Ladder Networks for Emotion Recognition: Using Unsupervised Auxiliary Tasks to Improve Predictions of Emotional Attributes
Ladder Networks for Emotion Recognition: Using Unsupervised Auxiliary Tasks to Improve Predictions of Emotional Attributes
复制标题
DOI：
10.21437/interspeech.2018-1391
复制
发表时间：
2018-04
期刊：
影响因子：
--
通讯作者：
Srinivas Parthasarathy;C. Busso
中科院分区：
文献类型：
--
作者：
Srinivas Parthasarathy;C. Busso
识别情绪使用几个属性维度，如唤醒，效价和优势提供了灵活性，有效地表示复杂的情绪行为范围。学习这些情感描述符的传统方法主要集中在单独的模型上来识别这些属性中的每一个。最近的工作表明，一起学习这些属性可以使模型规则化，从而获得更好的特征表示。本研究通过添加无监督辅助任务来重建隐藏层表示，从而探索了新的正则化形式。这个辅助任务需要对自动编码器每一层的隐藏表示进行去噪。该框架依赖于梯形网络，利用编码器和解码器层之间的跳过连接来学习情感维度的强大表示。结果表明，与单独学习每个属性的基线和传统的去噪自动编码器相比，梯形网络提高了系统的性能。此外，无监督辅助任务具有很好的潜力，可用于半监督设置，其中很少有标记的句子。
Recognizing emotions using few attribute dimensions such as arousal, valence and dominance provides the flexibility to effectively represent complex range of emotional behaviors. Conventional methods to learn these emotional descriptors primarily focus on separate models to recognize each of these attributes. Recent work has shown that learning these attributes together regularizes the models, leading to better feature representations. This study explores new forms of regularization by adding unsupervised auxiliary tasks to reconstruct hidden layer representations. This auxiliary task requires the denoising of hidden representations at every layer of an auto-encoder. The framework relies on ladder networks that utilize skip connections between encoder and decoder layers to learn powerful representations of emotional dimensions. The results show that ladder networks improve the performance of the system compared to baselines that individually learn each attribute, and conventional denoising autoencoders. Furthermore, the unsupervised auxiliary tasks have promising potential to be used in a semi-supervised setting, where few labeled sentences are available.