论文标题
以生物学为灵感的人类运动序列的持续学习
Biologically-Inspired Continual Learning of Human Motion Sequences
论文作者
论文摘要
这项工作提出了一个模型，用于持续学习涉及时间序列的任务，特别是人类动作。它通过构建具有生物学启发的条件时间变异自动编码器（BI-CTVAE）来改善了最近提出的脑启发的重播模型（BI-R），该模型（BI-CTVAE）实例化了高斯的潜在混合物以进行课堂表示。我们研究了一种新型的持续学习对生成（CL2GEN）场景，该场景该模型会生成不同类别的运动序列。模型的生成准确性在一组任务上进行了测试。依次学习所有动作类别后，双分区在人类运动数据集上的最终分类精度均为78％，比使用NO-Replay高63％，仅比最先进的离线训练的GRU模型低5.4％。
This work proposes a model for continual learning on tasks involving temporal sequences, specifically, human motions. It improves on a recently proposed brain-inspired replay model (BI-R) by building a biologically-inspired conditional temporal variational autoencoder (BI-CTVAE), which instantiates a latent mixture-of-Gaussians for class representation. We investigate a novel continual-learning-to-generate (CL2Gen) scenario where the model generates motion sequences of different classes. The generative accuracy of the model is tested over a set of tasks. The final classification accuracy of BI-CTVAE on a human motion dataset after sequentially learning all action classes is 78%, which is 63% higher than using no-replay, and only 5.4% lower than a state-of-the-art offline trained GRU model.
