Provided is a surgical imaging apparatus that includes a multi-link, multi-joint structure including a plurality of joints that interconnect a plurality of links to provide the multi-link, multi-joint structure with a plurality of degrees of freedom, at least one video camera being disposed on a distal end of the multi-link, multi-joint structure; at least one actuator that drives at least one of the plurality of joints; and circuitry that detects a joint force experienced at the at least one of the plurality of joints in response to an applied external force, and controls the at least one actuator based on the joint force so as to position the video camera.

Provided is a medical robot arm apparatus including a plurality of joint units configured to connect a plurality of links and implement at least 6 or more degrees of freedom in driving of a multi-link structure configured with the plurality of links, and a drive control unit configured to control driving of the joint units based on states of the joint units. A front edge unit attached to a front edge of the multi-link structure is at least one medical apparatus.

Embodiments disclosed herein include a method of training a DNN. A processor initializes an element of an A matrix. The element may include a resistive processing unit. A processor determines incremental weight updates by updating the element with activation values and error values from a weight matrix multiplied by a chopper value. A processor reads an update voltage from the element. A processor determines a chopper product by multiplying the update voltage by the chopper value. A processor directs storage of an element of a hidden matrix. The element of the hidden matrix may include a summation of continuous iterations of the chopper product. A processor updates a corresponding element of a weight matrix based on the element of the hidden matrix reaching a threshold state.