Understanding the mutual preferences between potential dating partners is core to the success of modern web-scale personalized recommendation systems that power online dating platforms. In contrast to classical user-item recommendation systems which model the unidirectional preferences of users to items, understanding the bidirectional preferences between people in a reciprocal recommendation system is more complex and challenging given the dynamic nature of interactions. In this paper, we describe a reciprocal recommendation system we built for one of the leading online dating applications in Japan. We also discuss the lessons learnt from designing, developing and deploying the reciprocal recommendation system in production.

A force display system is a kind of human-coexistent robot systems, which share the space with people while they are working, and which directly touch and display force-senses to their users. For such a robot system, it is important to estimate safety quantitatively and to ensure mechanical safety. In this paper, it is described that using MR (magneto-rheological) actuators, which is one clutch type actuators, can ensure safety. Moreover, the characteristics of MR actuators that are low inertia, high torque/inertia ratio and high responsibility contribute to improve the performance of force display systems. In this study, we developed an MR actuator with low inertia (2.6/spl times/10/sup -5/[kg/spl middot/ m/sup 2/]), high torque/inertia ratio (3.8/spl times/10/sup 5/[1/s/sup 2/]) and high responsibility. Torque/inertia ratio of this MR actuator is highest among MR actuators developed so far and much higher than those of conventional servo actuators. Then, a high-performance 2-D force display system using the MR actuator was developed. Maximum force displayed is 190 [N] and rigidity of the system is 5.9 [N/mm]. Moreover, backdrivability, dynamic range and collision sense are improved by using MR actuators in comparison with a system using ER actuators.