We develop a semantic parser that is trained in a grounded setting using pairs of videos captioned with sentences. This setting is both data-efficient, requiring little annotation, and similar to the experience of children where they observe their environment and listen to speakers. The semantic parser recovers the meaning of English sentences despite not having access to any annotated sentences. It does so despite the ambiguity inherent in vision where a sentence may refer to any combination of objects, object properties, relations or actions taken by any agent in a video. For this task, we collected a new dataset for grounded language acquisition. Learning a grounded semantic parser-turning sentences into logical forms using captioned videos-can significantly expand the range of data that parsers can be trained on, lower the effort of training a semantic parser, and ultimately lead to a better understanding of child language acquisition.

No abstract

Small-scale, out-of-plane actuators can enable tactile interfaces; however, achieving sufficient actuator force and displacement can require larger actuators. In this work, 2-mm2 out-of-plane microactuators were created, and were demonstrated to output up to 6.3 µm of displacement and 16 mN of blocking force at 170 V. The actuators converted in-plane force and displacement from a piezoelectric extensional actuator into out-of-plane force and displacement using robust, microelectromechanical systems (MEMS)-enabled, half-scissor amplifiers. The microscissors employed two layers of lithographically patterned SU-8 epoxy microstructures, laminated with a thin film of structural polyimide and adhesive to form compact flexural hinges that enabled the actuators’ small area. The self-aligned manufacture minimized assembly error and fabrication complexity. The scissor design dominated the actuators’ performance, and the effects of varying scissor angle, flexure thickness, and adhesive type were characterized to optimize the actuators’ output. Reducing the microscissor angle yielded the highest actuator performance, as it maximized the amplification of the half-scissor’s displacement and minimized scissor deformation under externally applied loads. The actuators’ simultaneously large displacements and blocking forces for their size were quantified by a high displacement-blocking force product per unit area of up to 50 mN·µm/mm2. For a linear force–displacement relationship, this corresponds to a work done per unit area of 25 mN·µm/mm2.