Artificial intelligence is generally defined in relation with the human ability to figure out situations and solve problems. The term may also put the focus on an hypothetical disembodied artificial brain.
Taking the brain perspective makes for easier outlining:
Artificial brains can build and process symbolic and non symbolic representations, respectively with semantic and neural networks.
Artificial brains can solve problems applying logic and abstraction or data analytics, respectively to symbolic and non symbolic representations.
Like human ones, artificial brains combine sensory-motor capabilities with purely cognitive ones.
Like human ones, and apart from sensory-motor capabilities, artificial brains support a degree of cognitive plasticity and versatility between symbolic and non symbolic representation and processing.
These generic capabilities are at the root of the wide-ranging and dramatic advances of machine learning.
The upsurge in the scope and performances of artificial brains sometimes brings a new light on human cognition. Semantic layers and knowledge graphs offer a good example of a return to classics, in that case with Greek philosophers’ ontologies.
According to their philosophical origins, ontologies are systematic accounts of existence for whatever can make sense in an universe of discourse. From that starting point four basic observations can be made:
Ontologies are structured set of names denoting symbolic (aka cognitive) representations.
These representations can stand for whatever may be taken into consideration: terms (nothing is represented), ideas (sets of terms), instances of objects or phenomena, categories (sets of instances), documents.
Ontologies are solely dedicated to the validity and internal consistency of the representations.
Ontologies are not contingent on truth (epistemic) status, i.e external consistency. As a corollary, they are not meant to support emprical purposes.
That makes models a refinement of ontologies as they are meant to be externally consistent and serve a purpose.
As championed by a brave writer, should we see the Web as a crib for born again narratives, or as a crypt for redundant texts.
Once Upon A Time
Borrowing from Einstein, “the only reason for time is so that everything doesn’t happen at once.” That befits narratives: whatever the tale or the way it is conveyed, stories take time. Even if nothing happens, a story must be spelt in tempo and can only be listened to or read one step at a time.
In So Many Words
Stories have been told before being written, which is why their fabric is made of words, and their motifs weaved by natural languages. So, even if illustrations may adorn printed narratives, the magic of stories comes from the music of their words.
A Will To Believe
To enjoy a story, listeners or readers are to detach their mind from what they believe about reality, replacing dependable and well-worn representations with new and untested ones, however shaky or preposterous they may be; and that has to be done through an act of will.
Stories are make-beliefs: as with art in general, their magic depends on the suspension of disbelief. But suspension is not abolition; while deeply submerged in stories, listeners and readers maintain some inward track to the beliefs they left before diving; wandering a cognitive fold between surface truths and submarine untruths, they seem to rely on a secure if invisible tether to the reality they know. On that account, the possibility of an alternative reality is to transform a comforting fold into a menacing abyss, dissolving their lifeline to beliefs. That could happen to stories told through the web.
Stories & Medium
Assuming time rendering, stories were not supposed to be affected by medium; that is, until McLuhan’s suggestion of medium taking over messages. Half a century later internet and the Web are bringing that foreboding in earnest by melting texts into multimedia documents.
Tweets and Short Message Services (SMS) offer a perfect illustration of the fading of text-driven communication, evolving from concise (160 characters) text-messaging to video-sharing.
That didn’t happen by chance but reflects the intrinsic visual nature of web contents, with dire consequence for texts: once lording it over entourages of media, they are being overthrown and reduced to simple attachments, just a peg above fac-simile. But then, demoting texts to strings of characters makes natural languages redundant, to be replaced by a web Esperanto.
Web Semantic Silos
With medium taking over messages, and texts downgraded to attachments, natural languages may lose their primacy for stories conveyed through the web, soon to be replaced by the so-called “semantic web”, supposedly a lingua franca encompassing the whole of internet contents.
As epitomized by the Web Ontology Language (OWL), the semantic web is based on a representation scheme made of two kinds of nodes respectively for concepts (squares) and conceptual relations (circles).
Concept nodes are meant to represent categories specific to domains (green, right); that tallies with the lexical level of natural languages.
Connection nodes are used to define two types of associations:
Semantically neutral constructs to be applied uniformly across domains; that tallies with the syntactic level of natural languages (blue, left).
Domain specific relationships between concepts; that tallies with the semantic level of natural languages (green, center).
The mingle of generic (syntactic) and specific (semantic) connectors induces a redundant complexity which grows exponentially when different domains are to be combined, due to overlapping semantics. Natural languages typically use pragmatics to deal with the issue, but since pragmatics scale poorly with exponential complexity, they are of limited use for semantic web; that confines its effectiveness to silos of domain specific knowledge.
But semantic silos are probably not the best nurturing ground for stories.
Stories In Cobwebs
Taking for granted that text, time, and suspension of disbelief are the pillars of stories, their future on the web looks gloomy:
Texts have no status of their own on the web, but only appear as part of documents, a media among others.
Stories can bypass web practice by being retrieved before being read as texts or viewed as movies; but whenever they are “browsed” their intrinsic time-frame and tempo are shattered, and so is their music.
If lying can be seen as an inborn human cognitive ability, it cannot be separated from its role in direct social communication; such interactive background should also account for the transient beliefs in fictional stories. But lies detached from a live context and planted on the web are different beasts, standing on their own and bereft of any truth currency that could separate actual lies from fictional ones.
That depressing perspective is borne out by the tools supposed to give a new edge to text processing:
Hyper-links are part and parcel of internet original text processing. But as far and long as stories go, introducing links (hardwired or generated) is to hand narrative threads over to readers, and by so transforming them into “entertextment” consumers.
Machine learning can do wonders mining explicit and implicit meanings from the whole of past and present written and even spoken discourses. But digging stories out is more about anthropology or literary criticism than about creative writing.
As for the semantic web, it may work as a cobweb: by getting rid of pragmatics, deliberately or otherwise, it disables narratives by disengaging them from their contexts, cutting them out in one stroke from their original meaning, tempo, and social currency.
The Deconstruction of Stories
Curiously, what the web may do to stories seems to reenact a philosophical project which gained some favor in Europe during the second half of the last century. To begin with, the deconstruction philosophy was rooted in literary criticism, and its objective was to break the apparent homogeneity of narratives in order to examine the political, social, or ideological factors at play behind. Soon enough, a core of upholders took aim at broader philosophical ambitions, using deconstruction to deny even the possibility of a truth currency.
With the hindsight on initial and ultimate purposes of the deconstruction project, the web and its semantic cobweb may be seen as the stories nemesis.
Turning thoughts into figures faces the intrinsic constraint of dimension: two dimensional representations cannot cope with complexity.
So, lest they be limited to flat and shallow thinking, mind cartographers have to introduce the cognitive equivalent of geographical layers (nature, demography, communications, economy,…), and archetypes (mountains, rivers, cities, monuments, …)
Nodes: What’s The Map About
Nodes in maps (aka roots, handles, …) are meant to anchor thinking threads. Given that human thinking is based on the processing of symbolic representations, mind mapping is expected to progress wide and deep into the nature of nodes: concepts, topics, actual objects and phenomena, artifacts, partitions, or just terms.
It must be noted that these archetypes are introduced to characterize symbolic representations independently of domain semantics.
Connectors: Cognitive Primitives
Nodes in maps can then be connected as children or siblings, the implicit distinction being some kind of refinement for the former, some kind of equivalence for the latter. While such a semantic latitude is clearly a key factor of creativity, it is also behind the poor scaling of maps with complexity.
A way to frame complexity without thwarting creativity would be to define connectors with regard to cognitive primitives, independently of nodes’ semantics:
References connect nodes as terms.
Associations: connect nodes with regard to their structural, functional, or temporal proximity.
Analogies: connect nodes with regard to their structural or functional similarities.
At first, with shallow nodes defined as terms, connections can remain generic; then, with deeper semantic levels introduced, connectors could be refined accordingly for concepts, documentation, actual objects and phenomena, artifacts,…
Semantics: Extensional vs Intensional
Given mapping primitives defined independently of domains semantics, the next step is to take into account mapping purposes:
Extensional semantics deal with categories of actual instances of objects or phenomena.
Intensional semantics deal with specifications of objects or phenomena.
That distinction can be applied to basic semantic archetypes (people, roles, events, …) and used to distinguish actual contexts, symbolic representations, and specifications, e.g:
Car (object) refers to context, not to be confused with Car (surrogate) which specified the symbolic counterpart: the former is extensional (actual instances), the latter intensional (symbolic representations)
Maintenance Process is extensional (identified phenomena), Operation is intensional (specifications).
Reservation and Driver are symbolic representations (intensional), Person is extensional (identified instances).
It must be reminded that whereas the choice is discretionary and contingent on semantic contexts and modeling purposes (‘as-it-is’ vs ‘as-it-should-be’), consequences are not because the choice is to determine abstraction semantics.
For example, the records for cars, drivers, and reservations are deemed intensional because they are defined by business concerns. Alternatively, instances of persons and companies are defined by contexts and therefore dealt with as extensional descriptions.
Abstractions: Subsets & Sub-types
Thinking can be characterized as a balancing act between making distinctions and managing the ensuing complexity. To that end, human edge over other animal species is the use of symbolic representations for specialization and generalization.
That critical mechanism of human thinking is often overlooked by mind maps due to a confused understanding of inheritance semantics:
Strong inheritance deals with instances: specialization define subsets and generalization is defined by shared structures and identities.
Weak inheritance deals with specifications: specialization define sub-types and generalization is defined by shared features.
The combination of nodes (intension/extension) and inheritance (structures/features) semantics gives cartographers two hands: a free one for creative distinctions, and a safe one for the ensuing complexity. e.g:
Intension and weak inheritance: environments (extension) are partitioned according to regulatory constraints (intension); specialization deals with subtypes and generalization is defined by shared features.
Extension and strong inheritance: cars (extension) are grouped according to motorization; specialization deals with subsets and generalization is defined by shared structures and identities.
Intension and strong inheritance: corporate sub-type inherits the identification features of type Reservation (intension).
Mind maps built on these principles could provide a common thesaurus encompassing the whole of enterprise data, information and knowledge.
Intelligence: Data, Information, Knowledge
Considering that mind maps combine intelligence and cartography, they may have some use for enterprise architects, in particular with regard to economic intelligence, i.e the integration of information processing, from data mining to knowledge management and decision-making:
Data provide the raw input, without clear structures or semantics (terms or aspects).
Categories are used to process data into information on one hand (extensional nodes), design production systems on the other hand (intensional nodes).
Abstractions (concepts) makes knowledge from information by putting it to use.
Along that perspective mind maps could serve as front-ends for enterprise architecture ontologies, offering a layered cartography that could be organized according to concerns:
Enterprise architects would look at physical environments, business processes, and functional and technical systems architectures.
Knowledge managers would take a different perspective and organize the maps according to the nature and source of data, information, and knowledge.intelligence w
As demonstrated by geographic information systems, maps built on clear semantics can be combined to serve a wide range of purposes; furthering the analogy with geolocation assistants, layered mind maps could be annotated with punctuation marks (e.g ?, !, …) in order to support problem-solving and decision-making.
“For things to remain the same, everything must change”
Lampedusa, “The Leopard”
Whatever the understanding of the discipline, most EA schemes implicitly assume that enterprise architectures, like their physical cousins, can be built from blueprints. But they are not because enterprises have no “Pause” and “Reset” buttons: business cannot be put on stand-by and must be carried on while work is in progress.
Systems & Enterprises
Systems are variously defined as:
“A regularly interacting or interdependent group of items forming a unified whole” (Merriam-Webster).
“A set of connected things or devices that operate together” (Cambridge Dictionary).
“A way of working, organizing, or doing something which follows a fixed plan or set of rules” (Collins Dictionary)
“A collection of components organized to accomplish a specific function or set of functions” (TOGAF from ISO/IEC 42010:2007)
While differing in focus, most understandings mention items and rules, purpose, and the ability to interact; none explicitly mention social structures or interactions with humans. That suggests where the line should be drawn between systems and enterprises, and consequently between corresponding architectures.
Architectures & Changes
Enterprises are live social entities made of corporate culture, organization, and supporting systems; their ultimate purpose is to maintain their identity and integrity while interacting with environments. As a corollary, changes cannot be carried out as if architectures were just apparel, but must ensure the continuity and consistency of enterprises’ structures and behaviors.
That cannot be achieved by off-soil schemes made of blueprints and step-by-step processes detached from actual organization, systems, and processes. Instead, enterprise architectures must be grown bottom up from actual legacies whatever their nature: technical, functional, organizational, business, or cultural.
Insofar as enterprise architectures are concerned, legacies are usually taken into account through one of three implicit assumptions:
No legacy assumptions ignore the issue, as if the case of start-ups could be generalized. These assumptions are logically flawed because enterprises without legacy are like embryos growing their own inherent architecture, and in that case there would be no need for architects.
En Bloc legacy assumptions take for granted that architectures as a whole could be replaced through some Big Bang operation without having a significant impact on business activities. These assumptions are empirically deceptive because, even limited to software architectures, Big Bang solutions cannot cope with the functional and generational heterogeneity of software components characterizing large organizations. Not to mention that enterprise architectures are much more that software and IT.
Piecemeal legacies can be seen as the default assumption, based on the belief that architectures can be re-factored or modernized step by step. While that assumption may be empirically valid, it may also miss the point: assuming that all legacies can be dealt with piecemeal rubs out the distinction pointed above between systems and enterprises.
So, the question remains of what is to be changed, and how ?
EA as a Work In Progress
As with leopard’s spots and identity, the first step would be to set apart what is to change (architectures) from what is to carry on (enterprise).
Maps and territories do provide an overview of spots’ arrangement, but they are static views of architectures, whereas enterprises are dynamic entities that rely on architectures to interact with their environment. So, for maps and territories to serve that purpose they should enable continuous updates and adjustments without impairing enterprises’ awareness and ability to compete.
That shift from system architecture to enterprise behavior implies that:
The scope of changes cannot be fully defined up-front, if only because the whole enterprise, including its organization and business model, could possibly be of concern.
Fixed schedules are to be avoided, lest each and every unit, business or otherwise, would have to be shackled into a web of hopeless reciprocal commitments.
Different stakeholders may come as interested parties, some more equal than others, possibly with overlapped prerogatives.
So, instead of procedural and phased approaches supposed to start from blank pages, EA ventures must be carried out iteratively with the planning, monitoring, assessment, and adjustment of changes across enterprises’ businesses, organizations, and systems. That can be represented as an extension of the OODA (Observation, Orientation, Decision, Action) loop:
Actual observations from operations (a)
Data analysis with regard to architectures as currently documented (b).
Changes in business processes (c).
Changes in architectures (d).
Moreover, due to the generalization of digital flows between enterprises and their environment, decision-making processes used to be set along separate time-frames (operational, tactical, strategic, …), must now be weaved together along a common time-scale encompassing internal (symbolic) as well as external (actual) events.
It ensues that EA processes must not only be continuous, but they also must deal with latency constraints.
Changes & Latency
Architectures are by nature shared across organizational units (enterprise level) and business processes (system level). As a corollary, architecture changes are bound to introduce mismatches and frictions across business-specific applications. Hence the need of sorting out the factors affecting the alignment of maps and territories:
Elapsed time between changes in territories and maps updates (a>b) depends on data analytics and operational architecture.
Elapsed time between changes in maps and revised objectives (b>c) depends on business analysis and organization.
Elapsed time between changes in objectives and their implementation (c>d) depends on engineering processes and systems architecture.
Elapsed time between changes in systems and changes in territories (d>a) depends on applications deployment and technical architectures.
On that basis it’s possible to define four critical lags:
Operational: data analytics can be impeded by delayed, partial, or inaccurate feedback from processes.
Mapping: business analysis can be impeded by delays or discrepancies in data analytics.
Engineering: development of applications can be impeded by delays or discrepancies in business analysis.
Processes: deployment of business processes can be impeded by delays in the delivery of supporting applications.
These lags condition the whole of EA undertakings because legacy structures, mechanisms, and organizations are to be continuously morphed into architectures without introducing misrepresentations that would shackle activities and stray decision-making.
EA Latency & Augmented Reality
Insofar as architectural changes are concerned, discrepancies and frictions are rooted in latency, i.e the elapsed time between actual changes in territories and the updating of relevant maps.
As noted above, these lags have to be weighted according to time-frames, from operational days to strategic years, so that the different agents could be presented with the relevant and up-to-date views befitting to each context and concerns.
That could be achieved if enterprises architectures were presented through augmented reality technologies.
Compared to virtual reality (VR) which overlooks the whole issue of reality and operates only on similes and avatars, augmented reality (AR) brings together virtual and physical realms, operating on apparatuses that weaves actual substrates, observations, and interventions with made-up descriptive, predictive, or prescriptive layers.
On that basis, users would be presented with actual territories (EA legacy) augmented with maps and prospective territories.
Composition and dynamics of maps and territories (actual and prospective) could be set and edited appropriately, subject to latency constraints.
Self-driving vehicles should behave like humans, here is how to teach them so.
The eventuality of sharing roads with self-driven vehicles raises critical issues, technical, social, or ethical. Yet, a dual perspective (us against them) may overlook the question of drivers’ communication (and therefore behavior) because:
Contrary to smart cars, human drivers don’t use algorithms.
Contrary to humans, smart cars are by nature unethical.
If roads are to become safer when shared between human and self-driven vehicles, enhancing their collaboration should be a primary concern.
Driving Is A Social Behavior
The safety of roads has more to do with social behaviors than with human driving skills, as it depends on human ability, (a) to comply with clearly defined rules and, (b) to communicate if and when rules fail to deal with urgent and exceptional circumstances. Given that self-driving vehicles will have no difficulty with rules compliance, the challenge is their ability to communicate with other drivers, especially human ones.
What Humans Expect From Other Drivers
Social behavior of human drivers is basically governed by clarity of intent and self-preservation:
Clarity of intent: every driver expects from all protagonists a basic knowledge of the rules, and the intent to follow the relevant ones depending on circumstances.
Self-preservation: every driver implicitly assumes that all protagonists will try to preserve their physical integrity.
As it happens, these assumptions and expectations may be questioned by self-driving cars:
Human drivers wouldn’t expect other drivers to be too smart with their interpretation of the rules.
Machines have no particular compunction with their physical integrity.
Mixing human and self-driven cars may consequently induce misunderstandings that could affect the reliability of communications, and so the safety of the roads.
Why Self-driving Cars Have To Behave Like Human Drivers
As mentioned above, driving is a social behavior whose safety depends on communication. But contrary to symbolic and explicit driving regulations, communication between drivers is implicit by necessity; if and when needed, it is in urgency because rules are falling short of circumstances: communication has to be instant.
So, since there is no time for interpretation or reasoning about rules, or for the assessment of protagonists’ abilities, communication between drivers must be implicit and immediate. That can only be achieved if all drivers behave like human ones.
Turing’s Imitation Game Revisited
Alan Turing designed his Imitation Game as a way to distinguish between human and artificial intelligence. For that purpose a judge was to interact via computer screen and keyboard with two anonymous “agents”, one human and one artificial, and to decide which was what.
Extending the principle to drivers’ behaviors, cars would be put on the roads of a controlled environment, some driven by humans, others self-driven. Behaviors in routine and exceptional circumstances would be recorded and analyzed for drivers and protagonists.
Control environments should also be run, one for human-only drivers, and one with drivers unaware of the presence of self-driving vehicles.
Drivers’ behaviors would then be assessed according to the nature of protagonists:
H / H: Should be the reference model for all driving behaviors.
H / M: Human drivers should make no difference when encountering self-driving vehicles.
M / H: Self-driving vehicles encountering human drivers should behave like good human drivers.
Ma / Mx: Self-driving vehicles encountering self-driving protagonists and recognizing them as such could change their driving behavior providing no human protagonists are involved.
Ma / Ma: Self-driving vehicles encountering self-driving protagonists and recognizing them as family related could activate collaboration mechanisms providing no other protagonists are involved.
Such a scheme could provide the basis of a driving licence equivalent for self-driving vehicles.
Self-driving Vehicles & Self-improving Safety
If self-driving vehicles have to behave like humans and emulate their immediate reactions, they may prove exceptionally good at it because imitation is what machines do best.
When fed with data about human drivers behaviors, deep-learning algorithms can extract implicit knowledge and use it to mimic human behaviors; and with massive enough data inputs, such algorithms can be honed to statistically perfect similitude.
That could set the basis of a feedback loop:
A limited number of self-driving vehicles (properly fed with data) are set to learn from communicating with human drivers.
As self-driving vehicles become better at the imitation game their number can be progressively increased.
Human behaviors improve influenced by the growing number of self-driving vehicles, which adjust their behavior in return.
That is to create a virtuous feedback for roads safety.
PS: A Contrary Evidence
A trial in Texas demonstrates a contrario the potency of the argument by adopting the alternative policy, making clear that self-driving vehicles are indeed machines. Apart for being introduced as a public transport within a defined area and designed stops, two schemes are used to inhibit the imitation game:
Appearance: design and color enable immediate recognition.
Communication: external screens are used for textual (as opposed to visual) notification to pedestrians and other vehicles.
Future will tell if that policy is just a tactical step or a more significant shift towards a functional distinction between artificial brains and humans ones.
“Clocks slay time… time is dead as long as it is being clicked off by little wheels; only when the clock stops does time come to life.”
The melting of digital fences between enterprises and business environments is putting a new light on the way time has to be taken into account.
The shift can be illustrated by the EU GDPR: by introducing legal constraints on the notifications of changes in personal data, regulators put systems’ internal events on the same standing as external ones and make all time-scales equal whatever their nature.
Ontological Limit of WC3 Time Recommendation
The W3C recommendation for OWL time description is built on the well accepted understanding of temporal entity, duration, and position:
While there isn’t much to argue with what is suggested, the puzzle comes from what is missing, namely the modalities of time: the recommendation makes use of calendars and time-stamps but ignores what is behind, i.e time ontological dimensions.
Out of the Box
As already expounded (Ontologies & Enterprise Architecture) ontologies are at their best when a distinction can be maintained between representation and semantics. That point can be illustrated here by adding an ontological dimension to the W3C description of time:
Ontological modalities are introduced by identifying (#) temporal positions with regard to a time-frame.
Time-frames are open-ended temporal entities identified (#) by events.
It must be noted that initial truth-preserving properties still apply across ontological modalities.
Conclusion: OWL Descriptions Should Not Be Confused With Ontologies
Languages are meant to combine two primary purposes: communication and symbolic representation, some (e.g natural, programming) being focused on the former, other (e.g formal, specific) on the latter.
The distinction is somewhat blurred with languages like OWL (Web Ontology Language) due to the versatility and plasticity of semantic networks.
That apparent proficiency may induce some confusion between languages and ontologies, the former dealing with the encoding of time representations, the latter with time modalities.