The information contained in this report is not legal or financial advice and is for informational purposes only. The projects included in Satoshi•Fund ICO Reports vary significantly in terms of investment risks. Before considering any investment, make sure you read the report below carefully and take time to do your own due diligence.
The Neuron network will be a decentralized artificial intelligence platform on the Ethereum blockchain for training neural network models by the network participants. The first application will be personalized biology. Users will be able to bring their data onto the platform, have them analyzed by data scientists on the network and sell them to interested participants. This innovative project has the potential to boost medical research and contribute to people’s health benefit. The Neuron network token NRN is a utility token used for payments on the platform transferred via Ethereum smart contracts. The incentive system is clear in broad terms, but some key issues of the token economics are not specified yet. Only parts of the source code will be made public at a later date. The current technical specification is scarce and only a few demos are available. The management team are top professionals, although not in blockchain, and are supported by a few blockchain experts and practitioners. ICO investors are offered the right to buy the NRN tokens at a later date - when the network launches which will be no sooner than in 6 months. ICO information disclosure is high except for the unclear distribution of the incentives fund which accounts for 33% of tokens. Pre-ICO and ICO investors will not get much (37% of tokens) and the team will have 30%. Use of proceeds is not specified. For potential investors, it looks like a choice between a promising concept and the lack of any solid proof of it. The delayed token distribution is likely to attract the project’s fans and discourage those seeking quick gain.
The Neuron network is being developed by the DOC.AI company. It’s a conversational platform for on-demand, quantified biology that uses an advanced natural language dialog system. The Neuron network will be a decentralized artificial intelligence (AI) platform for training neural network models. The first application area the company will focus on will be personalized biology. Using the platform’s token NRN, users will be incentivized to bring the medical data to the platform where they can store it securely and incentivize data scientists called Neuron miners to analyze the data making predictions and suggestions on various issues. Data collected on the platform can also be used in research on important medical issues.
Implementation
Problem being solved
The project seeks to provide an easy access to AI for everyone so that any person can put their (personal) data to the network via an app and basically train their own AI. The first application area chosen by the team is personalized biology. The platform will allow to analyze users’ data and make predictions and suggestions on everything from medication renewals to users’ sleep schedule. For example, users will be able to get a quantified, documented profile (about their health) that would help them to choose the right therapy. Another area of focus of the projects is data privacy: data will be stored on the end user’s device (a tablet, smartphone or computer).
Platform users
Users interested in training their own AI models using the Neuron network: e.g. individuals interested in their own general health benefits or individuals who wish to join a common clinical data trial uniting with others in an ad hoc research organization (RO); ROs will be able to involve individuals in their data trials by rewarding them with NRN
Users and their families (“participants”, or data owners) with rare diseases who wish to connect their devices and data to the network earning NRN for that (the data stays encrypted and anonymous); they will also be able to incentivize Neuron miners (described in the next paragraph) with payments in NRN to come up with discoveries related to their specific health problem
Neuron miners (data scientists) will be able to solve medical mysteries in a contest with other miners earning NRN tokens for their work, which is similar to competitions held on Kaggle, a platform for predictive modelling where data miners compete to produce the best models for predicting and describing the data
Companies looking for data that can benefit from using AI and data driven healthcare, such as insurance or pharmaceutical companies, will be able to get users’ medical data paying them in NRN
How it will work and development roadmap
Phase 1: AI-protocol for deconstructing medical records (6 months after ICO end date)
Pilot (scalability testing) on top of the Ethereum blockchain; tokens should be released shortly after that
Participants can access the platform, connect their data to the platform for no fee and get their data aggregated and structured; in case they want a prediction on the spot by Neuron miners, they will be able to purchase one with NRN
Data grants covering all costs will be awarded to participants from an inclusion list which will be compiled later
Technology level: a doc.ai application on top of the Neuron network that uses a hybrid of natural language and faceted search allowing users to explore a collection of information by applying multiple filters
Phase 2: Miners community is formed (6 months after NEURON becomes operational)
Individual participants will be able to broadcast their data to the network to recruit other volunteer participants forming an ad hoc RO. The RO can then initiate a competition between Neuron miners to solve a scientific problem related to their data. Users will have to pay in NRN for broadcasting their proposal to miners.
The winning miners get the prize, or bounty, in NRN, and a data report will be published to the group or individual participant which they can later use. After that miners will lose access to the sandbox (the environment) in which they worked with users data for data protection
Phase 3: BIO Wallet is online (12 months after the Network becomes operational)
Data owners can get registered on the network to make their data available for sale in NRN to potential buyers
Phase 4 (about 24 months after NEURON is operational)
When the AI model is validated clinically, the RO that decide to open source it for everyone’s benefit or to buy a license for NRN
See Chapter 5 of this document for a detailed workflow for three different use-cases
Implemented using the Ethereum blockchain and Ethereum smart contracts which will be used for payments on the platform; other usage is not specified
Source code will be partially open later when the network is launched, it is not yet known which parts exactly will be open
Several non-blockchain components that teach users how to construct datasets of their health and where and how to access these datasets, are in development, and at this stage only simple demos are available on the website
Economics and token
Platform token: NEURON (NRN), ERC20 compatible
Total supply: 860,000,000 NRN tokens
Token utility:
Means of payment on the platform granting access to the platform’s functionality; the exception is the connection of users’ data to the network which will be free of charge
Incentives for developers, end users, and research organizations to participate
Fees and rewards
NRN will be a currency used in payments on the platform; the activities associated with payments in NRN are described in Implementation section above
The platform will operate as an open market, e.g. the proposal bounty for a training will be set by an individual or organization making the proposal and the payment will be transferred through a smart contract
A mechanism to reduce the risk of DDoS or Sybil attacks will be developed which will require users to pay a deposit (“buy-in”) for using the network which will be returned to them after the end of the training session; this should increase a monetary barrier to use the network to discourage malicious users
The details of the token economics are said to be further developed in the coming months
Team and advisors
The team is 18 people, the names of the management and key staff are disclosed, the rest is not
CEO Walter De Brouwer founded Scanadu Inc., a venture-backed mobile health company, Gen.Life an insurtech company, specialized in machine learning and natural language, and DOC.AI
CTO Apurv Mishra co-founded Datawallet, a consumer-to-business marketplace where users get paid for data they share with companies, worked on context-aware technologies at Glavio Labs and was a vice president at Hypios, the company which solves specific R&D problems of corporations and develops software for in-depth semantic analysis of each problem
Pre-pre-ICO (Advisory Sale): held from 15 August, 2017 to 6 September, 2017 with a 25% discount, documentation
Pre-ICO: closed on 24 September, 2017 with a 10% discount (1 NRN = $0.09 USD), a minimal individual contribution was $50,000 (see documentation for more detail)
ICO will be available for registered participants only
Pre-registration was available several days prior to ICO and will re-open at ICO start date; investors need to use their Civic App ID and pass KYC (Know Your Customer) and AML (Anti-Money Laundering) checks
ICO starts on 28 September, 2017, at 13:00 UTC and will continue until 12 October, 2017 or until the hard cap is reached, whichever earlier
Hard cap (pre-ICO and ICO): around $30,000,000 (287,866,667 NRN)
ICO token price: 1 NRN = $0.1 USD, no bonuses
Accepted currencies: BTC and ETH
Token distribution
37% - token sale investors (both pre-ICO periods and ICO)
33% - incentives for the ecosystem, but the distribution and the ownership are not specified further except they will be used to reward the community “for
providing data, maintaining blockchain, and more”
30% - retained by the DOC.AI company with vesting
Vesting for the company’s tokens:
⅓ of tokens will be locked-up for 1 year starting from ICO end date
Distribution date: tokens are expected to be issued around spring 2018 when the Neuron network launches
Legal overview
Level of information accessibility: HIGH
DOC.AI information disclosure may in large part be considered a good example of compliance and clarity for the blockchain industry
DOC.AI uses a Simple Agreement for Future Tokens (the SAFT) (users buy not the token itself, but a right to get the tokens at a later date); the Confidential Private Placement Memorandum (the CPPM) can be accessed here, it incorporates the SAFT by reference
DOC.AI discloses detailed information about the Advisory Sale, pre-sale and crowdsale; the information about the company’s stake and the incentive fund is provided in the Term Sheet of the CPPM, however, the exact distribution of the incentives fund and its ownership are not specified
The information about DOC.AI Incorporated may be found at Delaware State companies register here; the information about the company indicated in the CPPM is valid and correct
The CPPM contains the list of risk factors that may influence the sale and the operation of the project; it provides specific notices for investors from key jurisdictions
DOC.AI do not provide sufficient information about the use of proceeds; they emphasize that the funds will be spent at their own discretion
Legal assessment of the token: MEDIUM
The platform’s tokens (NRN tokens) are the core of the incentive system of the project and their function will be a means of payment
NRN tokens are not registered as securities in any jurisdiction
The CPPM distinguishes the right for purchase of NRN tokens and NRN tokens themselves; the company warns buyers about the risk of the emission registration in various jurisdictions
It is highly likely that NRN tokens will pass the Howey Test. They do not guarantee any fixed income in future. They may be used by insurance companies and research centers as a means of payment for medical data. At the same time, DOC.AI is uncertain about “achieving a balance between supply and demand” (page 2 of the CPPM) and therefore about the future value of the token. NRN Tokens may fail to pass Risk Capital Test, which can be applied to them in several states of the US.
Conclusion
Highly innovative concept bringing together AI and blockchain technology
The network may serve the individual health benefit of users as well as contribute to the advancement of medicine
Network’s token is a means of payment on the platform, which will function as an open market, and a way to incentivize users; its value will grow while the network grows and users add more and more data to their profile
Incentives structure is clear in general, but the token economics lacks specification on several issues (distribution of the incentives fund, the buy-in mechanism to mitigate the risks of Sybil attacks etc.)
Source code is not available and only parts of it will be open to the public (no sooner than in 6 months)
Technical specification is not well detailed in the whitepaper
Several parts of the project are in development, only four basic demos are available for view right now
Co-founders and key staff have a solid professional record in relevant areas, although not in blockchain; advisors include a few practitioners from the blockchain community
ICO will be held using the instrument SAFT whereby investors buy the right to get the token later but not the token itself; tokens will be distributed when the network should be launched (no sooner than in 6 months)
Level of ICO information disclosure by the company is high
Token distribution model is not very decentralized and clear: investors get 37% of tokens (both pre-ICO and ICO) against 30% going to the team; the community should get 33% of tokens distributed as incentives, but neither the ownership nor the exact allocation of this pool are clearly specified
Neuron token is unlikely to fall under a security
Data scientists on the platform are said to sign a legally binding data usage agreement with users providing data (in order not to use tokens for other purposes), but the implementation is not specified
