While recent neural encoder-decoder models have shown great promise in modeling open-domain conversations, they often generate dull and generic responses. Unlike past work that has focused on diversifying the output of the decoder at word-level to alleviate this problem, we present a novel framework based on conditional variational autoencoders that captures the discourse-level diversity in the encoder. Our model uses latent variables to learn a distribution over potential conversational intents and generates diverse responses using only greedy decoders. We have further developed a novel variant that is integrated with linguistic prior knowledge for better performance. Finally, the training procedure is improved by introducing a bag-of-word loss. Our proposed models have been validated to generate significantly more diverse responses than baseline approaches and exhibit competence in discourse-level decision-making.

Highly flexible transparent capacitive sensors have been demonstrated for the detection of deformation and pressure. The elastomeric sensors employ a pair of compliant electrodes comprising silver nanowire networks embedded in the surface layer of polyurethane matrix, and a highly compliant dielectric spacer sandwiched between the electrodes. The capacitance of the sensor sheets increases linearly with strains up to 60% during uniaxial stretching, and linearly with externally applied transverse pressure from 1 MPa down to 1 kPa. Stretchable sensor arrays consisting of 10 × 10 pixels have also been fabricated by patterning the composite electrodes into X-Y addressable passive matrix.

No abstract

No abstract