Representation, Reasoning, and Learning for a Relational Influence Diagram Applied to a Real-Time Geological Domain

Mining companies typically process all the material extracted from a mine site using processes which are extremely consumptive of energy and chemicals. Sorting the good material from the bad would effectively reduce required resources by leaving behind the bad material and only transporting and processing the good material. We use a relational influence diagram with an explicit utility model applied to the scenario in which an unknown number of rocks in unknown positions with unknown mineral compositions pass over 7 sensors toward 7 diverters on a high-throughput rock-sorting machine developed by MineSense Technologies Ltd. After receiving noisy sensor data, the system has 400 ms to decide whether to activate diverters which will divert the rocks into either a keep or discard bin. We learn the model offline and do online inference. Our result improves over the current state-of-the-art. This paper considers the problem of sorting rock ore based on mineralogy, separating the valuable, high-grade rocks from the low-grade rocks as they pass over an array of electromagnetic sensors. By sorting more effectively, we reduce costs and help preserve the environment, because the amount of material sent to further downstream mining processes is reduced. MineSenseTM is a Vancouver-based company developing conductivity-based sensing and sorting systems to sort ore more effectively and have developed a rock sorting platform called SortOreTM pictured in 1. We have developed an anytime algorithm (Zilberstein 1996) for sorting called Rock Predictor Sorting Algorithm (RPSA) which uses the SortOreTM rock sorting platform on which we have performed training and evaluation. In SortOreTM, rocks are dumped on a conveyor belt, moving downward on the y-axis as in the schematic in Figure 2. As rocks travel, they pass over a sensor array consisting of 7 electromagnetic coils which measure magnetic and electromagnetic flux and respond to materials which have conductive or magnetic properties. Each sensor takes readings of the magnitude of the magnetic field disrupted by the rock. The computer processes the sensor data while the rocks travel to the end of the conveyor belt where they fall Copyright c 2014, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. Figure 1: MineSenseTM SortOreTM