TRUST IS CRUCIAL TO DIGITAL RELATIONSHIP ECONOMICS

This document provides an overview of the HUMANS reputation system, its use of blockchain, and our execution plan.

Just like many other talent-on-demand platforms, the HUMANS Platform started off by offering its users an ordinary search tool based on simple star ratings and reviews, the functionality of which has been gradually expanded with the addition of filters, recommendations and predictive scenarios. Whilst being useful as primary filtering tools these systems have significant limitations.

First of all, ratings and scores are not sufficiently informative. In most cases platforms use cumulative average scores or ratings, which on their own are not enough to make informed decisions when choosing another platform user to engage with. Reviews, on the other hand, can be more informative. However, very often they are biased or incomplete, tending to focus only on one aspect of the service or help provided — for example, a good taxi driver may receive a bad review from one of his passengers for the music he was playing during the ride. This is an example of a qualitative characteristic i.e. an emotional assessment of certain aspects of service, as opposed to a quantitative assessment, that considers a whole variety of objective parameters (e.g. punctuality, safety, car condition, driving skills, cleanliness, etc.).

Secondly, existing systems require that users spend time navigating through ratings and analyzing each and every profile manually to find the one that suits best. This often goes against the main idea of open platforms, which are able to make the search process more convenient and significantly faster.

Finally, rating and review systems are susceptible to fraud. Users can create fake accounts to review and positively comment on their main pages, thereby boosting their rating by providing misleading and false information. This then erodes trust, which is crucial to digital relationship economics, and creates numerous problems and deficiencies for users.

Interactive matching system

To address these problems, the HUMANS Platform is developing its own reputation system that would help its users find an ideal match for their requests within seconds by delegating the search process and navigation through large amounts of data to a sophisticated AI system, which analyzes thousands of different parameters. The HUMANS Platform is essentially an information ecosystem whose users generate large amounts of data, such as preferences, choices, emotional ratings and overall comments. On top of that, every action by the Platform’s users and every transaction between them is recorded providing data about all participants.

All this information is then used to create in-depth user profiles which are subsequently assessed and categorized to help match users against each other and against their requests. The more information users provide, the easier it is for us to match their requests. The Platform essentially relies on a ‘help us help you’ approach, wherein each user can enhance his/her (and others’) search results by providing more information and by being proactive on the Platform. Data collection and processing is described in the following diagram:

Once data is gathered and stored, AI analyzes it to match Registered Users and their requests.

The ecosystem continues recording data about Registered Users when they submit requests, communicates with other Platform participants and reviews them.

Humans reputation system

The next step in the development of the Platform will be aimed at increasing its capacity, ensuring transparency and enhancing our search and matching capabilities. To do so we plan to implement a socalled Decentralised Reputation System (DRS), basing it on the existing Interactive Matching System. The HUMANS Reputation System will help us structure the data we have about all HUMANS Platform participants by segmenting them into five groups:

HUMAN (user) — any individual Registered User of the HUMANS Platform. Users can interact with the Platform, create and manage their own accounts as well as interact with and control other groups.

COMPANY — a group that has a special account on the Platform with extended functionality. A Company is managed by admins and can interact with other groups.

COMMUNITY — similar to ‘Company’ but may include both Users and Companies.

ASSET — a dependent group that can not interact with other groups by itself. It can be owned and managed by other groups on their own behalf or on behalf of other Platform participants (Companies, Communities, Service).

SERVICE — service provided by Users. Here, users can create a Company or a Community that will focus on parts of the service (e.g. Service - Translation, a Company within this Service specializes in translations from English to Spanish).

For each of the five groups we will aim to create a unique relative assessment model that will be based on a Multimodal Preferences Vector concept. A Vector is a notional aggregate of all the information contained by the Platform about each individual User and group. Each Vector is permanent and represents a User’s/group’s reputation by considering all of the information about them, including preferences, communication and interaction with other Users/groups, viewed pages, navigation chain, recommendations, impressions and clicks. The information is also compiled into personal credibility reports that each User/group can find in his/her own account.

Vectors allow for a relative comparison of Users’ characteristics and concrete information against other Users, groups and requests.

The HUMANS information ecosystem is represented by thousands of Vectors pertaining to the Platform’s Users and other groups. Each Vector in turn is a high-dimensional dataset consisting of a bulk of random variables. Therefore, to compare Vectors we must first bring them to a uniform standard and reduce their dimensions by stripping off unnecessary information and obtaining a set of principal variables. To achieve this, we plan to use convolutional neural nets and machine learning. Given the size of each dataset, only these technologies have the capacity to rightly identify the most important variables within each Vector. We understand that dimensionality reduction is of the utmost importance and will gradually become even more important as our network grows. This is why we will ensure that our Platform is flexible enough to accommodate new and more efficient technologies.

The next stage of the comparison process would involve identification of Community/Service value Vectors. These are the Vectors that represent a relative value of each individual Community/Service and are based on past experiences, preferences and reviews made about them by Platform Users. When a User makes a request for a service, his/her individual Vector is compared against the respective Community/Service value Vector. Community/Service value Vectors are essentially used to ‘filter’ through users’ Vectors to find ideal matches for their requests. Here too we plan to employ neural nets to ensure that each Community/Service value Vector is composed accurately. Due to the large amounts of information left by Users and the Platform about each Community/Service, not even the most sophisticated data analyst would be capable of identifying the most important characteristics of Communities/Services. Machine Learning and AI technologies are therefore crucial to the Platform.

This process is illustrated in the image below.

To understand how the comparison mechanism works, let us take the taxi driver example. A taxi driver’s Vector is an aggregate of concrete factors such as punctuality, cleanliness, interactions with customers, music preferences etc. On the other hand, we have a User who would like to book a taxi. If he is a regular taxi rider his personal Vector will contain his preferences such as music, car type and driving style preferences etc. However, both of these Vectors also include all kinds of other data about Users’ activities on the Platform. Thus, comparing the two Vectors against each other is not helpful, as we have to first identify the relevant information within each Vector. To do that we ‘filter’ both Vectors through respective Community/Service value Vectors: a User’s Vector would be compared against a general taxi Service value Vector, whilst a taxi driver’s Vector would be compared against taxi rider’s preference Community value Vector. This enables the comparison to result in an ideal match.

By establishing a Vector system instead of an ordinary rating/review system we will ensure that no service provider or User is ever unjustly discriminated or not included in the search results because of bad reviews or a low rating. The Vector system will match Users and entities with similar preferences.

This process is illustrated in the image below.

Role of blockchain

We call it a Decentralised Reputation System as we plan to implement the system on blockchain. Blockchain will allow us to ensure the transparency and safety of data stored on the Platform as well as solve a number of data management problems. Furthermore, to establish trust among the Platform Users all data must be verified. An open data ledger is therefore crucial to the functionality of the HUMANS Platform as all new transactions are added and verified according to protocols that ensure consensus.

Blockchain in the HUMANS Decentralized Reputation System is responsible for storing data about every transaction, including Users’ comments, preferences and their past experiences with other Platform participants. Such information about each provided service is public and it details what service was performed, the duration of the service and its quality. Every detail of every transaction can be traced back to the moment a contract was signed. The details continue to be tracked and publicly recorded through to the completion of the service, as specified by a smart contract.

The arbitration system and fraud prevention will be performed by our powerful AI-based search engine based on the behavioral history of each User. Smart contracts will contain the following data about Users: identity, history about all communication and interactions, previous transactions, behavioral traits and matching statistics. The node consensus will operate between the Humans IDS and client GUI of decentralized applications (DApps).

The HUMANS public blockchain will store all client data and content for predicting future search requests. The content that has been stored on the blockchain will be processed by AI using the HUMANS Multimodal Preferences Vector concept and used to provide fast and accurate suggestions, enhanced search results and reputation scoring.

The target architecture of the HUMANS ecosystem is described in the following diagram:

Execution plan

The initial HUMANS Reputation System architecture is based on numerous backend modules which store User’s action data such as Platform logins, posts publication, search requests and messages. DataLake modules are used for analytics and data-driven solutions of operational services.

The system also consists of Humans IDS (Intrusion detection system) and Client GUI (Graphic User Interface). Logs are connected in an event-based orchestration and choreography layer.

The existing Reputation System works on standard and centralized server operation types, but we are preparing to run tests of the blockchain infrastructure based on smart contracts of a public blockchain. However, we are planning to develop a new blockchain that will be able to process an ever-growing number of transactions within the Platform.

For the pilot stage we have adopted the ERC-20 technical standard, which is usually used for cryptocurrencies and other fungible and divisible entities. In the future we plan to use a non-fungible token standard such as ERC-721 or its future modifications to collect and store HUMANS Reputation, given that our Reputation concept assumes that information pertaining to groups' reputation cannot be changed, transferred or split in any way.

Apart from the fact that ERC-721 is non-fungible it also has another significant advantage — on top of storing such information as names, balances, token supply and symbols it can also record large amounts of metadata about assets and ownership information which is crucial to our concept of Trust and the overall security of our Decentralized Reputation System.

The migration of the HUMANS Platform onto the blockchain is a challenging process and will happen gradually:

  1. I Implantation of the Ethereum blockchain on a testing clone of the HUMANS ecosystem creating the first version of the data storage with a limited amount of data.
  2. II Closed alpha-test of reputation modelling for early adopters and qualified Users which will be rewarded with HUMANS GEN Tokens
  3. III Integration of blockchain solutions into the running HUMANS Platform
  4. IV Beta-test of the HUMANS Reputation System on the blockchain
  5. V Analysis and requirements of blockchain
  6. VI Development of own HUMANS blockchain protocol

Ethereum transactions are relatively costly and slow. We believe that at some point during the development stage, the Ethereum blockchain will run out of capacity to service the Platform needs. We therefore aim to develop our own HUMANS blockchain which must match certain technological requirements of the HUMANS Reputation System and the Platform:

Number of DAU (Daily Active Users) and MAU (Month Active Users)

As of October 2018, the DAU volume average overlays 3,000 users. We predict that by 2022 DAU will exceed 10,000,000 with MAU volume of more than 320,000,000.

Number of daily user requests

25% of MAU will use smart contracts for their transactions, where each deal would require 6 addresses per smart contract. By 2022, we expect that the average number of daily user requests will be up to 16,000,000 with at least 1,330,000 deals daily.

Transaction form of hash information

The node should be able to store the required amount of information.

Number of user requests per minute

We expect that by 2022 user requests per minute would average 11,000, thus fast response time will become essential to ensure a positive user experience.

Multi-signature wallet

Availability for atomic transactions and storage of desired currencies.

Applicable currencies

During the testing period USD, BTC and ETH will be available.

Affordable transaction costs, without user fees

Fees will be covered by the HUMANS Platform itself.

The main milestone of developing our own blockchain will be building a comparison tool for data rating. Asset characteristics are not usually similar pieces of information.

Therefore, blockchain technology should be used to compare nodes with unique written information, based on an infinite continuum.

Disclaimer

ICO.HUMANS.NET WEBSITE TOGETHER WITH ALL CONTENT, FUNCTIONALITY AND SERVICES THEREON IS OPERATED AND MADE AVAILABLE BY HUMANS LICENSE PTE. LTD. (”WE”, “OUR” “COMPANY”). WE ARE A LIMITED COMPANY REGISTERED IN SINGAPORE ON JULY 12, 2016 UNDER COMPANY NUMBER 201618971E AND HAVE OUR REGISTERED OFFICE AT 171 TRAS STREET, #10-179 UNION BUILDING, SINGAPORE 079025. WE ARE THE OWNER OR THE LICENSEE OF ALL INTELLECTUAL PROPERTY RIGHTS AND CONTENT IN OUR SITE, AND IN THE MATERIAL PUBLISHED ON IT.

WE ARE AN AFFILIATE OF (1) HUMANS NET, INC., A DELAWARE CORPORATION INCORPORATED ON JULY 29, 2016, OPERATING AT 37 WEST 19TH ST, 2ND FLOOR NEW YORK, NY 10011 (REFERRED TO IN THIS WHITE PAPER AS “HUMANS PLATFORM OPERATOR”) AND (2) HUMANS NET EUROPE LTD, REGISTERED ON 7 MARCH 2018 UNDER REGISTRATION NUMBER 11240485 AND REGISTERED ADDRESS AT 64 KNIGHTSBRIDGE, LONDON, UNITED KINGDOM, SW1X 7JF (REFERRED TO IN THIS WHITE PAPER AS “HUMANS UK”). HUMANS NET, INC. OPERATES THE SITE HUMANS.NET TOGETHER WITH ALL CONTENT, FUNCTIONALITY AND SERVICES OFFERED THEREON (REFERRED TO IN THIS WHITE PAPER AS “HUMANS PLATFORM” OR THE “PLATFORM”). HUMANS UK SERVES AS A PAYMENT AGENT OF THE COMPANY.

THE COMPANY, HUMANS PLATFORM OPERATOR AND HUMANS UK ARE PART OF THE HUMANS GROUP OF COMPANIES (COLLECTIVELY, WITH OTHER GROUP COMPANIES REFERRED TO IN THIS WHITE PAPER, “HUMANS”) WITH RELATED, BUT SEPARATE FUNCTIONS. HUMANS PLATFORM OPERATOR IS THE OPERATOR AND PROMOTER OF HUMANS PLATFORM. HUMANS UK RECEIVES PROCEEDS OF SAFT OFFERING AND CROWDSALE ON OUR BAHALF.

WE OWN AND DEVELOP THE PLATFORM AND A SYSTEM OF RELATED DECENTRALIZED AI AND BLOCKCHAIN DRIVEN APPLICATION AND SYSTEM SOFTWARE WITH A VIEW TO EVENTUALLY EXPAND THE PLATFORM INTO HUMANS OPEN MULTI-FEATURED ECOSYSTEM (REFERRED TO IN THIS WHITE PAPER AS “HOME” OR “THE ECOSYSTEM”). WE ARE ALSO PARTNERING WITH HUMANS PLATFORM OPERATOR TO DEVELOP AND INTEGRATE THE “HUMANS GEN” TOKENS AND INCENTIVE PROGRAM (AS DESCRIBED HEREIN) INTO THE PLATFORM. WE INTRODUCE HUMANS GEN AND PROGRAM IN ORDER TO BRING ENHANCED FUNCTIONALITY AND ATTRACT NEW ACTIVE, REGISTERED USERS TO THE PLATFORM (“REGISTERED USERS”).

THE COMPANY RETAINS ALL RIGHTS TO THE PLATFORM, HOME, HUMANS GEN INFRASTRUCTURE AND ANY CONTENT RELATED THERETO.

THIS WHITE PAPER HAS BEEN PREPARED BY THE COMPANY SOLELY FOR REGISTERED USERS OF THE PLATFORM AND SOLELY FOR THE PURPOSE OF DESCRIBING THE PROPOSED TECHNICAL IMPLEMENTATION OF, AND ARCHITECTURE FOR, THE NEXT STAGE OF HUMANS PLATFORM'S DEVELOPMENT. THIS WHITE PAPER IS NON-BINDING IN ALL RESPECTS, MAY BE SUBJECT TO CHANGES WITHOUT ANY NOTICE AND DOES NOT CONSTITUTE AN OFFER OR SALE OF HUMANS GEN TOKENS DESCRIBED HEREIN, NOR DOES THIS WHITE PAPER CONSTITUTE AN INVITATION OR SOLICITATION FOR INVESTMENT PURPOSES.

THE ULTIMATE IMPLEMENTATION OF THE HUMANS GEN INFRASTRUCTURE, INCENTIVE PROGRAM AND THE ECOSYSTEM IS DEPENDENT UPON FACTORS AND RISKS OUTSIDE OF THE CONTROL OF HUMANS, INCLUDING REGULATORY RISKS, CONTRIBUTOR PARTICIPATION, AND THE CONTINUED WIDESPREAD ADOPTION AND EVOLUTION OF BLOCKCHAIN TECHNOLOGY. THERE IS A SUBSTANTIAL RISK THE SOFTWARE AND TECHNOLOGIES DESCRIBED HEREIN MAY NEVER COME TO FRUITION OR ACHIEVE THE OBJECTIVES SPECIFIED IN THIS WHITE PAPER. PROSPECTIVE REGISTERED USERS OF HUMANS PLATFORM ARE ADVISED TO CONTRIBUTE INTO AND PARTICIPATE IN IT AT THEIR OWN RISK AND WITHOUT RELYING ON ANY STATEMENT CONTAINED IN THIS WHITE PAPER.