Timezone: »

Oral
Model-Level Dual Learning
Yingce Xia · Xu Tan · Fei Tian · Tao Qin · Nenghai Yu · Tie-Yan Liu

Wed Jul 11 06:20 AM -- 06:30 AM (PDT) @ Victoria
Many artificial intelligence tasks appear in dual forms like English$\leftrightarrow$French translation and speech$\leftrightarrow$text transformation. Existing dual learning schemes, which are proposed to solve a pair of such dual tasks, explore how to leverage such dualities from data level. In this work, we propose a new learning framework, model-level dual learning, which takes duality of tasks into consideration while designing the architectures for the primal/dual models, and ties the model parameters that playing similar roles in the two tasks. We study both symmetric and asymmetric model-level dual learning. Our algorithms achieve significant improvements on neural machine translation and sentiment analysis.