Decentralized nation, solving the web identity crisis

Abstract

The web of today whether you prefer to call it web 2.0, web 3.0, web 5.0 or even the metaverse is at a critical stage of evolution and challenge, largely centered around its crisis of identity. Like teenagers who cannot assess properly their reason for being and do not seem ready to take responsibility for their actions, we are constantly blaming the very system we are trying to get away from. To truly realize the benefits from innovation and technology, this crisis has to be resolved, not just through tactical solutions but through developments that enhance the sustainability of the web and its benefits. Significant strides are being made in the evolution of digital services enabled by technology, regulation, and the sheer pace of societal change. The journey to the decentralized web is mirroring the convergence of the physical and digital worlds across all economies and is increasingly embracing the digital native world. Technology has provided the foundational platform for individuals and entities to create and manage wealth, potentially without the need for big institutions. Ironically, despite all of the advancements, we are still facing an unprecedented and increasing wealth gap. Clearly, the system is broken, not just around the edges but at the very core of the democratic underpinning of our society. In this whitepaper, we propose how artificial intelligence on blockchain can be used to generate a new class of identity through direct human computer interaction. We demonstrate how this, combined with new perspectives for sustaining community and governance embedded within the use of blockchain technology, will underpin a sustainable solution to protect identity, authorship and privacy at the same time while contributing to restore trust amongst members of a future decentralized nation and hence contribute to solving the web most significant identity crisis.

Full text

A decentralized nation beyond the centralized web | October 2022
Decentralized Nation: Solving the Web’s Identity Crisis
“In the digital world, the most valuable things I have are
who I am – my identity – and what I do – my reputation!”
Frederic A. Jumelle1, Timothy J. Pagett2, Ryan S. Lemand3
Ydentity Organization1,2, Neovision Wealth Management3
{f.jumelle1,t.pagett2}@ydentity.org ryan.lemand@neovision-wealth.com3
Abstract
The Web of today whether you prefer to call it Web 2.0, Web 3.0,
Web 5.0 or even the metaverse is at a critical stage of evolution
and challenge – largely centered around its crisis of identity. Like
teenagers who cannot assess properly their reason for being and
do not seem ready to take responsibility for their actions, we are
constantly blaming the very system we are trying to get away
from. To truly realize the benefits from innovation and
technology – this crisis has to be resolved – not just through
tactical solutions but through developments that enhance the
sustainability of the web and its benefits.
Significant strides are being made in the evolution of digital
services enabled by technology, regulation, and the sheer pace of
societal change. The journey to the decentralized Web is
mirroring the convergence of the physical and digital worlds
across all economies and is increasingly embracing the digital
native world. Technology has provided the foundational platform
for individuals and entities to create and manage wealth,
potentially without the need for big institutions. Ironically,
despite all of the advancements, we are still facing an
unprecedented and increasing wealth gap. At the core of this, we
believe, is the imbalance of power – a physical world problem
that has now been conveniently replicated in a digital version of
that world representing a more pervasive threat to our democratic
existence much more so than that we faced within our existing
physical world.
Clearly, the system is broken – not just around the edges but at
the very core of the democratic underpinning of our society.
In this whitepaper, we propose how artificial intelligence on
blockchain can be used to generate a new class of identity through
direct human-computer interaction. We demonstrate how this,
combined with new perspectives for sustaining community and
governance embedded within the use of blockchain technology,
will underpin a sustainable solution to protect identity, authorship
and privacy at the same time while contributing to restore trust
amongst members of a future decentralized nation and hence
contribute to solving the Web’s most significant identity crisis.
Keywords: Web2, 3, 5; identity crisis; human rights; civil and
political rights; human-computer interaction; digital democracy;
decentralized nation; social graph; nested GNN
Decentralized Nation:
Solving the Web’s Identity Crisis
From birth, each individual has the right to an identity. The
identity of an individual is the assertion of their existence in a
society. It is also a matter of recognition of their individuality and
what differentiates them from their peers. Having an identity is a
fundamental human right which allows each individual the ability
to enjoy all of their rights. [1] Exercising and maintaining identity
as a fundamental human right is a critical challenge in
cyberspace balancing the need for a digital version of identity that
allows access to technology and the services provided by
technology to mediate relationships.
Early generations of digital identity do little more than capture a
person’s identity dimensions as dictated by the information
required to support Know Your Customer expectations (name,
address, education, credentials etc.) driven by regulators,
marketers and product developers. This so called Personal
Identifiable Information (PII) has increasingly become the
subject of regulation and protection from abusive practices that
are particularly prevalent within the web. Our next generation
identity technology, Ydentity ID, captures signals directly from
the user without the need for capturing so called PII and computes
the attributes of this natural person to define and serve them better.
In this paper, we will explain to you why you need to act now and
think differently to remain relevant in the future. A new digital
democracy is coming – with identity at the core. We believe
that Ydentity technology gives users the right to redefine civil
society [2] in the digital form of a decentralized nation. Through
connectivity and empowerment, we can together recreate the
original democracy we always wanted to be part of, and the Web
will rise again.
1. Legal Identity
A critical place to start any consideration of sustainable and true
identity in cyberspace is with the basic principles that underpin
identity from a legal perspective.
1.1 Identity of a natural person
This right to an identity is enshrined in Article 6 of the Universal
Declaration and reiterated by QHRC Section 15 of the Human
Rights Act 2019:
"Everyone has the right to recognition everywhere as a
person before the law." [3]
From birth, each individual has the right to an identity. The
identity of an individual is the assertion of his or her existence in
a society. It is also a matter of recognition of their individuality
and what differentiates them from their peers.
Having an identity is a fundamental human right which allows
each individual the ability to enjoy all of their rights.
Identity encompasses the family name, the surname, date of birth,
gender and nationality of the individual. Through these details, an
individual will hold rights and obligations specific to their status
(woman, man, child, handicapped, refugee, etc.).
From birth, each individual has the right to have a name and a
surname. In most countries parents have a duty to declare the
name, the surname and date of birth of a newborn to the
authorities in charge. By recording this birth, the State officially
recognizes the existence of the child and formalizes their status in
the eyes of the law. In addition, through this name and recording
in the Registration of Births and Deaths, the child will be able to
establish filiations, that it is to say links of blood relations linking
the child to their father and/or mother.
From birth, the child also has the right to a nationality.
Nationality can be obtained in two different ways:
• Jus sanguinis (by blood): the child will have the same
nationality as his/her parents.

A decentralized nation beyond the centralized web | October 2022
• Jus soli (by birth): The child will have the nationality of
the territory on which he/she was born, even if his/her
parents have a different nationality.
Nationality is confirmed through the issuing of a birth certificate.
It is an important aspect of a person’s life, as it is an attribute of
citizenship. Nationality allows establishment of the affiliation for
an individual to a nation.
1.2 Management of individuals’ identities
Separate from an individual’s right to an identity is the concept of
state authority over its citizens. Every country has some form of
registering an individual as part of its management and control
over society whether through birth certificates, passports, ID
cards, national insurance numbers or any combination of the
above.
This confirms an individual’s status as a citizen of the relevant
jurisdiction but also serves the societal function of allowing a
state to track its citizens’ activities through their actions and
financial transactions. The tension between the individual’s right
to freedom and privacy and the state’s reasonable requirement to
monitor its citizens for security, protection and other purposes is
at the heart of one of the most important philosophical debates of
our time.
Currently, more than half of the world’s countries operate a
National Identity Card System. Some countries use biometric ID
cards, others have biometric passports. Currently India has one of
the most advanced and comprehensive systems of national ID
called Aadhaar. [4]
1.3 Refugee/migrant problem
World Bank ID4D Global Dataset 2017 suggests that 1.1 billion
people worldwide do not possess sufficient documentation or
transactional evidence to prove their own identity to a level
sufficient to be acceptable for regulatory standards.
This gives rise to economic and social consequences for the
person lacking a formal identity as formal identification is usually
required to access healthcare, education and financial institutions
as well as for gender equality. Similarly, a proper identity register
is necessary for governments to provide social welfare services.
It is widely considered to be one of the greatest contributors to
continued financial exclusion.
In 2015, the UN adopted the 2030 Sustainable Development
Goals and Goal 16.9 [5] sets the provision that by that year the
UN shall provide legal identity for all persons including free birth
registrations.
This is widely recognized as a challenging commitment with
debate continuing at pace across and between many signatories as
to what constitutes legal identity. There is some hope that
digitization may provide the basis for accelerating the
achievement of this objective – in the form of a hybrid sovereign
identity solution. However, skepticism remains as to the true
intention of a universally applied identity – be it digital, hybrid or
physical.
1.4 Digital identity
A critical challenge in cyberspace is knowing with whom one is
interacting. Using static identifiers such as passwords and email
there is no way to precisely determine the identity of a person in
a digital space, largely because this information can be stolen,
manipulated or used by many individuals acting in concert or as
one. To overcome this, digital identity has evolved based on
dynamic entity relationships that can be captured from behavioral
history across multiple websites and mobile apps can verify and
authenticate an identity with high accuracy.
By comparing a set of entity relationships between a new event
(e.g., login) and past events, a pattern of convergence can verify
or authenticate the identity as legitimate where divergence
indicates an attempt to mask an identity. Data used for this type
of digital identity is generally anonymized using a one-way hash,
thereby avoiding privacy concerns. Because it is based on
behavioral history, a digital identity is nearly impossible to fake
or steal.
A digital identity is information on an entity used by computer
systems to represent an external agent. That agent may be a
person, organization, application, or device. ISO/IEC 24760-1
defines identity as "set of attributes related to an entity". [6]
The information contained in a digital identity allows for
assessment and authentication of a user interacting with a
business or social ecosystem on the web, without the involvement
of human operators. Digital identities allow our access to
technology and the automated services they provide and make it
possible for technology to mediate all forms of personal and
business relationships.
The term "digital identity" has also come to denote aspects of
civil and personal identity that have resulted from the widespread
use of identity information to represent people and their actions
in digital form. Digital identity is now often used in ways that
require data about persons stored in digital form to be linked to
their civil, or national, identities. Furthermore, the use of digital
identities is now so widespread that many discussions refer to
"digital identity" as the entire collection of information generated
by a person or entity’s online activity.
This includes usernames and passwords, online search activities,
birth date, social security, and purchasing history. Especially
when that information is publicly available and not anonymized
and can be used by others to discover that person's civil identity.
In this wider sense, a digital identity is a version, or facet, of a
person's social identity. This may also be referred to as an online
identity.
The legal and social effects of digital identity are complex and
challenging. However, they are simply a consequence of the
increasing use of technology, and the need to provide technology
with information that can be used to identify external agents.
ID2020 [7] considers the core features and advantages of digital
ID to be as follows:
• personal: unique to the owner and the owner only.
• persistent: lives with the owner from birth to death.
• private: only the owner can control his/her own identity,
and the owner can choose what to share and with whom.
• portable: accessible anywhere the owner happens to be
through multiple form-factors.
Digital identity fundamentally requires digital identifiers –
strings or tokens that are unique within a given scope (globally
or locally within a specific domain, community, directory,
application, etc.).
Identifiers are the key used by the parties to an identification
relationship to agree on the entity being represented. Identifiers
may be classified as omnidirectional and unidirectional.
Omnidirectional identifiers are intended to be public and easily
discoverable, while unidirectional identifiers are intended to be
private and used only in the context of a specific identity
relationship.
Identifiers may also be classified as resolvable or non-resolvable.
Resolvable identifiers, such as a domain name or email address,
may be dereferenced into the entity they represent, or some
current state data providing relevant attributes of that entity.
Non-resolvable identifiers, such as a person's real-world name, or
a subject or topic name, can be compared for equivalence but are
not otherwise machine understandable.
There are many different schemes and formats for digital
identifiers. The most widely used is Uniform Resource Identifier
(URI) and its internationalized version Internationalized
Resource Identifier (IRI), the standard for identifiers on the World
Wide Web.

A decentralized nation beyond the centralized web | October 2022
Lightweight Identity (LID) and OpenID are web authentication
protocols that use standard HTTP URIs (often called URLs), for
example. [8] [9]
1.5 Privacy and security
Businesses and governments, alike, fight for control over access
to user information in order to maximize advertising dollars,
maintain political power, optimize their positioning within, and
sometimes even monopolize, marketplaces. Until the rise of the
sharing economy platforms, large traditional businesses appeared
to maintain a complete monopoly in a variety of industries, such
as tourism, hospitality, retail and personal banking, investing and
capital markets, and taxi/ride hailing services.
Following the revelations of Edward Snowden on mass
surveillance by the NSA [10] in 2013 and the Facebook-
Cambridge Analytica data collection scandal [11] in 2018, many
businesses have, through a combination of peer, regulatory and
social pressure, shifted power back toward the consumer
through allowing individuals to assume the role of these entities
in peer-to-peer transactions. Already, the sharing economy has
seen extremely rapid growth and will continue to become a part
of everyday life.
Despite the disruption the sharing economy has generated, there
is still significant control over wide-spread user information in the
hands of much fewer private businesses. While the sharing
economy has helped to reduce the power of large, traditional
businesses, individuals must still place a lot of trust in the security
promised by the organizations who design, create and build any
given sharing platform. New technologies continue to emerge,
however, affording more and more individuals even greater and
more secure control over their own data and digital presence.
1.6 Reputation services and the blockchain
Usually, user ratings and reviews are tied to a single platform
which owns their content and individuals have to start from
scratch if they want to build their reputation on another sharing
platform. Reputation response can be set for email identifiers and
used by a reputation services provider. [12] Thus, third party
platforms afford users greater control over their data by allowing
them to utilize their entire digital reputation as leverage to prove
their trustworthiness across peer-to-peer marketplaces.
Users can voluntarily connect their social media and sharing
platform accounts in order to use their trust profile which is based
on the ratings and reviews they have received on connected
accounts. All of this happens when users sign up to a new
platform and third-party technology extends the sign-up flow
with a seamless integration where users can build their trust
profile.
Mixing user identity and rating services with blockchain
technology could result in an even better final product.
As mentioned above, ratings and reviews are usually related to a
specific service and are owned by the company/person behind
it. The structure of the current digital reality is based on clusters
of information, made available by large companies.
Online existence has historically been dependent on the access
provided by centralized services, and our resulting digital
identities and reputation are still owned by a handful of
organizations that provide these services.
We pay them with information about ourselves, oblivious to how
that information is actually used (and by whom).
It is probably fair to say that the organizations that have truly
monetized the value of data held in digital form are the BigTech
businesses that control access to technology.
However, in a move that could potentially challenge the business
model of data ownership, blockchain (or distributed ledger)
technology provides an alternative way of storing information
without the need of a central authority whilst still serving and
connecting users.
As an example, the Bitcoin blockchain stores information about
the balance of each Bitcoin address.
The same approach could be applied to many other types of
information. Data about anything that could be represented
digitally can be stored on a blockchain and made accessible by
anyone with the proper decipher tool. Immutable, public and, at
the same time, encrypted information could be the key to
enhancing identity and reputation solutions.
Solutions that provide users with the ability to have improved
control over their digital privacy and the ability to sharing that
personal information only when they choose to or are legally
required to are a critical need as Web 3.0 evolves.
1.7 Ydentity, a decentralized ID
Traditional forms of digital identity generally do little more than
capture the necessary components required to support compliance
with regulatory requirements, to support models and analytics
driven by digital platforms and ecosystems or to aid in social
connectivity. Most, if not all comprise digital representations of
physical documentation (e.g. name, address, education,
credentials).
Ydentity’s creator, Dr Frederic Jumelle, considers that this does
not capture the attributes of a natural person therefore does not
define nor serve the person anywhere near as well as it serves the
entity requiring the identity.
Ydentity uses a method and system for neuropsychological
performance testing based on a self-interview on portable
device during which device sensors capture the user’s cognitive
and emotional responses to a selection of 30 questions selected
randomly from a proprietary database of over 600 questions. The
Artificial Intelligence (AI) comprises a cellular architecture with
several artificial neural networks for processing the signals
captured by sensors during the interview. Ydentity uses an edge
computing approach which includes the preselection and
preprocessing of most informative data directly on the device and
the destruction of all materials after processing. AI modules have
been pre-trained using international human emotions datasets and
will benefit of online learning when new profiles are generated.
Each user’s metadata file comprises 7 scores (tolerance,
credibility, maturity, autonomy, emotional state, worthiness, and
w-range in reference to the tested population) along with 5
demographics (age, gender, ambition, job level and education
level) and their YDR reputation level.
A record of the user’s 128-dim matrix of facial features is also
uploaded onchain to allow an automatic matchmaking between
the record and the user’s face at login preventing duplicate and
deep fakes. Each Ydentity ID can be used to check identity and
authenticate, authorize, certify and upload/download documents
on IPFS, or create a digital twin for risk management and protect
the rights of the user/owner across various blockchains.
The Ydentity proprietary login system includes 2 layers of anti-
spoofing technology based on video capture using Multi-task
Cascaded CNN for gatekeeping unique access to the dApp and
wallet of the user, making it a Self-Sovereign and Decentralized
Identity. [13]
1.8 The Ydentity.org Association
The Ydentity.org Association was created in Switzerland, Canton
de Genève, [14] to guarantee and protect the right of every natural
person or citizen to obtain, own and use a Ydentity non-fungible
token ID.
While the Ydentity dApp enables users to self-generate an
individual Ydentity based on signal capture and processing of the
attributes of each user by an artificial intelligence unit, the
association is lobbying to:
• facilitate the use of such an Alt-ID by every citizen for
identification and authentication during internet
navigation; and

A decentralized nation beyond the centralized web | October 2022
• create and maintain a database of active members that
can represent the world population in terms of diversity
to support the development and generalization of Alt-ID
in the fields of online identification of persons, KYC,
and risk management oracle services; and
• warrant the principle of neutrality and independence
related to individual identity.
The Association is an association of stakeholders and followers
and has no profit purpose.
The Association may pursue all lawful activities to achieve its
purpose. In particular, the Association may undertake the active
and international recruitment of its members to meet its purpose;
advertisement and development of a marketing strategy for its
purpose; the formation of a lobby to influence governments, non-
profit organizations and other associations in relation to its
purpose and implementation.
1.9 The right to own a decentralized ID
A Ydentity token ID is a non-fungible-token that can be
generated on decentralized applications deployed on the
Ethereum and Rootstock blockchains. Owners are able to use it to
identify, communicate and interact with one another.
A Ydentity token ID can be used to create a unique decentralized
Identity Document that is not transferable and cannot be traded.
It contains a PII-free file (without Personally Identifiable
Information) that is uploaded on-chain when minted. This on-
chain file is non-encrypted and can be searched on the dApp by
other Ydentity users. The result of a search is a list of token
addresses that can be used to connect via secret email (ex:
ethmail.cc) and build a Ycommunity. It can also be used by third
parties for other purposes such as statistics, polls, decision-
making protocols, etc.
Each Ydentity token is held in its owner’s wallet along with
Ydentity reputation tokens (YDR).
If a Ydentity ID owner decides to burn his/her Ydentity ID, there
will be consequences resulting from this action up to being
removed from the Ycommunity or the DAO in certain cases.
Reinstatement would be possible but subject to a new full
registration and Network Governance approval.
1.10 YDR, the new currency of trust
If identity is who I am then reputation is what I do – with my
identity.
Each Ydentity ID owner can earn reputation in the form of
Ydentity reputation tokens (YDR). [15] These ERC20 tokens are
earned when owners share their unique Ydentity signature (proof-
of-identity) for example when underwriting transactions. The
number of YDR tokens awarded for a transaction will depend on
the smart contract underpinning the transaction especially the
definition of the criteria of satisfaction between parties.
Ydentity ID owners will find ways to spend their YDR in the
DAO for example while searching for peers and building
Ycommunities or joining certain focus groups, clubs or buying
privileges to invest in complex products or borrowing cryptos or
gaining access to certain social graphs.
A certain level of YDR tokens will be necessary to join some
groups but also multiply the gain from these groups. In the case
where a Ydentity ID owner behaves poorly, YDR tokens can be
lost or cancelled, and the reputation level of this individual will
decrease. Because of this loss of status, the individual can be
denied access to certain groups or services until they regain
tokens. Reinstatement will be decided by Network Governance
especially in cases where the reputation level has reached zero.
2. Ydentity Technology
2.1 Ydentity (NFT) token ID
Ydentity is a non-fungible token identity that stores the
person/owner’s attributes on a digital ledger or blockchain which
certifies that this digital asset, the user’s Ydentity, is a unique
decentralized identity (DID) [16] and a self-sovereign identity
(SSI). [17][18]
Each Ydentity token ID is by definition unique, non-fungible,
ownable, non-transferable and burnable. It tokenizes the physical,
cognitive and affective attributes of a person captured
from/directly on the user by himself/herself during a self-
administered and timed Video Emotion Recognition-based
Interview or VERI.
The Ydentity DAO stage will be a platform offering members the
chance to mint their Ydentity IDs and start interacting with other
holders, building groups, clusters and Ycommunities, to earn and
spend YDR in numerous dApps deployed by the Ydentity
Foundation and other partners of the Foundation licensing the
Ydentity technology. The DAO can also invite developers to
build dApps for the DAO, improve its algorithms, accept
challenges for which they receive grants in the form of YDR
tokens.
2.2 Ydentity in a decentralized economy
Ydentity ID will become the key element of risk and decision
management of a decentralized economy. Risk and decision
management are important parts of both an individual’s and a
corporation’s daily existence. We all practice risk and decision
management to varying degrees and for corporations’ assessment
and mitigation of risk has been performed for decades. Entities,
whether individuals or corporations, have been classified in broad
categories regarding their relation to risk. Risk managers call
upon these categories throughout their career when assessing risk,
mitigating risk, selling products and the acceptance of the risk in
the first place.
Introducing a new type of attribute-based digital identity with a
stack of scores available for virtual interactions is a must that
could outsmart the system of physical and digital identity cards.
Similarly to a government-issued ID, all kind of rights can be
attached to a Ydentity ID but conversely Ydentity can keep a log
of background history and offer legacy rights to the owner(s).
Developing the technology where users can build their own
attribute-based identity, in the form of an immutable digital file,
own it to earn and trade while sharing only a unique signature was
a necessary step to protect the uniqueness of one’s identity.
An NFT is a non-fungible unit of data stored on a blockchain that
can offer a representation of the user during web sessions. Users
can earn reputation tokens (fungible) during transactions since
trust is removed from the equation, only reputation matters.
The Ydentity profiling framework using a Video Emotion
Recognition Interview [19][20] was developed in partnership
with a research team of the Hong Kong University of Science and
Technology – Neuromorphic Interactive System Lab, and is
constantly upgraded to allow users to self-assess and become their
own oracle to predict the effect of their (potential) decisions on
communities; or evaluate the systemic effect of their decisions on
platforms, graphs or markets; or on pricing of products they want
to trade in before taking the decisions; or evaluate co-investing
risks and the implication of sharing risks on certain products or
services before, during and after taking decisions.
Blockchain oracle services currently include many prominent
projects and numerous decentralized finance (DeFi) dApps that
could work with Ydentity API and benefit from the Ydentity
technology at large while using onchain data as first party oracle
for risk management services.
2.3 Neuropsychological testing at the core
The technology contributing to the generation of Ydentity ID is a
patented method and system for neuropsychological performance
test. [21][22][23][24] It is based on a terminal device used to
interact with a cloud server which only stores user processed data
and is logged into by the users through their terminal device. A
test module comprises the user information, which is stored in the
cloud server or can be downloaded from the cloud server and is

A decentralized nation beyond the centralized web | October 2022
directly accessed through said terminal device and is trained by
the artificial neural network. User information comprises user
performance metrics and the terminal device can display the
neuropsychological performance test results before they are
minted to a non-fungible token ID.
Ydentity technology can be applied to pre-screening, remote
screening and onboarding of human resources; sorting of personal
accounts, fraud prevention and forensics for social media;
matchmaking in client relation management and dating;
onboarding and remote onboarding of new customers in the
financial services industry for compliance Know Your Client
( “KYC” ) or Customer Due Diligence ( “CDD” ) regulations and
also any new virtual services to persons including providing smart
ID for smart cities. Compatible with other identification and
identity authentication/verification technologies, Ydentity
technology allows the creation of true personal identity by
using personal metrics with a high degree of accuracy and
security. Ydentity technology enables some of the biggest
challenges of the internet such as the large amount of fake
accounts including the fake social media accounts that can pose a
threat to society to be easily solved.
2.4 Recognizers on edge to enhance privacy
Based on a grant for pioneer research titled “Bandwidth Aware
Video Emotion Recognition in the Cloud”, we are developing
new recognizers for the Ydentity dApp frontend that can enable
edge computing of the captured signals and guarantee privacy in
5G or 6G environments.
Our arousal network (shallow neural network) can pick up the key
frames in the interview package and deliver the total emotion
metric. It is more effective to provide measurements of emotional
states that are more finely grained than discrete categories. It also
works from video input.
The team at Hong Kong University of Science and Technology
has an established track record in applying deep neural network
(DNN) technology to address this area (Deng et al., 2020; Deng,
Chen & Shi, 2020; Zhou, Pi & Shi, 2017; Zhou & Shi, 2017).
[25][26]
Deep neural networks developed by the team have yielded state
of the art performance on international benchmarks, e.g. winning
international challenges, such as the 2017 Facial Expression
Recognition and Analysis (FERA) AU Intensity Estimation
Challenge (Zhou, Pi & Shi, 2017) and the Valence-Arousal and
the Action Unit Tasks in the 2020 Affect Recognition in-the-wild
Challenge (Deng, Chen &Shi, 2020). [27] Optimization also
includes Iterative Distillation for Better Uncertainty Estimates in
Multitask Emotion Recognition. [28]
In particular, the team has developed technology that classifies
emotions along the continuously valued dimensions of valence
and arousal, which can achieve any value between -1 and 1, rather
than only seven discrete values (Deng et al., 2020). This allows
for the more fine-grained distinctions required by user profiling.
It has also developed algorithms for Facial Action Unit intensity
estimation (Zhou, Pi & Shi, 2017). Action units are localized
facial muscle movements, which can be used to evaluate and
describe human mental states like depression and happiness. The
problem of intensity estimation seeks to determine the extent to
which a particular action unit is activated, as a number ranging
anywhere from zero to one. This fine-grained estimation provides
more information than just a single binary detection result of
whether or not the action unit is present.
2.5 From third party oracle to first party oracle for third
parties
Much of the evolution of traditional finance has increasingly
involved the development of data rich experience driven models
– in many cases provided through the use of third-party oracles.
Take the provision of lending finance driven by the scores
generated and held by credit reference agencies, where propensity
to repay is inferred by statistical correlations between actions,
observable and collected attributes and outcomes. The same is
true in the behavioral and statistical models that underpin much
of the algorithmic trading that dominates many global investment
and trading markets.
In both of these cases risk management is based on statistics, a
stochastic approach using a limited number of variables that are
captured from markets by sensors and feed agents which apply
algorithmic relationships between input and output.
AI/Machine Learning is being used with varied success to
overcome limitations in statistical using big data (de-personalized
and anonymized data) in sufficiently large volumes, with high
velocity and variety. This brings the benefit of correlations and
the power to take decisions with confidence even though the
accuracy at an individual node level is generally considered
mediocre. This accuracy challenge can be addressed through
AI/ML using personalized data in smaller volumes, but high
accuracy would prove very impactful. However, this would
contradict much of the privacy protection regulation and therefore
cannot be used or is just too expensive to overcome.
In decentralized finance or DeFi, many of the apps that have been
developed are looking to serve the users on a peer-to-peer basis
without the need to challenge privacy regulations and to avoid the
cost of relying on third party anonymized data such as credit
reports. Notwithstanding the way these models are configured
they still rely heavily on historical observations.
We are seeing emerging challenges with the DeFi apps and the
inherent limitations of basing behavioral models on static points
in time actions and attributes rather than dynamic analysis
combined with emotional positioning and response. The latter are
just too expensive to capture and analyze.
Despite the existence of Nobel winning research from the likes of
Daniel Kahneman and Richard H. Thaler that provides
definitive support that emotions play a critical role in driving
decision making, particularly in finance, many of the models in
use in finance applications today, whether traditional, digital or
DeFi, do not consider emotions as an input.
The Ydentity token ID incorporates the emotional component
with the cognitive and timing components of decision making to
provide a unique combination of attributes for each Ydentity
holder. When held in sufficient volumes the population of
Ydentity holders can form the basis for assisting them to “self-
appraise their relative risk level” whilst at the same time as
empowering them to share this for their own benefit such as
getting a better loan rate or accessing a selection of equity funds.
The technology and methodology supporting the Ydentity token
ID, is an effective “first party” input to allow “third party”
indefinite real time risk profiling as well as allowing users to
decide whether to link users’ data to the market and get feedback
for online learning without interruption. Ydentity can be used to
provide risk management data and first party oracle services in
peer-to-peer interaction and applications without the need for
costly intermediaries leading to enhanced profit opportunities.
2.6 Ydentity onchain data for DeFi (by third party dApps
in the DAO)
Ydentity is based on a technology using AI neural networks for
profiling and risk and decision management. In this way, it can be
used as a critical input to a model contributes to predict
customer’s propensity to uptake a product and predict the
correlated risk of certain products. Through the unique and
immutable features of Ydentity, we believe that it could be used
to set up a feedback mechanism to learn directly from the users.
From a reputation point of view, the actions of the user could be
used as a trigger alert for management to take timely mitigation
actions.
Although Ydentity can currently be used to provide highly
accurate predictions based on neural networks, the method and
system (especially the hidden layers or LSTM mechanisms)
remain difficult to explain to the majority of regulators therefore

A decentralized nation beyond the centralized web | October 2022
it cannot currently be used to achieve regulatory compliance and
for reporting to regulators. However, we believe that this will
change as the use of this type of approach grows exponentially on
the Web.
Ydentity has the potential to be used for marketing and decision
making such as business decisions with confidence without
having the need for total accuracy. Because companies’ profile,
user’s risk profiles and markets can evolve, recalibration with
online training of the engines is important in risk management
and is Ydentity’s main feature. AI/ML on personal data enables
first party oracle services provided directly to the user from the
users’ data, standalone or for pooling. Ydentity can contribute
data and be a first party oracle on its own regarding pricing,
underwriting and modelling individual or group decision. It can
possibly target systemic impact and analytics at an advanced
stage.
2.7 First party oracle and sustainable value creation and
transfer
Human decision making applies the decision theory while human
planning applies the theory of choice and game theory.
A first party oracle is fed by the user(s) for self-serving interest
and authorship such as optimization of decision making, planning
and value transfer based on:
• Interaction between decision makers, i.e. co-investing
Ydentity owners.
• Intertemporal choice, the Ydentity owners enter the
Web3 sphere at different times.
• Non-zero-sum game theory applied to Web-based
investment markets where the sum is not zero because
the interaction between decision makers is generating a
new intrinsic value comparable to the enthalpy of
formation which can take the form of a reward in YDR
reputation tokens and other earnings and enable value
transfer. In the opposite direction, penalties can be given
in a form related to the enthalpy of dissolution. This
feature will be developed as a test against a zero-sum
game also called a Ponzi Scheme.
Any mechanism for sustainably creating value from one’s identity
needs to simultaneously consider the mechanism for the ultimate
transfer and realization of that value.
The value of identity in almost all existing Web 2.0 applications
is not based on what you do with your identity – it is more based
on what you did with your identity – and the attributes that “define”
your identity. You are considered more valuable if you did the
things that models/AI said were valuable – borrowed money, paid
it back, bought certain goods and services, accumulated wealth
by a certain age, lived in certain areas, had certain ethnicity, went
to the right schools, got the right job. All attributes that are
dictated and measured from the “center” – controlled by those that
have sufficient data to define what is “right”.
However, consider a digital native world where what you do,
where you go and how you act are only capable of capture in a
digital form and are really only controlled by what you want to
do. In both this and the converging digital and physical world,
digital currencies, crypto currencies, tokenization and fractional
ownership are being progressively recognized as mechanisms for
creating both incentive and reward – providing a new basis for
creating “transferable” or “realizable” value without the need for
traditional monetization.
In this world – what is considered good?
Much of the progressive thinking on this tends toward the creation
of smart contracts that recognize and reward “good” behavior
through granting access of digital rights/assets whilst
simultaneously penalizing bad behavior – largely through loss of
access or digital rights/assets. Through empowered users, whose
positive choices become the self-fulfilling governance
mechanism to weed out and isolate bad actors, the need for
centralized control becomes obsolete.
Reputation – what you do – becomes equally as valuable in a
digital world as who you are.
Add a mechanism to transfer the value of reputation to create
even more value – only by successful contractual actions – one
can never “buy” reputation but only earn, use and lose reputation.
The velocity of usage will provide the momentum for new ways
of working, where you are only rewarded for the contribution that
you make – no entitled wealth creation – just reward for
performance considered through consensus to be of value to the
community. This has the basis to fundamentally change the way
in which reward and incentive mechanisms operate, not just in the
digital native world but in the physical world as well.
This is analogous to the current positioning that copyright and
intellectual property plays in both the physical and digital world.
The recognition of effort – both tangible and intangible – is given
credence through the successful application of “protection” –
allowing for the owner of that protection to not only capture
appreciation of the intangible value of the reputation being
protected, but also through royalties paid for the use of that
intangible value by others. Equate this to a digital native world
and effectively your reputation is the intangible value of how you
act in the digital world. However, in a digital native world there
is no “ready-made” mechanism to protect the value of your
reputation or what you do with your reputation – unless it is
attached to an immutable form of identity that is protected and
controlled by the owner of that identity. A logical extension of
this is then the real future of work, where contribution to work is
proven by the consensus mechanisms inherent within blockchain
technology. When combined with immutable identity there is a
new way of recognizing and rewarding contribution – creating
tangible rewards and intangible value – in the same way, but far
more efficiently, than was created in a physical world. The value
from what you do, not from who you are.
What you do has no value if it is not backed by the authenticity
of who you are. It is critical that identity be simultaneously
verifiable and recognizable. Value transfer mechanisms need to
be based on a better version of identity – one that represents the
convergence of the physical and digital worlds – through human
computer interaction. Not one developed on a static basis, that is
really just a digital representation of a physical form but one that
is a dynamic representation of the core combination of an
individual’s identity – the physical, cognitive and emotional
elements working in concert. Not one that is controlled centrally,
but one created and controlled by the true holder of that identity
– on an immutable basis – for the good of only that holder.
3. Decentralized Nation as a Solution
3.1 Definition of a nation
A nation is a large body of people united by common descent,
history, culture, or language, inhabiting a particular country or
territory, according to Oxford Languages and Google.
In other dictionaries, nation refers to:
• All human beings living in the same territory, having a
community of origin, history, culture, tradition,
sometimes language, and constituting a political
community; or
• Abstract, collective and indivisible entity, distinct from
the individuals who comprise it and the holder of
sovereignty; or
• In biblical literature, “nations refer to pagan peoples, as
opposed to the chosen people.”
3.2 Definition of a decentralized autonomous nation (DAN)
A decentralized nation is likely to be an abstract, autonomous,
collective and indivisible organization distinct from the
individuals who comprise it and who are the holders of
sovereignty.
As a new type of nation pioneering in a field of abstract concepts,
a decentralized nation should look for the best available systems

A decentralized nation beyond the centralized web | October 2022
of political, social and economic governance before inception in
order to avoid the pitfalls and failures of systems that are proven
inefficient, oppressive and misleading. We believe there is a
sufficient number of indicators for such failures to identify what
works and what is not working. The interest of looking for
mathematical formulation of systems is that mathematics allow
computation and computation can automate the rules of execution
of contracts.
3.3 Why do we need one or more DANs?
Let's start with the definition of "Security":
• a state of being free from danger or threat; or
• something deposited or pledged as a guarantee of the
fulfilment of an undertaking or the repayment of a loan,
to be forfeited in case of default.
Are the two definitions equivalent or ambivalent or something
from the past?
Are all financial assets considered "securities" for the sake of the
security of the state which aims to protect its citizens?
A security (definition 2 above) is considered security (definition
1 above) because it is guaranteed by a hard asset, a business that
generates or purports to generate cashflow or value, or anything
that can be sold to recover a certain amount of money in case of
the default of the aforementioned security's issuer, hence where
the name came from. In Finance, we use the word "security" to
generically describe financial instruments such as stocks, bonds,
units in mutual funds, and others. However, in recent years, this
definition has become blurred with the arrival of digital assets,
namely digital coins and cryptocurrencies.
Many have even argued that cryptocurrencies cannot even be
called "assets". In fact, until the arrival of digital assets, in
Finance, the words financial assets and securities were frequently
used interchangeably.
Many others argue that cryptocurrencies cannot be considered as
securities because they fail the Howey test although the Securities
and Exchange Commission (SEC) of the United States position
seems to be that many cryptocurrencies may be securities based
on recent actions they have taken and cases they have brought.
Others have argued that cryptocurrencies are not only assets and
securities and more but most importantly a new way of life.
In fact, for years now, proponents of cryptocurrencies, have
preached that they represent a form of deliverance from political
and financial oppression as well as the end of the surveillance
of our very life by a paranoid "Big Brother".
Cryptocurrencies have become a way out of sorts, a means of
escape if you will, that represents the core beliefs of its users
giving the opportunity to become their identity throughout and
beyond government identification.
Everything that surrounds us and is related to us from a socio-
economic perspective has become increasingly represented in a
digital form: our medical records in the cloud are readily available
to medical practitioners, our credit scores can be consulted by
creditors to verify our financial worthiness, our biometric
passports contain our biometric data, etc.
However, the usage of fiat currency remains flexible and
anonymous if one so chooses. For example, when we use a $20
bill to buy groceries, this transaction will not appear anywhere in
the digital world. Hence one of the reasons for developing loyalty
cards and other point of sale survey questions such as postcode,
phone or name, as a means of collecting information on
transactions and customers. You do not have to provide the
information – to conclude the transaction if it involves cash.
Contrary to this, if the same transaction is made though digital
means, contactless smartphone to terminal or a card whether by
fiat currencies or cryptocurrencies, it will be recorded somewhere
in the digital world and could be traced to a credit card, a wallet,
a bank account and to you the owner or holder.
The paradigm of a digitized nation has already been adopted and
standardized in many developed countries, as well as in some
emerging countries engaged in a rapid digital transformation
linked to the digital identity of their citizens.
These digitized nations are centralized around the core aim to
connect their citizens’ digital identities and data with government
services, medical services, financial institutions, and economic
policy in a way that can make them proactive rather than reactive.
Taken one step further, a state-controlled digital nation can
decide to assign a social score to each digital identity holder and
pretend that is the mirror of the person's actions, behaviors, risk
management and financial management. If this social score also
includes individual medical records and consumption habits of
the person, this person will witness the disappearance of their
personality and privacy entirely, the depth of which has only been
seen in the most somber hours of humanity.
We can safely say that the industrial revolution has given way to
the digital revolution, which started around the year 2000 with the
Internet maturing and becoming the backbone of our society and
it is still ongoing like a maelstrom swallowing everything known
to us from personal to societal, financial, economic, etc.
Where does this leave us?
What will happen to our privacy and individual freedoms if we
continue to trade our socioeconomic identity for the sake of
spurious security?
Is this first version of digital nation simply creating a heard of
sheep that can be steered electronically through push technology?
At Ydentity, we believe that humans need autonomy and freedom
to mature and be creative, to have independence and the stamina
to make the smart and the right decisions.
We think that a truly digital nation cannot be deployed by the
current governments digitizing their control but by their
citizens, not only the citizens of one nation but the citizens of
multiple nations together, transcending the very concept of a
nation that has failed so many times throughout history.
There is a possible way between the paradise of the dreamers and
the hell of a Kafkian society.
It implies laws made for this new generation of DAN wrapped in
legal entities with frameworks designed to guarantee their
existence in the middle of the other moving parts of the coming
new world order.
It is inevitable because these DAN represent hope, the aspiration
to have a better future, to bring the borders down, to avoid war,
to save the planet when others are already thinking to leave Earth
to Mars or wherever they believe they can have a future without
the burden of taking care of we have now in our hands.
3.4 Political system and consensus by Proof-of-Authority
A DAN should make the choice of a political system based on the
weighted arithmetic mean W which allows the users with higher
reputation level to contribute more than the ones with lower level
and exclude religious beliefs and leader centricity from the
system governance. [29][30]
 




 
where W is the weighted average; n is the number of terms to be
averaged; wi are the weights applied to x values; Xi is the data
values to be averaged. A political system allowing weighted vote
will avoid irrationality in the decision-making process that can
integrate with the economy for the purpose of sharing the
common good. [31]
With a DAN operated by an organic Ydentity blockchain where
trust is distributed, Proof-of-Authority (PoA) for consensus [32]
will be adopted by the Ydentity network.
In the Ydentity PoA-based network, transactions and blocks are
validated by approved accounts, known as validators. To become

A decentralized nation beyond the centralized web | October 2022
a Ydentity validator, one must own more than 1,000,000 YDR
tokens. Ydentity validators run software allowing them to put
transactions in blocks. The process is automated and does not
require validators to be constantly monitoring their computers,
but it does require maintaining the computer (‘authority’ node)
uncompromised. Because individuals earn the right to become
validators, there is an incentive to retain the position that they
have gained. By attaching good reputation represented by YDR
reputation tokens to their Ydentity ID, validators are incentivized
to uphold the transaction process, as they do not wish to have their
Ydentity attached to a negative reputation represented by a low
level of YDR or none.
In simple words, if validators uphold transactions they earn YDR
tokens and when they do not they lose YDR tokens.
PoA consensus is considered more robust than Proof-of-Stake
because PoA only allows non-consecutive block approval from
any one validator, meaning that the risk of serious damage is
centralized to the authority node. Robustness of PoA is increased
with the number of validators and our recommendation is to have
a minimum of 23 validators.
Nevertheless, the results of the most advanced research on the
future of blockchain are leaning toward removing the consensus
mechanism entirely when autonomous entities will be “living” in
the network, they will be always right by definition.
3.5 Econodynamics
Econodynamics is the application of the mathematics of
dynamic statistical mechanics and chaos to the study of
economics. [33][34][35] If a DAN is looking for a controlled
chaos, it should make a choice of an economic system following
the logic of the thermodynamics applied to the economy and be
based on non-zero-sum game theory. This theory describes
situations where one decision maker’s gain (or loss) does not
necessarily result in the other decision maker’s loss (or gain). In
other words, it allows creation of a value based on the energy
generated by the interaction or transfer of data between parties.
The enthalpy is a property of a thermodynamic system that
could fit this purpose. An enthalpy change describes the change
in enthalpy observed in the constituents of a thermodynamic
system when undergoing a transformation. This process is
specified solely by their initial and final states.
3.5.1 Enthalpy of formation, energy of creation of bonds also
called “connections” in a nation
A common enthalpy change is the enthalpy of formation H or heat
of formation which is the change of enthalpy happening when one
substance is formed from its initial constituents which also
increases the entropy of the surroundings.
 󰇛󰇜󰇛󰇜
where  is the standard enthalpy of formation for a chemical
reaction at standard temperature and pressure (STP); v is the
coefficient of each respective reactant or product in the balanced
chemical reaction; (products) is the sum of enthalpy of
each individual product in the balanced chemical reaction;
(reactants) is the sum of the enthalpy of each individual
reactant in the balanced chemical reaction.
This property could work well to describe the change of energy
happening in a non-zero-sum game if the decentralized nation
economic system behaves hypothetically like a fully connected
system created from the initial constituents of the nation.
3.5.2 Enthalpy of atomization, energy of dissolution of bonds
in a nation
Atomization is the change in enthalpy when all bonds between
atoms of a compound are broken in a way that they become atoms
and are incapable to recreate the broken bonds. For diatomic
compounds, enthalpy of atomization is equal to enthalpy of total
dissociation. This is usually what is happening when a group or
an entire society or a nation becomes incapable to maintain the
bonds between its members.
3.5.3 Entropy of the nation
Entropy is the measure of disorderliness of the system. Entropy
generally means disorderliness which is the analog to the variance
in arrangements of particles or assets or users. Entropy is
represented by S. In case of solid blocks, the users are very close
to each other because they are arranged in regular order, so solid
blocks have less entropy than liquids. The case of liquid matters
or liquidity is the intermediate state between solids and gases in
which case users are the farthest away from each other. The more
away from each other, the larger and more positive is the entropy.
A very organized system has a low or even negative entropy that
means there a potential to expand and there are also risks in this
expansion if it is not controlled.
S(solid/real estate/gold) < S(liquid/stock/fiat) < S(gas/defi/crypto)
When entropy increases, usually enthalpy decreases. However,
they can both increase if the process is endothermic i.e. does not
create heat. The entropy of the universe is always increasing.
Systems based on maximum entropy can discount prior beliefs
and compute markets feedback accurately but there is an
associated cost for that. [36]
3.6 System optimization by learning agent
Selective machine learning or “gating” consists of introducing an
intelligent learning agent for processing incoming signals such as
a user’s performance score and timing; for computing a selection
at the gate; for computing the decisions made after the gate by
selected individuals or teams; and for monitoring the retro-signals
also known as the effect of the decisions of the users on the
ecosystem. [37]
The agent is getting its knowledge from a double-loop
mechanism between the networks and the gate on one side, and
the ecosystem on the other side. The agent is designed for a
partially observable environment, stochastic (random in nature),
semi-dynamic (the environment itself does not change or very
slowly but the agent performance does), continuous (unlimited
perceptions), multi-agent operating in the quasi-known
environment.
This type of agent has the advantage to start operating in unknown
or quasi-known environments and to become more competent
than its initial knowledge alone might allow. The most important
distinction is made between the learning (element) by short term
gate looping which is responsible for making improvements at the
gate, and the performance (element or retrograde signaling)
learning by long term looping which is responsible for selecting
external action’s effects and sending a retrograde signal to the
gate to improve selection parameters.
3.7 Social graph of the Nation, a galaxy of many nested
graphs
A graph neural network (GNN) is a class of neural network for
processing data best represented by graph data structures. Based
on the assumption that transactions between users reflect social
interactions of a social network and that the density and frequency
of transaction are a measurable activity of these users, we are
proposing a model of GNN to optimize interactions and increase
transactions volume and monitoring between Ydentity ID users
in a nested graph. This GNN will be used to predict the density
of transactions (gas fees paid for consumed transactions) and
their frequency between the nodes composing a particular nested
graph wherein the node features are made from the user-profile
which is a Non-Fungible-Token ID number and scores (Ydentity
token ID) stored at the user’s wallet address.
3.7.1 Problem formulation of nested GNN
Consider a social graph deployment comprising N user-profiles.
We denote by 󰇝 󰇞  a sequence of
density of transaction measurements (fees) over T timestamps up
to the current time t, where  is the transactions snapshot across

A decentralized nation beyond the centralized web | October 2022
all user profiles, observed over an interval 󰇟   󰇠, i.e., 
󰇝

󰇞, where is the density of transactions (fees) at the i-
th profile, and  is the temporal granularity of transaction
observations configurable by a graph administrator. The social
graph is represented as a directed, weighted and dynamic graph.
At the t-th time step, we define a graph 󰇛󰇜, where 
 is the graph signal and C is the number of features (i.e.,
density information, user’s profile information), and 
the adjacency matrix where an element 
 represents the
frequency of transactions between user-profiles i and j observed
at that time. Our objective is to predict the most likely density of
transactions in the next H time steps, given the past T
observations, i.e.,

t+1, … , 
t+H = 
 󰇛  󰇜.
3.7.2 YDentityNet
Inspired from SGDNet [38], we propose YDentityNet, a deep
neural network that solves the density of transactions forecasting
problem posed in Section 3.7.1. YDentityNet captures spatio-
temporal correlations among density information at different
nodes and dynamic adjacency matrices modelled from frequency
information. SGDNet (shown in Fig. 1) consists of a feature
extraction block followed by several spatiotemporal (ST) blocks.
Each block comprises temporal layers that handle graph signals
and adjacency matrices, and spatial layers for dynamic graph
convolution. [38] We spotted the similarity between the mobile
traffic forecasting problem that SGDNet aims to solve and our
nested GNN problem, so we adopted SGDNet as a solution in our
case. In what follows we explain these different modules in the
model in more detail.
a) Spatiotemporal feature extraction: The first block generates
feature maps for the next module by capturing spatiotemporal
correlations from graph signals    and the
dynamic adjacency matrix    . The feature
dimension of A is first reduced by a Convolution Neural
Network (CNN) layer before concatenating with V, so that
features pertaining to A do not dominate in the concatenated
matrix. Then the outcome is passed through another CNN
layer to extract feature maps for the subsequent
spatiotemporal block.
b) Gated TCN: Each ST block encompasses two gated temporal
convolution networks (TCNs) and a Dynamic Graph
Convolution Network (DGCN). We adopt gated 1-D dilated
causal convolution [38] as the Temporal convolution layer to
capture complex temporal dependencies. Dilated causal
convolution works by sliding over inputs and skipping
elements with a periodically increasing step, and it is able to
handle long-term sequences in a non-recursive manner. We
stacked several dilated causal convolution layers together.
Given an input    and filter   , the dilated causal
convolution operation step is represented as:
󰇛󰇜󰇛󰇜󰇛   󰇜

 ,
where d is the dilation factor determining the length of the
skipping step. Then we leverage a gating mechanism to control
the information flow through layers, [38] as follows:
󰇛󰇜 󰇛 󰇛󰇜󰇜󰇛󰇛󰇜󰇜,
where b and c are model weights,  is the element-wise
multiplication,  (·) is an activation function, and  (·) is the
sigmoid function which controls the information passed to the
next layer. We apply Gated TCNs on both inputs V and A, to learn
their temporal dependencies while reducing the temporal
dimension of the propagated output.
c) Dynamic Graph Convolution Network: to obtain accurate
predictions both short- and long-term, we combine spectral
graph convolution and DCRNN into a dynamic graph
convolution network. Applying the spectral graph
convolution of    and the adjacency matrix  
 along the time dimension is not effective, because
these T outcome snapshots share one weight and thus lose
temporal correlations. To circumvent this issue, we adopt
EvolveGCN, [38] where we assign a weight to each snapshot,
and these weights are temporally related by a Gated recurrent
unit (GRU). Mathematically, for every snapshot xt and its
corresponding adjacency matrix ,
  
,
 󰇛󰇜,
where 
 

 and  is the weight of t-th snapshot.
Each GCN operation has a weight, which is generated from the
weight in the last snapshot. Compared to the Long Short Term
Memory (LSTM), GRU has fewer gates and therefore is faster to
train and uses less memory. We initialize  at the beginning. For
the following recurrent steps, we use the last output as both a
hidden state and the input to the GRU. Finally, we concatenate
the output from every snapshot as the final output of EvolveGCN.
We denote the EvolveGCN operator as . We express the
DGCN operation in matrix form:
󰇛󰇜     










 ,
where  denotes concatenation, 
 

, 
󰆒,  󰆒. C′ is the dimension of the hidden states.
Finally, the formula of the l-th ST block given the input graph
signal  󰆒 and input adjacency matrix  , is
given by:
  󰇛󰇜󰇛󰇜;   󰇛󰇜.
d) Training Loss: the purpose of training is to minimize the
Mean Square Error (MSE) of every snapshot, i.e.,



   
󰆹





 ,
where H is the number of prediction steps and N is the number of
user-profiles in the deployment.
Figure 1. SDGNet [38]
Nested GNN will be initially used to optimize the transaction
volume and to monitor inside a small graph and will be later
generalized to anticipate onboarding rules of new users in the
DAN.
Message Passing Neural Networks (MPNN) are models proposed
to optimize GNNs for use on larger graphs and apply them to
domains such as social graph networks and online communities.

A decentralized nation beyond the centralized web | October 2022
4. Conclusion
As a reminder of the past, the Tower of Babel [39] is a myth
referring to misunderstanding between peoples whether it was the
wish of a God or the fate of different geographics does not matter,
only the lesson remains. We wish we could understand each other,
and the web is the network that was built for this particular
purpose. In 2000, the internet made headlines as if it “may be a
passing fad as millions give up on it”. [40] It seems we have
passed that test and the majority is using the internet, but
questions remain: what do we want to do with it next? And who
does it belong to?
The future of value transfer, the future of authorship, the future of
work and investment and the future of the meaning of life are at
stake.
A Neo-Existentialism [41] is on its way and about to enter our
lives. The time for a reboot has come.
A new form of identity is at the core, but sustainability will come
from creating the sense of connectivity required in any contract
– be it social, political, economic personal or legal – learning from
the properties of connectivity observed in the mathematics of
thermal dynamics and systems and operated by graphs.
Ydentity enables the creation of an ecosystem or DAN where
each Ydentity holder can generate their own "reputation" in the
form of YDR reputation tokens that represent the value created
by a positive sum game – of digital life in a decentralized
autonomous nation. Users may choose to liquidate their
reputation, but this decision has consequences because YDR
tokens can only be earned, and liquidation is irreversible like in
real life.
5. Acknowledgements
We are grateful to Michael Buxton and Yat Wan Lui for their
thoughtful feedback and comments.
Special acknowledgment to PhD candidate Yini Fang for her
contribution to the section 3.7 YDentityNet.
All errors and views are our own.
6. About the Authors
Dr. Frederic Jumelle, MD, IEEE Senior Member, is a
neuroscientist advocating for the revaluation of the approach to
personal identity in the digital world for the future of the internet.
Tim Pagett, MAppFin, formerly Deloitte Asia Pacific Financial
Services Industry leader, is advocating for the future of value
transfer.
Dr. Ryan Lemand, PhD, is an economist advocating for a
blockchain economy and a regulated management of crypto assets.
References
[1] Universal Declaration of Human Rights, United Nations
General Assembly resolution 217A (III), Paris, December
10, 1948
[2] International Covenant on Civil and Political Rights, United
Nations General Assembly resolution 2200A (XXI),
December 16, 1966
[3] Queensland Human Rights Commission, Section 15 of the
Human Rights Act 2019
[4] Aadhaar, Unique Identification Authority of India,
Government of India, https://uidai.gov.in
[5] Sustainable Development Solutions Network, UN
Secretary-General Ban Ki-moon, August 2012. SDG 16.9,
“by 2030 provide legal identity for all including free birth
registrations.”
[6] ISO/IEC 24760-1:2019 “IT Security and Privacy – A
framework for identity management – Part 1: Terminology
and concepts” May 2019, https://www.iso.org
[7] ID2020 is a digital ID initiative funded by Microsoft and
Accenture, https://id2020.org
[8] OpenID, a project to provide the “internet identity layer”,
https://openid.net/foundation
[9] Lightweight Identity (LID), URL-based identity
management system developed by NetMesh and followed
by OpenID, OAuth and XDI
[10] G. Greenwald, E. MacAskill, L. Poitras, “Edward Snowden:
the whistleblower behind the NSA surveillance revelations”,
The Guardian, June 11, 2013
[11] C. Cadwalladr, E. Graham-Harrison, “Revealed: 50 million
Facebook profiles harvested for Cambridge Analytica in
major data breach”, The Guardian, March 17, 2018
[12] N. Borenstein, M. Kucherawy, “ A Reputation Response Set
for Email Identifiers” and “An Architecture for Reputation
Reporting”, Internet Engineering Task Force, November
2013, https://www.rfc-editor.org/info/rfc7070
[13] Ydentity ID decentralized applications are available at
https://web3.ydentity.org and https://web3.rsk.ydentity.info
[14] Association Ydentity.org, Registre du Commerce d u
Canton de Genève, Switzerland,
https://ge.ch/hrcintapp/externalCompanyReport.action?co
mpanyOfsUid=CHE-448.613.114&lang=FR
[15] Ydentity Reputation Tokens, symbol YDR, are ERC-20
tokens deployed on contract:
https://etherscan.io/address/0x32779a598c7820ad5f13031c
0d91e66dceda31bd
[16] Decentralized Identity Foundation (DIF) is a project in favor
of a decentralized identity ecosystem and funded by an
international consortium of companies and organizations
[17] Self-Sovereign Identities (SSI) are identities controlled by
individuals such as the Decentralized Identifiers of W3C
[18] M. S. Ferdous, F. Chowdhury and M. O. Alassafi, "In Search
of Self-Sovereign Identity Leveraging Blockchain
Technology," in IEEE Access, vol. 7, pp. 103059-103079,
2019, doi: 10.1109/ACCESS.2019.2931173
[19] F. A. Jumelle, Y. W. Lui, “Ydentity (Ydentity Cognitive
Theory-based Survey EN/SC)”, China work copyright
certificate No. 2020-A-01215857
[20] F. A. Jumelle, Y. W. Lui, “Ydentity v1.0 (Ydentity
Cognitive Theory-based Survey Design Software)”, China
software copyright certificate No. 2021SR0038528
[21] F. A. Jumelle, Y. W. Lui, “Method and System for
Neuropsychological Performance Test”, International
Patent (WIPO) published on January 14, 2021, No.
WO2021003681A1
[22] F. A. Jumelle, “Method and System for Neuropsychological
Performance Test”, US Patent application No.17/625,568 of
January 7, 2022, as USA national entry of PCT patent No.
WO2021003681
[23] F. A. Jumelle, Y. W. Lui, “Method and System for
Neuropsychological Performance Test”, Hong Kong SAR
standard patent certificate No. HK 40052998 A (2021) of
January 28, 2022, as published in China on June 18, 2021
[24] F. A. Jumelle, Y. W. Lui, “Method and System for
Neuropsychological Performance Test”, China Patent
application No. 201980072840.X of June 18, 2021, as China
national entry of the PCT patent No. WO2021003681Hong
Kong Short-Patent no. HK30003521 A was published on
19th May 2020
[25] Y. Zhou and B. E. Shi, “Photorealistic Facial Expression
Synthesis by the Conditional Difference Adversarial
Autoencoder,” International Conference on Affective
Computing and Intelligent Interaction, San Antonio, TX,
USA, Oct. 2017
[26] D. Deng, Z. Chen, B. E. Shi, “Multitask Emotion
Recognition with Incomplete Labels,” IEEE International
Conference on Automatic Face and Gesture Recognition, in
Buenos Aires, Argentina, Nov. 2020. (Winner of the
Valence-Arousal and the Action Unit Challenges in the FG-
2020 Workshop "Affect Recognition in-the-wild:
Uni/Multi-
Modal Analysis
& VA-AU-Expression Challenges)
[27] Y. Zhou, J. Pi and B. E. Shi, “Pose-independent Facial
Action Unit Intensity Regression Based on Multi-task Deep
Transfer Learning,” IEEE International Conference on
Automatic Face & Gesture Recognition, Washington, DC,
USA, May 2017 (Winner of 2017 Facial Expression

A decentralized nation beyond the centralized web | October 2022
Recognition and Analysis (FERA) AU Intensity Estimation
Challenge)
[28] D. Deng, L. Wu, B. E. Shi, “Iterative Distillation for Better
Uncertainty Estimates in Multitask Emotion Recognition,”
IEEE Xplore, November 24, 2021
[29] O. Gokalp and E. Tasci, "Weighted Voting Based Ensem-
ble Classification with Hyper-parameter Optimization,"
2019 Innovations in Intelligent Systems and Applications
Conference (ASYU), 2019, pp. 1-4, doi:
10.1109/ASYU48272.2019.8946373
[30] Scharfenaker, E. (2022), “Statistical equilibrium methods
in analytical political economy,” Journal of Economic
Surveys, 36: 276-309. https://doi.org/10.1111/joes.12403
[31] Richard H. Thaler [2017 Nobel Memorial Prize in Economic
Sciences], Cass R. Sunstein, “Nudge: Improving Decisions
About Health, Wealth, and Happiness”, Penguin Books,
February 24, 2009 (first published April 8, 2008)
[32] Gavin Wood (2015), “Proof of Authority (PoA) Private
Chains”, GitHub
[33] Anand Banerjee, Victor M. Yakovenko. Universal patterns
of inequality. New Journal of Physics. 2010; 12 (7):1.
[34] Ao, P., “Boltzmann Gibbs distribution of fortune and broken
time reversible symmetry in Econodynamics”, Communica-
tions in Nonlinear Science and Numerical Simulations, vol.
12, no. 5, pp. 619–626, 2007. doi: 10.1016/j.cnsns.
2005.07.004.
[35] Victor M. Yakovenko, Jr J. Barkley Rosser. Colloquium:
Statistical mechanics of money, wealth, and income.
Reviews of Modern Physics. 2009; 81 (4):1703-1725
[36] Benjamin Patrick Evans, Mikhail Prokopenko, A Maximum
Entropy Model of Bounded Rational Decision-Making with
Prior Beliefs and Market Feedback, Entropy,
10.3390/e23060669, 23, 6, (669), (2021).
[37] S. Sadeghi, M. Amiri, and F. M. Mooseloo, "Artificial
Intelligence and Its Application in Optimization under
Uncertainty", in Data Mining - Concepts and Applications.
London, United Kingdom: IntechOpen, 2021 [Online].
Available: https://www.intechopen.com/chapters/78237 doi:
10.5772/intechopen.98628
[38] Y. Fang, S. Ergüt and P. Patras, "SDGNet: A Handover-
Aware Spatiotemporal Graph Neural Network for Mobile
Traffic Forecasting," in IEEE Communications Letters,
vol. 26, no. 3, pp. 582-586, March 2022, doi:
10.1109/LCOMM.2022.3141238.
[39] The “Tower of Babel” narrative, Genesis 11:1-9
[40] Daily Mail, “Internet “may be just a passing fad as millions
give up on it’, December 5, 2000
[41] Markus Gabriel, “Neo-Existentialism”, John Wiley & Sons,
November 26, 2018