Recognition in planning seeks to find agent intentions, goals or activities given a set of observations and a knowledge library (e.g. goal states, plans or domain theories). In this work we introduce the problem of Online Action Recognition. It consists in recognizing, in an open world, the planning action that best explains a partially observable state transition from a knowledge library of first-order STRIPS actions, which is initially empty. We frame this as an optimization problem, and propose two algorithms to address it: Action Unification (AU) and Online Action Recognition through Unification (OARU). The former builds on logic unification and generalizes two input actions using weighted partial MaxSAT. The latter looks for an action within the library that explains an observed transition. If there is such action, it generalizes it making use of AU, building in this way an AU hierarchy. Otherwise, OARU inserts a Trivial Grounded Action (TGA) in the library that explains just that transition. We report results on benchmarks from the International Planning Competition and PDDLGym, where OARU recognizes actions accurately with respect to expert knowledge, and shows real-time performance.
Online Knowledge Distillation via Mutual Contrastive Learning for Visual Recognition
Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）；University of Chinese Academy of Sciences（中国科学院大学）；Institute of Artificial Intelligence, Beihang University（北京航空航天大学人工智能研究院）；Zhongguancun Laboratory（中关村实验室）；Horizon Robotics（地平线机器人）
Online Multi-level Contrastive Representation Distillation for Cross-Subject fNIRS Emotion Recognition
School of Electronic and Information Engineering, South China University of Technology（华南理工大学电子与信息工程学院）；Pazhou Lab（琶洲实验室）；School of Future Technology, South China University of Technology（华南理工大学未来技术学院）
Actional-Structural Graph Convolutional Networks for Skeleton-based Action Recognition
Cooperative Medianet Innovation Center, Shanghai Jiao Tong University（上海交通大学合作媒体网络创新中心）；Mitsubishi Electric Research Laboratories（三菱电机研究实验室）；Huawei Noah’s Ark Lab（华为诺亚方舟实验室）
Semi-Supervised Online Structure Learning for Composite Event Recognition
Inst. of Informatics & Telecom., National Center for Scientific Research “Demokritos”（国家科学研究中心“德莫克利特”信息与电信研究所）；Dept. of Informatics & Telecom., National and Kapodistrian University of Athens（雅典国立卡波迪斯特里亚大学信息与电信系）；Dept. of Maritime Studies, University of Piraeus（比雷埃夫斯大学海事研究系）
Skeleton-DML: Deep Metric Learning for Skeleton-Based One-Shot Action Recognition
University of Koblenz-Landau（科布伦茨-兰道大学）
单次动作识别允许仅用单个训练样本来识别人类执行的动作。这可以通过使机器人能够对以前看不见的行为做出反应来积极影响人机交互。我们将一次性动作识别问题表述为深度度量学习问题，并提出了一种新的基于图像的骨架表示，该表示在度量学习设置中表现良好。因此，我们训练一个模型，将图像表示投影到嵌入空间中。在嵌入空间中，相似的动作具有较低的欧氏距离，而不相似的动作具有较高的距离。一次性动作识别问题变成了一组活动参考样本中的最近邻搜索。我们评估我们提出的表示对其他各种基于图像的图像表示的性能。此外，我们提出了一个消融研究，显示了不同的嵌入向量的大小，损失和扩增的影响。我们的方法在相当的训练设置下，将NTU RGB + D 120数据集上的一次性动作识别协议的最先进水平提高了3.3%。通过额外的增强，我们的结果提高了7.7%以上。
