The easyMINE mine management platform promises to effectively remove the need for human supervision from the cryptocoin mining equation. “What we have in mind reaches far beyond simple automation,” explains Dr. Andrzej Buller, the easyMINE team’s machine learning expert. “Our system will be capable of ‘learning’ from prior ‘experience,’ drawing informed conclusions from accumulated data and applying them to new situations.”
The underlying technology—dubbed the Autonomous Cryptocurrency Mining Controller (ACMC) —will utilize a three-level semi-neural architecture inspired by current research in machine psychodynamics and fuzzy knowledge processing.
The first level of ACMC will be responsible for crash prevention and response. This is intended to minimize unnecessary system downtime by anticipating and averting freezes, resets, and shutdowns. Any action taken at this level can be overridden by a second-level decision.
On the second level, ACMC will actively seek to improve the overall efficiency of the mining process by modifying various operating parameters and evaluating the resultant patterns of behavior. Second-level algorithms will allow ACMC to deliberately perform risky actions—in other words, to experiment—in order to learn from its own mistakes or increase the chances of finding an unexpected solution. In contrast to classic automation paradigms, ACMC’s method of negotiating contradictory perceptions will be modeled after the human decision-making process.
The third level will be responsible for the implementation of long-term operating strategies. This will involve selecting optimal configuration settings based on general performance recommendations provided by a human operator. The recommendations may concern, for instance, the relative importance of energy saving or the degree of risk-taking to be employed by the system.
Significantly, individual ACMCs will be able to share their experience with one another in order to improve the performance of the entire mine. For instance, the analysis of the interdependencies between various GPU settings in any individual machine—and their impact on the overall effectiveness of the mining process—will allow these settings to be optimized not only in that particular unit, but also in other units sharing the same or similar hardware configuration. “We are certain that, thanks to these [self-analysis and self-optimization] algorithms, easyMINE will deliver excellent hash rates, high levels of stability, and reduced operating costs,” sums up Andrzej Belczak, the project’s co-founder and CFO.
easyMINE is scheduled to launch in 2018. To fund further development of the project, 27,000,000 easyMINE tokens (EMT), based on the Ethereum Classic platform, will be offered for sale during a 30-day Initial Coin Offering event later this summer. The tokens will be listed on cryptocurrency exchanges in the third quarter of 2017.
More information about the project is available at: https://easymine.io