Toward Machine Interpreting: Lessons from Human Interpreting Studies