comparison still further, data would be likened to liquidity, information to fixed income investment, and knowledge to capital ventures. To summarize, assets whatever their nature lose value when left asleep and bear fruits when kept awake; that’s doubly the case for data and information:

  • Digitized business flows accelerates data obsolescence and makes it continuous.
  • Shifting and porous enterprises boundaries and markets segments call for constant updates and adjustments of enterprise information models.

But assessing the business value of knowledge has always been a matter of intuition rather than accounting, even when it can be patented; and most of knowledge shapes up well beyond regulatory reach. Nonetheless, knowledge is not manna from heaven but the outcome of information processing, so assessing the capabilities of such processes could help.

Admittedly, traditional modeling methods are too stringent for that purpose, and looser schemes are needed to accommodate the open range of business contexts and concerns; as already expounded, that’s precisely what ontologies are meant to do, e.g:

  • Systems modeling,  with a focus on integration, e.g Zachman Framework.
  • Classifications, with a focus on range, e.g Dewey Decimal System.
  • Conceptual models, with a focus on understanding, e.g legislation.
  • Knowledge management, with a focus on reasoning, e.g semantic web.

And ontologies can do more than bringing under a single roof the whole of enterprise knowledge representations: they can also be used to nurture and crossbreed symbolic assets and develop innovative ones.

Ontologies Benefits

Knowledge is best understood as information put to use; accounting rules may be disputed but there is no argument about the benefits of a canny combination of information, circumstances, and purpose. Nonetheless, assessing knowledge returns is hampered by the lack of traceability: if a part of knowledge is explicit and subject to symbolic representation, another is implicit and manifests itself only through actual behaviors. At philosophical level it’s the line drawn by Wittgenstein: “The limits of my language mean the limits of my world”;  at technical level it’s AI’s two-lanes approach: symbolic rule-based engines vs non symbolic neural networks; at corporate level implicit knowledge is seen as some unaccounted for aspect of intangible assets when not simply blended into corporate culture. With knowledge becoming a primary success factor, a more reasoned approach of its processing is clearly needed.

To begin with, symbolic knowledge can be plied by logic, which, quoting Wittgenstein again, “takes care of itself; all we have to do is to look and see how it does it.” That would be true on two conditions:

  • Domains are to be well circumscribed. 
  • A water-tight partition must be secured between the logic of representations and the semantics of domains.

That could be achieved with modular and specific ontologies built on a clear distinction between common representation syntax and specific domains semantics.

As for non-symbolic knowledge, its processing has for long been overshadowed by the preeminence of symbolic rule-based schemes, that is until neural networks got the edge and deep learning overturned the playground. In a few years’ time practically unlimited access to raw data and the exponential growth in computing power have opened the door to massive sources of unexplored knowledge which is paradoxically both directly relevant yet devoid of immediate meaning:

  • Relevance: mined raw data is supposed to reflect the geology and dynamics of targeted markets.
  • Meaning: the main value of that knowledge rests on its implicit nature; applying existing semantics would add little to existing knowledge.

Assuming that deep learning can transmute raw base metals into knowledge gold, enterprises would need to understand, assess, and improve the refining machinery. That could be done with ontological frames.

Further Reading

External Links


Open Ontologies: From Silos to Architectures

January 1, 2018

To be of any use for enterprises, ontologies have to embrace a wide range of contexts and concerns, often ill-defined for environments, rather well expounded for systems.

Circumscribed Contexts & Crossed Concerns (Robert Goben)

And now that enterprises have to compete in open, digitized, and networked environments, business and systems ontologies have to be combined into modular knowledge architectures.

Ontologies & Contexts

If open-ended business contexts and concerns are to be taken into account, the first step should be to characterize ontologies with regard to their source, justification, and the stability of their categories, e.g:

  • Social: No authority, volatile, continuous and informal changes.
  • Institutional: Regulatory authority, steady, changes subject to established procedures.
  • Professional: Agreed upon between parties, steady, changes subject to established procedures.
  • Corporate: Defined by enterprises, periodic, changes subject to internal decision-making.
  • Personal: Customary, defined by named individuals (e.g research paper).

Assuming such an external taxonomy, the next step would be to see what kind of internal (i.e enterprise architecture) ontologies can be fitted into, as it’s the case for the Zachman framework.

The Zachman’s taxonomy is built on well established concepts (Who,What,How, Where, When) applied across architecture layers for enterprise (business and organization), systems (logical structures and functionalities), and platforms (technologies). These layers can be generalized and applied uniformly across external contexts, from well-defined (e.g regulations) to fuzzy (e.g business prospects or new technologies) ones, e.g:

Ontologies, capabilities (Who,What,How, Where, When), and architectures (enterprise, systems, platforms).

That “divide to conquer” strategy is to serve two purposes:

  • By bridging the gap between internal and external taxonomies it significantly enhances the transparency of governance and decision-making.
  • By applying the same motif (Who,What, How, Where, When) across the semantics of contexts, it opens the door to a seamless integration of all kinds of knowledge: enterprise, professional, institutional, scientific, etc.

As can be illustrated using Zachman concepts, the benefits are straightforward at enterprise architecture level (e.g procurement), due to the clarity of supporting ontologies; not so for external ones, which are by nature open and overlapping and often come with blurred semantics.

Ontologies & Concerns

A broad survey of RDF-based ontologies demonstrates how semantic overlaps and folds can be sort out using built-in differentiation between domains’ semantics on one hand, structure and processing of symbolic representations on the other hand. But such schemes are proprietary, and evidence shows their lines seldom tally, with dire consequences for interoperability: even without taking into account relationships and integrity constraints, weaving together ontologies from different sources is to be cumbersome, the costs substantial, and the outcome often reduced to a muddy maze of ambiguous semantics.

The challenge would be to generalize the principles as to set a basis for open ontologies.

Assuming that a clear line can be drawn between representation and contents semantics, with standard constructs (e.g predicate logic) used for the former, the objective would be to classify ontologies with regard to their purpose, independently of their representation.

The governance-driven taxonomy introduced above deals with contexts and consequently with coarse-grained modularity. It should be complemented by a fine-grained one to be driven by concerns, more precisely by the epistemic nature of the individual instances to be denoted. As it happens, that could also tally with the Zachman’s taxonomy:

  • Thesaurus: ontologies covering terms and concepts.
  • Documents: ontologies covering documents with regard to topics.
  • Business: ontologies of relevant enterprise organization and business objects and activities.
  • Engineering: symbolic representation of organization and business objects and activities.

Ontologies: Purposes & Targets

Enterprises could then pick and combine templates according to domains of concern and governance. Taking an on-line insurance business for example, enterprise knowledge architecture would have to include:

  • Medical thesaurus and consolidated regulations (Knowledge).
  • Principles and resources associated to the web-platform (Engineering).
  • Description of products (e.g vehicles) and services (e.g insurance plans) from partners (Business).

Such designs of ontologies according to the governance of contexts and the nature of concerns would significantly reduce blanket overlaps and improve the modularity and transparency of ontologies.

On a broader perspective, that policy will help to align knowledge management with EA governance by setting apart ontologies defined externally (e.g regulations), from the ones set through decision-making, strategic (e.g plate-form) or tactical (e.g partnerships).

Open Ontologies’ Benefits

Benefits from open and formatted ontologies built along an explicit distinction between the semantics of representation (aka ontology syntax) and the semantics of context can be directly identified for:

Modularity: the knowledge basis of enterprise architectures could be continuously tailored to changes in markets and corporate structures without impairing enterprise performances.

Integration: the design of ontologies with regard to the nature of targets and stability of categories could enable built-in alignment mechanisms between knowledge architectures and contexts.

Interoperability: limited overlaps and finer granularity are to greatly reduce frictions when ontologies bearing out business processes are to be combined or extended.

Reliability: formatted ontologies can be compared to typed programming languages with regard to transparency, internal consistency, and external validity.

Last but not least, such reasoned design of ontologies may open new perspectives for the collaboration between cognitive humans and pretending ones.

Further Reading

External Links

2018: Clones vs Octopuses

December 4, 2017

In the footsteps of robots replacing workmen, deep learning bots look to boot out knowledge workers overwhelmed by muddy data.

Cloning Knowledge (Tadeusz Cantor, from “The Dead Class”)

Faced with that , should humans try to learn deeper and faster than clones, or should they learn from octopuses and their smart hands.

Machine Learning & The Economics of Clones

As illustrated by scan-reading AI machines, the spreading of learning AI technology in every nook and cranny introduces something like an exponential multiplier: compared to the power-loom of the Industrial Revolution which substituted machines for workers, deep learning is substituting replicators for machines; and contrary to power looms, there is no physical limitation on the number of smart clones that can be deployed. So, however fast and deep humans can learn, clones are much too prolific: it’s a no-win situation. To get out of that conundrum humans have to put their hand on a competitive edge, e.g some kind of knowledge that cannot be cloned.

Knowledge & Competition

Appraising humans learning sway over machines, one can take from Spinoza’s categories of knowledge with regard to sources:

  1. Senses (views, sounds, smells, touches) or beliefs (as nurtured by the supposed common “sense”). Artificial sensors can compete with human ones, and smart machines are much better if prejudiced beliefs are put into the equation.
  2. Reasoning, i.e the mental processing of symbolic representations. As demonstrated by AlphaGo, machines are bound to fast extend their competitive edge.
  3. Philosophy which is by essence meant to bring together perceptions, intuitions, and symbolic representations. That’s where human intelligence could beat its artificial cousin which is clueless when purposes are needed.

That assessment is bore out by evolution: the absolute dominance established by humans over other animal species comes from their use of knowledge, which can be summarized as:

  1. Use of symbolic representations.
  2. Ability to formulate and exchange representations of contexts, concerns, and policies.
  3. Ability to agree on stakes and cooperate on policies.

On that basis, the third dimension, i.e the use of symbolic knowledge to cooperate on non-zero-sum endeavors, can be used to draw the demarcation line between human and artificial intelligence:

  • Paths and paces of pursuits as part and parcel of the knowledge itself. The fact that both are mostly obviated by search engines gives humans some edge.
  • Operational knowledge is best understood as information put to use, and must include concerns and decision-making. But smart bots’ ubiquity and capabilities often sap information traceability and decisions transparency, which makes room for humans to prevail.

So humans can find a clear competitive edge in this knowledge dimension because it relies on a combination of experience and thinking and is therefore hard to clone. Organizations should make sure that’s where smart systems take back and humans take up.

Organization & Innovation

Innovation being at the root of competitive edge, understanding the role played by smart systems is a key success factor; that is to be defined by organization.

As epitomized by Henry Ford, industrial-era thinking associated innovation with top-down management and the specialization of execution:

  • At execution level manual tasks were to be fragmented and specialized.
  • At management level analysis and decision-making were to be centralized and abstracted.

That organizational paradigm puts a double restraint on innovation:

  • On execution side the fragmentation of manual tasks prevents workers from effectively assessing and improving their performances.
  • On management side knowledge is kept in conceptual boxes and bereft of feedback from actual uses.

That railing between smart brains and dumb hands may have worked well enough for manufacturing processes limited to material flows and subject to circumscribed and predictable technological changes. It didn’t last.

First, as such hierarchies necessarily grow with processes complexity, overheads and rigidity force repeated pruning. Then, flat hierarchies are of limited use when information flows are to be combined with material ones, so enterprises have to start with matrix organization. Finally, with the seamless integration of digital and material flows, perpetuating the traditional line between management and execution is bound to hamstring innovation:

  • Smart tools may be able to perform a wide range of physical tasks without human supervision, but the core of innovation core as well as its front lines are where human and machines collaborate in processing a mix of material and information flows, both learning from the experience.
  • Hierarchies and centralized decision-making are being cut out from feeders when set in networked business environments colonized by smart bots on both sides of corporate boundaries.

Not surprisingly, these innovation trends seem to tally with the social dimension of knowledge.

Learning from the Octopus

The AI revolution has already broken all historical records of footprint (everything is affected) and speed (a matter of years). Given the length of human education cycles, appraising the consequences comes with some urgency, beginning with the disposal of two entrenched beliefs:

At individual level the new paradigm could be compared to the nervous system of octopuses: each arm gets its brain and neurons, and so its own touch of knowledge and taste of decision-making.

On a broader (i.e enterprise) perspective, knowledge should be supported by two organizational layers, one direct and innovation-driven between trusted co-workers, the other networked and knowledge-driven between remote workers, trusted or otherwise.

Further Reading

External Links

Transcription & Deep Learning

September 17, 2017

Humans looking for reassurance against the encroachment of artificial brains should try YouTube subtitles: whatever Google’s track record in natural language processing, the way its automated scribe writes down what is said in the movies is essentially useless.

A blank sheet of paper was copied on a Xerox machine.
This copy was used to make a second copy.
The second to make a third one, and so on…
Each copy as it came out of the machine was re-used to make the next.
This was continued for one hundred times, producing a book of one hundred pages. (Ian Burn)

Experience directly points to the probable cause of failure: the usefulness of real-time transcriptions is not a linear function of accuracy because every slip can be fatal, without backup or second chance. It’s like walking a line: for all practical purposes a single misunderstanding can throw away the thread of understanding, without a chance of retrieve or reprieve.

Contrary to Turing machines, listeners have no finite states; and contrary to the sequence of symbols on tapes, tales are told by weaving together semantic threads. It ensues that stories are work in progress: readers can pause to review and consolidate meanings, but listeners have no other choice than punting on what comes to they mind, hopping that the fabric of the story will carry them out.

So, whereas automated scribes can deep learn from written texts and recorded conversations, there is no way to do the same from what listeners understand. That’s the beauty of story telling: words may be written but meanings are renewed each time the words are heard.

Further Reading

Why Virtual Reality (VR) is Late

July 25, 2017


Whereas virtual reality (VR) has been expected to be the next breakthrough for IT human interfaces, the future seems to be late.

Detached Reality (N.Ghesquiere, G.Coddington)

Together with the cost of ownership, a primary cause mentioned for the lukewarm embrace is the nausea associated with the technology. Insofar as the nausea is provoked by a delay in perceptions, the consensus is that both obstacles should be overcame by continuous advances in computing power. But that optimistic assessment rests on the assumption that the nausea effect is to be uniformly decreasing.

Virtual vs Augmented

The recent extension of a traditional roller-coaster at SeaWorld Orlando illustrates the difference between virtual and augmented reality. Despite being marketed as virtual reality, the combination of actual physical experience (roller-coaster) with virtual perceptions (3D video) clearly belongs to the augmented breed, and its success may put some new light on the nausea effect.

Consciousness Cannot Wait

Awareness is what anchors living organisms to their environment. So, lest a confusion is introduced between individuals experience and their biological clock, perceptions are to be immediate; and since that confusion is not cognitive but physical, it will cause nausea. True to form, engineers initial answer has been to cut down elapsed time through additional computing power; that indeed brought a decline in the nausea effect, as well as an increase in the cost of ownership. Unfortunately, benefits and costs don’t tally: however small is the remaining latency, nausea effects are disproportionate.

Aesop’s Lesson

The way virtual and augmented reality deal with latency may help to understand the limitations of a minimizing strategy:

  • With virtual reality latency occurs between users voluntary actions (e.g moving their heads) and devices (e.g headset) generated responses.
  • With augmented reality latency occurs between actual perceptions and software generated responses.

That’s basically the situation of Aesop’s “The Tortoise and the Hare” fable: in the physical realm the hare (aka computer) is either behind or ahead of the tortoise (the user), which means that some latency (positive or negative) is unavoidable.

That lesson applies to virtual reality because both terms are set in actuality, which means that nausea can be minimized but not wholly eliminated. But that’s not the case for augmented reality because the second term is a floating variable that can be logically adjusted.

The SeaWorld roller-coaster takes full advantage of this point by directly tying up augmented stimuli to actual ones: augmented reality scripts are aligned with roller-coaster episodes and their execution synchronized through special sensors. Whatever the remaining latency, it is to be of a different nature: instead of having to synchronize their (conscious) actions with the environment feedback, users only have to consolidate external stimuli, a more mundane task which doesn’t involve consciousness.

Further Reading

External Links

Deep Blind Testing

March 21, 2017


Tests are meant to ensure that nothing will go amiss. Assuming that expected hazards can be duly dealt with beforehand, the challenge is to guard against unexpected ones.

Unexpected Outcome (Ariel Schlesinger)

That would require the scripting of every possible outcomes in an unlimited range of unknown circumstances, and that’s where Deep Learning may help.

What to Look For

As Donald Rumsfeld once famously said, there are things that we know we don’t know, and things we don’t know we don’t know; hence the need of setting things apart depending on what can be known and how, and build the scripts accordingly:

  • Business requirements: tests can be designed with respect to explicit specifications; yet some room should also be left for changes in business circumstances.
  • Functional requirements: assuming business requirements are satisfied, the part played by supporting systems can be comprehensively tested with respect to well-defined boundaries and operations.
  • Quality of service: assuming business and functional requirements are satisfied, tests will have to check how human interfaces and resources are to cope with users behaviors and expectations which, by nature, cannot be fully anticipated.
  • Technical requirements: assuming business and functional requirements are satisfied as well as users’ expectations for service, deployment, maintenance, and operations are to be tested with regard to feasibility and costs.

Automated testing has to take into account these differences between scope and nature, from bounded and defined specifications to boundless, fuzzy and changing circumstances.

Automated Software Testing

Automated software testing encompasses two basic components: first the design of test cases (events, operations, and circumstances), then their scripted execution. Leading frameworks already integrate most of the latter together with the parts of the former targeting technical aspects like graphical user interfaces or system APIs. Artificial intelligence (AI) and machine learning (ML) have also been tried for automated test generation, yet with a scope limited by dependency on explicit knowledge, and consequently by the need of some “manual” teaching. That hurdle may be overcame by the deep learning ability to get direct (aka automated) access to implicit knowledge.

Reconnaissance: Known Knowns

Systems are designed artifacts, with the corollary that their components are fully defined and their behavior predictable. The design of technical test cases can therefore be derived from what is known of software and systems architectures, the former for test units, the latter for integration and acceptance tests. Deep learning could then mine recorded log-files in order to identify critical cases’ events and circumstances.

Exploration: Known Unknowns

Assuming that applications must be tested for use during their expected shelf life, some uncertainty has to be factored in for future business circumstances. Yet, assuming applications are designed to meet specific business objectives, such hypothetical circumstances should remain within known boundaries. In that context deep learning could be applied to exploration as well as policies:

  • Compared to technical test cases that can rely on the content of systems log-files, business and functional ones have to look outside and mine raw data from business environments.
  • In return, the relevancy of observations can be assessed with regard to business objectives, improved, and feed the policy module in charge of defining test cases.

Blind Errands: Unknown Unknowns

Even with functional and technical capabilities well-tested and secured, quality of service may remain contingent on human quirks: instinctive or erratic behaviors that could thwart the best designed handrails. On one hand, and due to their very nature, such hazards are not to be easily forestalled by reasoned test cases; but on the other hand they don’t take place in a void but within known functional circumstances. Given that porosity of functional and cognitive layers, the validity of functional test cases may be compromised by unfathomable cognitive associations, and that could open the door to unmanageable regression. Enter deep learning and its ability to extract knowledge from insignificance.

Compared to business and functional test cases, hazards are not directly related to business activities. As a consequence, the learning process cannot be guided by business and functional test cases but has to chart unpredictable human behaviors. As it happens, that kind of learning combining random simulation with automated reinforcement is what makes the specificity of deep learning.

From Non-regression to Self-improvement

As a conclusion, if non-regression is to be the cornerstone of quality management, test cases are to be set along clear swim-lanes: business logic (independently of systems), supporting systems functionalities (for shared applications), users interfaces (for non shared interactions). Then, since test cases are also run across swim-lanes, it opens the door to feedback, e.g unit test cases reassessed directly from business rules independently of systems functionalities, or functional test cases reassessed from users’ behaviors.

Considering that well-defined objectives, sound feedback mechanisms, and the availability of massive data from systems logs (internal) and business environment (external) are the main pillars of deep learning technologies, their combination in integrated frameworks could result in a qualitative leap toward self-improving automated test cases.

Further Reading


Alternative Facts & Augmented Reality

February 5, 2017


Coming alongside the White House creative use of facts, the upcoming Snap’s IPO is to bring another perspective on reality with its Snapchat star product integrating augmented reality (AR) with media.


Layers of Reality (Marcel Duchamp)

Whatever the purpose, the “alternative facts” favored by the White House communication detail may bring to the fore two related issues of present-day relevancy: virtual and augmented reality on one hand, the actuality of George Orwell’s Newspeak on the other hand.

Facts and Fiction

To begin with, facts are not given but observed, and that can only be achieved through a mix of conceptual and technical apparatus, the former to design fact-finding vessels, the latter to fill them with actual observations. Based on that understanding, alternatives are less about the facts themselves than about the apparatuses used to collect them, which may be trustworthy, faulty, or deceitful. Setting flaws aside, trust is also what distinguishes augmented and virtual reality:

  • Augmented reality (AR) technologies operate on apparatuses that combine observation and analysis before adding layers of information.
  • Virtual reality (VR) technologies simply overlook the whole issue of reality and observation, and are only concerned with the design of trompe l’oeuils.

The contrast between facts (AR) and fiction (VR) may account for the respective applications and commercial advances: whereas augmented reality is making rapid inroads in business applications, its virtual cousin is still testing the water in games. More significantly perhaps, the comparison points to a somewhat unexpected difference in the role of language: necessary for the establishment of facts, accessory for the creation of fictions.

Speaking of Alternative Facts

As illustrated (pun intended) by virtual reality, fiction can do without words, which is not the case for facts. As a matter of fact (intended again), even facts can be fictional, as epitomized by Orwell’s Newspeak, the language used by the totalitarian state in his 1949 novel Nineteen Eighty-Four. Figuratively speaking, that language may be likened to a linguistic counterpart of virtual reality as its purpose is to bypass the issue of trusty discourse about reality by introducing narratives wholly detached from actual observations. And that’s when fiction catches up with reality: no much stretch of imagination is needed to recognize a similar scheme in current White House’s comments.

Language Matter

As far as humans are concerned, reality comes with semantic and social dimensions that can only be carried out through language. In other words truth is all about the use of language with regard to purpose: communication, information, or knowledge. Taking Trump’s inauguration crowd for example:


Data come from observations, Information is Data put in form, Knowledge is Information put to use.

  • Communication: language is used to exchange observations associated to immediate circumstances (the place and the occasion).
  • Information: language is used to map observations to mental representations and operations (estimates for the size of the audience).
  • Knowledge: language is use to associate information to purposes through categories and concepts detached of the original circumstances (comparison of audiences for similar events and political conclusions).

Augmented Reality devices on that occasion could be used to tally people on viewed portions of the audience (fact), figure out estimates for the whole audience (information), or decide on the best itineraries back home (knowledge). By contrast, Virtual Reality (aka “alternative facts”) could only be used at communication level to deceive the public.

Further Reading

Things Speaking in Tongues

January 25, 2017


Speaking in tongues (aka Glossolalia) is the fluid vocalizing of speech-like syllables without any recognizable association with a known language. Such experience is best (not ?) understood as the actual speaking of a gutted language with grammatical ghosts inhabited by meaningless signals.

The man behind the tongue (Herbert List)

Do You Hear What I Say ? (Herbert List)

Usually set in religious context or circumstances, speaking in tongue looks like souls having their own private conversations. Yet, contrary to extraterrestrial languages, the phenomenon is not fictional and could therefore point to offbeat clues for natural language technology.

Computers & Language Technology

From its inception computers technology has been a matter of language, from machine code to domain specific. As a corollary, the need to be in speaking terms with machines (dumb or smart) has put a new light on interpreters (parsers in computer parlance) and open new perspectives for linguistic studies. In due return, computers have greatly improve the means to experiment and implement new approaches.

During the recent years advances in artificial intelligence (AI) have brought language technologies to a critical juncture between speech recognition and meaningful conversation, the former leaping ahead with deep learning and signal processing, the latter limping along with the semantics of domain specific languages.

Interestingly, that juncture neatly coincides with the one between the two intrinsic functions of natural languages: communication and representation.

Rules Engines & Neural Network

As exemplified by language technologies, one of the main development of deep learning has been to bring rules engines and neural networks under a common functional roof, turning the former unfathomable schemes into smart conceptual tutors for the latter.

In contrast to their long and successful track record in computer languages, rule-based approaches have fallen short in human conversations. And while these failings have hindered progress in the semantic dimension of natural language technologies, speech recognition have strode ahead on the back of neural networks fueled by increasing computing power. But the rift between processing and understanding natural languages is now being fastened through deep learning technologies. And with the leverage of rule engines harnessing neural networks, processing and understanding can be carried out within a single feedback loop.

From Communication to Cognition

From a functional point of view, natural languages can be likened to money, first as medium of exchange, then as unit of account, finally as store of value. Along that understanding natural languages would be used respectively for communication, information processing, and knowledge representation. And like the economics of money, these capabilities are to be associated to phased cognitive developments:

  • Communication: languages are used to trade transient signals; their processing depends on the temporal persistence of the perceived context and phenomena; associated behaviors are immediate (here-and-now).
  • Information: languages are also used to map context and phenomena to some mental representations; they can therefore be applied to scripted behaviors and even policies.
  • Knowledge: languages are used to map contexts, phenomena, and policies to categories and concepts to be stored as symbolic representations fully detached of original circumstances; these surrogates can the be used, assessed, and improved on their own.

As it happens, advances in technologies seem to follow these cognitive distinctions, with the internet of things (IoT) for data communications, neural networks for data mining and information processing, and the addition of rules engines for knowledge representation. Yet paces differ significantly: with regard to language processing (communication and information), deep learning is bringing the achievements of natural language technologies beyond 90% accuracy; but when language understanding has to take knowledge into account, performances still lag a third below: for computers knowledge to be properly scaled, it has to be confined within the semantics of specific domains.

Sound vs Speech

Humans listening to the Universe are confronted to a question that can be unfolded in two ways:

  • Is there someone speaking, and if it’s the case, what’s the language ?.
  • Is that a speech, and if it’s the case, who’s speaking ?.

In both case intentionality is at the nexus, but whereas the first approach has to tackle some existential questioning upfront, the second can put philosophy on the back-burner and focus on technological issues. Nonetheless, even the language first approach has been challenging, as illustrated by the difference in achievements between processing and understanding language technologies.

Recognizing a language has long been the job of parsers looking for the corresponding syntax structures, the hitch being that a parser has to know beforehand what it’s looking for. Parser’s parsers using meta-languages have been effective with programming languages but are quite useless with natural ones without some universal grammar rules to sort out babel’s conversations. But the “burden of proof” can now be reversed: compared to rules engines, neural networks with deep learning capabilities don’t have to start with any knowledge. As illustrated by Google’s Multilingual Neural Machine Translation System, such systems can now build multilingual proficiency from sufficiently large samples of conversations without prior specific grammatical knowledge.

To conclude, “Translation System” may even be self-effacing as it implies language-to-language mappings when in principle such systems can be fed with raw sounds and be able to parse the wheat of meanings from the chaff of noise. And, who knows, eventually be able to decrypt languages of tongues.

Further Reading

External Links

NIEM & Information Exchanges

January 24, 2017


The objective of the National Information Exchange Model (NIEM) is to provide a “dictionary of agreed-upon terms, definitions, relationships, and formats that are independent of how information is stored in individual systems.”

(Alfred Jensen)

NIEM’s model makes no difference between data and information (Alfred Jensen)

For that purpose NIEM’s model combines commonly agreed core elements with community-specific ones. Weighted against the benefits of simplicity, this architecture overlooks critical distinctions:

  • Inputs: Data vs Information
  • Dictionary: Lexicon and Thesaurus
  • Meanings: Lexical Items and Semantics
  • Usage: Roots and Aspects

That shallow understanding of information significantly hinders the exchange of information between business or institutional entities across overlapping domains.

Inputs: Data vs Information

Data is made of unprocessed observations, information makes sense of data, and knowledge makes use of information. Given that NIEM is meant to be an exchange between business or institutional users, it should have no concern with data mining or knowledge management.

Data is meaningless, information meaning is set by semantic domains.

As an exchange, NIEM should have no concern with data mining or knowledge management.

The problem is that, as conveyed by “core of data elements that are commonly understood and defined across domains, such as person, activity, document, location”, NIEM’s model makes no explicit distinction between data and information.

As a corollary, it implies that data may not only be meaningful, but universally so, which leads to a critical trap: as substantiated by data analytics, data is not supposed to mean anything before processed into information; to keep with examples, even if the definition of persons and locations may not be specific, the semantics of associated information is nonetheless set by domains, institutional, regulatory, contractual, or otherwise.

Data is meaningless, information meaning is set by semantic domains.

Data is meaningless, information meaning is set by semantic domains.

Not surprisingly, that medley of data and information is mirrored by NIEM’s dictionary.

Dictionary: Lexicon & Thesaurus

As far as languages are concerned, words (e.g “word”, “ξ∏¥” ,”01100″) remain data items until associated to some meaning. For that reason dictionaries are built on different levels, first among them lexical and semantic ones:

  • Lexicons take items on their words and gives each of them a self-contained meaning.
  • Thesauruses position meanings within overlapping galaxies of understandings held together by the semantic equivalent of gravitational forces; the meaning of words can then be weighted by the combined semantic gravity of neighbors.

In line with its shallow understanding of information, NIEM’s dictionary only caters for a lexicon of core standalone items associated with type descriptions to be directly implemented by information systems. But due to the absence of thesaurus, the dictionary cannot tackle the semantics of overlapping domains: if lexicons alone can deal with one-to-one mappings of items to meanings (a), thesauruses are necessary for shared (b) or alternative (c) mappings.


Shared or alternative meanings cannot be managed with lexicons

With regard to shared mappings (b), distinct lexical items (e.g qualification) have to be mapped to the same entity (e.g person). Whereas some shared features (e.g person’s birth date) can be unequivocally understood across domains, most are set through shared (professional qualification), institutional (university diploma), or specific (enterprise course) domains .

Conversely, alternative mappings (c) arise when the same lexical items (e.g “mole”) can be interpreted differently depending on context (e.g plastic surgeon, farmer, or secret service).

Whereas lexicons may be sufficient for the use of lexical items across domains (namespaces in NIEM parlance), thesauruses are necessary if meanings (as opposed to uses) are to be set across domains. But thesauruses being just tools are not sufficient by themselves to deal with overlapping semantics. That can only be achieved through a conceptual distinction between lexical and semantic envelops.

Meanings: Lexical Items & Semantics

NIEM’s dictionary organize names depending on namespaces and relationships:

  • Namespaces: core (e.g Person) or specific (e.g Subject/Justice).
  • Relationships: types (Counselor/Person) or properties (e.g PersonBirthDate).

NIEM’s Lexicon: Core (a) and specific (b) and associated core (c) and specific (d) properties

But since lexicons know only names, the organization is not orthogonal, with lexical items mapped indifferently to types and properties. The result being that, deprived of reasoned guidelines, lexical items are chartered arbitrarily, e.g:

Based on core PersonType, the Justice namespace uses three different schemes to define similar lexical items:

  • “Counselor” is described with core PersonType.
  • “Subject” and “Suspect” are both described with specific SubjectType, itself a sub-type of PersonType.
  • “Arrestee” is described with specific ArresteeType, itself a sub-type of SubjectType.

Based on core EntityType:

  • The Human Services namespace bypasses core’s namesake and introduces instead its own specific EmployerType.
  • The Biometrics namespace bypasses possibly overlapping core Measurer and BinaryCaptured and directly uses core EntityType.
Lexical items are meshed disregarding semantics

Lexical items are chartered arbitrarily

Lest expanding lexical items clutter up dictionary semantics, some rules have to be introduced; yet, as noted above, these rules should be limited to information exchange and stop short of knowledge management.

Usage: Roots and Aspects

As far as information exchange is concerned, dictionaries have to deal with lexical and semantic meanings without encroaching on ontologies or knowledge representation. In practice that can be best achieved with dictionaries organized around roots and aspects:

  • Roots and structures (regular, black triangles) are used to anchor information units to business environments, source or destination.
  • Aspects (italics, white triangles) are used to describe how information units are understood and used within business environments.
nformation exchanges are best supported by dictionaries organized around roots and aspects

Information exchanges are best supported by dictionaries organized around roots and aspects

As it happens that distinction can be neatly mapped to core concepts of software engineering.

P.S. Thesauruses & Ontologies

Ontologies are systematic accounts of existence for whatever is considered, in other words some explicit specification of the concepts meant to make sense of a universe of discourse. From that starting point three basic observations can be made:

  1. Ontologies are made of categories of things, beings, or phenomena; as such they may range from simple catalogs to philosophical doctrines.
  2. Ontologies are driven by cognitive (i.e non empirical) purposes, namely the validity and consistency of symbolic representations.
  3. Ontologies are meant to be directed at specific domains of concerns, whatever they can be: politics, religion, business, astrology, etc.

With regard to models, only the second one puts ontologies apart: contrary to models, ontologies are about understanding and are not supposed to be driven by empirical purposes.

On that basis, ontologies can be understood as thesauruses describing galaxies of concepts (stars) and features (planets) held together by semantic gravitation weighted by similarity or proximity. As such ontologies should be NIEM’s tool of choice.

Further Reading

External Links

New Year: 2016 is the One to Learn

December 15, 2016

Sometimes the future is best seen through rear-view mirrors; given the advances of artificial intelligence (AI) in 2016, hindsight may help for the year to come.


Deep Mind Learning (J.Bosh)

Deep Learning & the Depths of Intelligence

Deep learning may not have been discovered in 2016 but Google’s AlphaGo has arguably brought a new dimension to artificial intelligence, something to be compared to unearthing the spherical Earth.

As should be expected for machines capabilities, artificial intelligence has for long been fettered by technological handcuffs; so much so that expert systems were initially confined to a flat earth of knowledge to be explored through cumbersome sets of explicit rules. But exponential increase in computing power has allowed neural networks to take a bottom-up perspective, mining for implicit knowledge hidden in large amount of raw data.

Like digging tunnels from both extremities, it took some time to bring together top-down and bottom-up schemes, namely explicit (rule-based) and implicit (neural network-based) knowledge processing. But now that it comes to fruition, the alignment of perspectives puts a new light on the cognitive and social dimensions of intelligence.

Intelligence as a Cognitive Capability

Assuming that intelligence is best defined as the ability to solve problems, the first criterion to consider is the type of input (aka knowledge) to be used:

  • Explicit: rational processing of symbolic representations of contexts, concerns, objectives, and policies.
  • Implicit: intuitive processing of factual (non symbolic) observations of objects and phenomena.

That distinction is broadly consistent with the one between humans, seen as the sole symbolic species with the ability to reason about explicit knowledge, and other animal species which, despite being limited to the processing of implicit knowledge, may be far better at it than humans. Along that understanding, it would be safe to assume that systems with enough computing power will sooner or later be able to better the best of animal species, in particular in the case of imperfect inputs.

Intelligence as a Social Capability

Alongside the type of inputs, the second criterion to be considered is obviously the type of output (aka solution). And since classifications are meant to be built on purpose, a typology of AI outcomes should focus on relationships between agents, humans or otherwise:

  • Self-contained: problem-solving situations without opponent.
  • Competitive: zero-sum conflictual activities involving one or more intelligent opponents.
  • Collaborative: non-zero-sum activities involving one or more intelligent agents.

That classification coincides with two basic divides regarding communication and social behaviors:

  1. To begin with, human behavior is critically different when interacting with living species (humans or animals) and machines (dumb or smart). In that case the primary factor governing intelligence is the presence, real or supposed, of beings with intentions.
  2. Then, and only then, communication may take different forms depending on languages. In that case the primary factor governing intelligence is the ability to share symbolic representations.

A taxonomy of intelligence with regard to cognitive (reason vs intuition) and social (symbolic vs non-symbolic) capabilities may help to clarify the role of AI and the importance of deep learning.

Between Intuition and Reason

Google’s AlphaGo astonishing performances have been rightly explained by a qualitative breakthrough in learning capabilities, itself enabled by the two quantitative factors of big data and computing power. But beyond that success, DeepMind (AlphaGo’s maker) may have pioneered a new approach to intelligence by harnessing both symbolic and non symbolic knowledge to the benefit of a renewed rationality.

Perhaps surprisingly, intelligence (a capability) and reason (a tool) may turn into uneasy bedfellows when the former is meant to include intuition while the latter is identified with logic. As it happens, merging intuitive and reasoned knowledge can be seen as the nexus of AlphaGo decisive breakthrough, as it replaces abrasive interfaces with smart full-duplex neural networks.

Intelligent devices can now process knowledge seamlessly back and forth, left and right: borne by DeepMind’s smooth cognitive cogwheels, learning from factual observations can suggest or reinforce the symbolic representation of emerging structures and behaviors, and in return symbolic representations can be used to guide big data mining.

From consumers behaviors to social networks to business marketing to supporting systems, the benefits of bridging the gap between observed phenomena and explicit causalities appear to be boundless.

Further Reading

External Links


Your content with a new angle at

IT Modernization < V.Hanniet

About IT Modernization

IT Modernization < V. Hanniet

software model driven approaches

Caminao's Ways

Do systems know how symbolic they are ?