The authors introduce a new clustering problem, mixed-modal clustering, and presents a solution leveraging a novel application of cycle consistency and adversarial learning. The method is well-explained and supported by experimental results.