1. “Decentralized Finance (DeFi) Projects: A Study of Key Performance Indicators in Terms of DeFi Protocols’ Valuations (Direct Link)” by Dominik Metelski and Janusz Sobieraj
TLDR:
DeFi is a vibrant and nascent area that has experienced tremendous growth over the past few years. Nevertheless, it can be challenging to measure this growth due to the lack of well-established metrics to compare and contrast DeFi protocols.
This paper evaluates the most popular metrics currently used by analysts, such as Total Value Locked (TVL), Protocol Revenue, and Token Inflation, amongst others.
The authors look specifically at the correlations between these metrics and protocol valuation to see if either can be predictive of returns.
2. “QLAMMP: A Q-Learning Agent for Optimizing Fees on Automated Market Making Protocols” by Dev Churiwala and Bhaskar Krishnamachari
TLDR:
One of the biggest issues preventing the proliferation of decentralized exchanges is the lack of liquidity in some key markets. This contributes to asset prices within these DEXs being suboptimal relative to their centralized exchange counterparts.
This is a particularly challenging issue because the entities currently providing liquidity to those markets consistently do so at a loss. This is predominantly due to a phenomenon referred to as Impermanent Loss (IL).
Many believe that in order to counter IL, the fee model employed by DEXs needs to be revamped. This paper, however, provides an interesting alternative. Rather than changing a DEX’s fee model, the authors propose a machine learning model that liquidity providers can use to optimize their fee revenue across markets.
3. “Demystifying Bitcoin Address Behavior via Graph Neural Networks” by Zhengjie Huang, Yunyang Huang, Peng Qian, Jianhai Chen, and Qinming He
TLDR:
Blockchains are pseudonymous in nature. Observers can see the addresses involved in all transactions, and these addresses function as pseudonyms.
At times, it is crucial to understand who is the entity behind an on-chain address, especially in the event of systemic failures, such as the collapse of FTX. Frequently, this is done via so-called address classifiers.
This paper introduces a new tool called BAClassifier, which can automatically classify (or cluster) Bitcoin addresses based on their behaviors. According to the authors, the BAClassifier outperforms existing Bitcoin address classifiers, with a precision-score and F1-score of 96% and 95%, respectively.
4. “Securing the Ethereum from Smart Ponzi Schemes: Identification Using Static Features” by Zibin Zheng, Weili Chen, Zhijie Zhong, Zhiguang Chen, and Yutong Lu
TLDR:
On-chain Ponzi schemes can be incredibly detrimental to the performance of a crypto network. For example, in 2018, an infamous project called FOMO 3D gamified Ponzi schemes and at times used upwards of 80% of Ethereum block space.
As such, tools that automatically identify Ponzis are critical to helping the ecosystem become more mature so that bad actors are pruned out, and block space is not wasted on toxic use cases.
This paper introduces an interesting set of on-chain heuristics that can automatically identify the patterns of a Ponzi scheme and flag users that are engaging in this activity.
5. “SATP: A simple and scalable protocol for virtual state channel networks” by Andrew Stewart, Colin Kennedy, Mike Kerzhner, George Knee, Matthias Geihs, and Sebastian Stammler
TLDR:
Payment Channel Networks (PCNs), such as Bitcoin’s Lightning Network, have popularized the idea of using off-chain coordination mechanisms for use cases such as payments.
However, there are multiple applications beyond payments that make use of the very same construct. For example, there have been experiments on Ethereum’s Raiden Network that showcased how this technology can be used for smart contracts as well.
This paper discusses yet another evolution of PCN designs, the advent of virtual channels, which can considerably improve efficiency and reduce the cost of utilizing these layer two networks.
6. “Safety Verification of Declarative Smart Contracts” by Haoxian Chen, Lan Lu, Brendan Massey, Yuepeng Wang, and Boon Thau Loo
TLDR:
Developing secure smart contracts can be a daunting task, especially given the intricacies of popular smart contract languages such as Solidity. Beyond development, proving that a smart contract is entirely secure may at times seem impossible.
In light of this challenge, there have been multiple approaches proposed to improve smart contract security, often through new programming languages. These programming languages use techniques such as formal verification to increase the security assurances of smart contracts.
This paper discusses an emerging, logic-based programming language called DeCon which has been specifically developed with security in mind. This language features a safety verification tool called DCV which can be used to provide security proofs for DeCon contracts.
Research collected and curated by @cipherix.
This newsletter is for informational purposes only and is not intended as legal, business, investment, or tax advice.
About SCRF
The Smart Contract Research Forum’s (SCRF) bold mission is to advance web3 through actionable research and knowledge-sharing. To this end, SCRF connects researchers and builders, sponsors projects, and constructs collaborative forums. SCRF’s community is an active, international network of academics, industry architects, and blockchain advocates.