Leading AI-powered cryptocurrency trading platform, OceanEX has published a paper today detailing how Shoreline, a deep-learning framework, will protect cryptocurrency exchanges from future hacks.
The paper titled “Shoreline: Data-Driven Threshold Estimation of Online Reserves of Cryptocurrency Trading Platforms” was co-documented by Jiayu Zhou, Co-founder and CTO of OceanEx, Xitong Zhang of Michigan State University, and He Zhu of OceanEx Labs.
The paper tends to determine the optimal threshold of exchanges’ hot wallets as OceanEx believes that if hot wallets are fully optimized, it could shield exchanges from malefactors’ activities.
OceanEx stated that the paper had been officially recognized by the Association for the Advancement of Artificial Intelligence (AAAI).
Exchanges’ opportunities and associated risks
Cryptocurrency exchanges are the most significant and profitable business in the cryptocurrency space. Since its inception, exchanges have proffered solutions to a variety of problems mostly associated with traditional financial institutions.
However, these opportunities are accompanied by their challenges as exchanges mostly suffer from hackers’ activities. Last year alone, exchanges lost more than $4 billion to hackers penetrating cryptocurrency exchanges and carefully maintained trading systems.
When hackers gain access to the private keys of an exchange, a program that signs transactions when customers request to withdraw funds, can easily steal the assets stored on the trading platform hot wallet to their wallet address.
Crypto exchanges mostly store a significant amount of assets on hot wallets to cater to its daily operational cost and prevent the exposure of its cold wallet private keys.
OceanEx’s Shoreline initiative explained
Several initiatives have been proposed over the years to help protect exchanges’ hot wallets from hacks. Still, OceanEx noted that Shoreline is the ideal solution to keep trading platforms’ hot wallets safe and secured.
As per the announcement, the Shoreline initiative uses machine learning techniques to provide a data-driven framework in order to determine the optimal threshold of an exchange’s hot wallets.
The solution temporarily deposits the exchange’s assets into a low-dimensional vector space based on the sampled sequences from temporal random walks on dynamic trading networks.
Notably, the hot wallets’ threshold is then estimated by combining multiple data modalities from the exchange, including historical trading activities, deposits, and withdrawals.
When concluded, the initiative would determine how exchanges should allocate funds in the hot wallets to mitigate risks from hackers while providing operational efficiency.
Meanwhile, Coinfomania reported that OceanEx launched the first-ever Vechain (VET) perpetual contracts, promising users 5x leverage while trading.
