Mechanisms of operation in quantum-dot light-emitting diodes (QLEDs) have been investigated theoretically and experimentally. Important factors governing the current efficiency of QLED were examined using a machine learning approach. High hole injection barrier to QD is the dominant efficiency limiting factor, and the machine learning result was confirmed experimentally. A mechanism of high current efficiency even in the presence of high hole injection was discussed in terms of device simulation