Deeplink: Making Blockchain Smarter and Safer

Introducing Deeplink

After over a decade since Bitcoin was conceived, blockchain related technologies have blossomed into billion-dollar industries and reached millions of users. While it is too early to say that the general population hasn’t reached consensus (yes, that pun was intended) on the viability of blockchain services compared to traditional centralised services, it is only a matter of time.

The lifespan of blockchain services has mimicked that of the dot-com era, each being met with early criticism and featuring big booms and busts. The evolution of the underlying technology and its uptake within society is also following a similar path. History is a great tool to help us understand the present and predict the future.

If you want to know the future, look at the past — Albert Einstein

Just like websites in the early stages of the internet, smart contracts are very static and lack adequate computational functionality. By their nature, smart contracts can only do something when called on by an external influence, and are heavily limited by gas fees as to what they can do. For services on the blockchain to truely compete with centralised, traditional competitors, this must change, just as it did for the internet. what this all will enable is the integration of deep learning technology into blockchain services. The power and benefits of deep learning is well established. Whether it be in the banking industry for mortgage assessments, or in supply chains predicting supply and demand, deep learning is an invaluable tool in almost every industry in the world. It has been responsible for in-comprehensible amount of extra profits for businesses that integrate it into their systems. We see this as the next evolutionary step in blockchain’s lifecycle, and essential to a sustainable, decentralised ecosystem.

Introducing Deeplink

Deeplink is an Ethereum based deep-learning protocol allowing blockchains to maintain Large Neural Networks on-chain while introducing programmability, machine learning and on-chain data science into blockchains. Protocols can use on-chain neural networks and agents to improve efficiency, on-chain execution capabilities and improve security.

When the blockchains become mainstream and complex, the underlining blockchain operation, efficiency and infrastructure optimisation will get harder. The abundance of unlabelled on-chain data will enable a new generation of on-chain statistics and predictive modelling. On-chain computational agents will learn to make decisions by trial and error.

On-chain agents can be allocated for specific environments (i.e. Uniswap) to develop neural networks by obtaining, collecting and simulating different scenarios to gain foundational knowledge and understanding of the use case, data sets and their participants (e.g. traders/LPs). The larger neural networks can be trained with more and more data, leading to increased performance.

The protocol owners can access the pre-existing on-chain knowledge/neural networks to better understand their ecosystem, spot inefficiencies, discover security vulnerabilities, make accurate predictions, and improve their core on-chain environments (e.g. improving liquidity construction and slippage). Users will be able to build advanced data models, algorithms and on-chain executions modules, and better manage risks. Developers can build ML models, train data for smart contracts, and develop programmable on-chain applications for the EVM.

The off-chain computation will be provided by Network Validators & off-chain Computation Providers. All on-chain and off-chain communication will go through the Deeplink rollup (ZK) to ensure security, and trustless, scalable operation.

Deeplink wants to make blockchains smarter, while maintaining decentralisation and security. It will be the first of its kind, offering on-chain neural networks and dev-ops for countless applications. With an SDK, developers can integrate powerful deep-learning tools located on-chain into their own smart contracts, providing the perks of deep learning without relying on any centralised authority, and making their products smarter and safer.

Governance

The protocol will feature two tokens on the Ethereum blockchain: DEEP (ERC-20) and DKeeper (ERC-721).

DEEP Token

The DEEP Token will be the core ecosystem value-anchoring token. It will be designed based on the ERC-20 standard and tradable just like any other ERC-20 token. The DEEP token however comes with the additional utility of governance voting for minting DKeepers, fees distribution (when protocol is active) and staking rewards.

DKeeper

The DKeeper is based on the ERC-721 NFT standard. Initially, these will be given to early supporters which will enable them to receive early rewards through staking, future fees distributions, and reward payouts. Most importantly however, DKeeper holders will act as Cluster Keepers. These participants are responsible for the management of clusters, and the allocation of nodes, agents and computation power for the use cases. These responsibilities will involve voting on data feed implementations and will naturally have fees and rewards attached.

Unlike the DEEP token, and many other tokens, DKeepers are NOT transferable and only ONE can ever be owned. Once received, it is impossible to sell or gift. This is done intentionally by design because they are not intended to be a speculative (or otherwise) investment vehicle — They are purely … utility functioning asset. This does not mean DKeeper holders are secure in that position — other DKeeper holders can vote to deactivate another DKeeper holder. If f DKeepers are not participating according to DAO rules (minimum participation requirements) then their rewards and fee distribution will be slashed. It is important to note here that DKeeper holders must stake their DKeeper to receive their DEEP token rewards — they are not automatically staked.

Tokenomics

DEEP Token will have a fixed supply of 100,000,000. The current tokenomics model has been designed with an emphasis on decentralisation and sustainability.

DKeeper will have an initial supply of 200, with the ability for more to be minted in the future as the protocol demands. Only 10 will be available to team members, with the remaining 190 available to the public through our initial R&D fundraising.

Roadmap

Q2 2022

• Project Launch
• R&D fundraising
• MVP — Deeplink PoC — the Beer Game
• Building academic research (on-chain bots, on-chain compression, on-chain credit system, supercluster architecture, metadata broadcast to L1 using ZK rollup)
• Talent acquisition

Q3 2022

• Continued academic research
• UniSwap cluster v1
• Deeplink Alpha: On-Chain bot (SMVP) — Smart AMM bot for UniSwap & Smart matching engine for DEXs

Q4 2022 — Q1 2023

• Project X connector integration (more info coming soon…)