In recent years, skeleton based action recognition is becoming an increasingly attractive alternative to existing video-based approaches, beneficial from its robust and comprehensive 3D information. In this paper, we explore an unsupervised representation learning approach for the first time to capture the long-term global motion dynamics in skeleton sequences. We design a conditional skeleton inpainting architecture for learning a fixed-dimensional representation, guided by additional adversarial training strategies. We quantitatively evaluate the effectiveness of our learning approach on three well-established action recognition datasets. Experimental results show that our learned representation is discriminative for classifying actions and can substantially reduce the sequence inpainting errors.

Unsupervised domain adaptation methods aim to alleviate performance degradation caused by domain-shift by learning domain-invariant representations. Existing deep domain adaptation methods focus on holistic feature alignment by matching source and target holistic feature distributions, without considering local features and their multi-mode statistics. We show that the learned local feature patterns are more generic and transferable and a further local feature distribution matching enables fine-grained feature alignment. In this paper, we present a method for learning domain-invariant local feature patterns and jointly aligning holistic and local feature statistics. Comparisons to the state-of-the-art unsupervised domain adaptation methods on two popular benchmark datasets demonstrate the superiority of our approach and its effectiveness on alleviating negative transfer.

Domain adaptation is an important technique to alleviate performance degradation caused by domain shift, e.g., when training and test data come from different domains. Most existing deep adaptation methods focus on reducing domain shift by matching marginal feature distributions through deep transformations on the input features, due to the unavailability of target domain labels. We show that domain shift may still exist via label distribution shift at the classifier, thus deteriorating model performances.To alleviate this issue, we propose an approximate joint distribution matching scheme by exploiting prediction uncertainty. Specifically, we use a Bayesian neural network to quantify prediction uncertainty of a classifier. By imposing distribution matching on both features and labels (via uncertainty), label distribution mismatching in source and target data is effectively alleviated, encouraging the classifier to produce consistent predictions across domains. We also propose a few techniques to improve our method by adaptively reweighting domain adaptation loss to achieve nontrivial distribution matching and stable training. Comparisons with state of the art unsupervised domain adaptation methods on three popular benchmark datasets demonstrate the superiority of our approach, especially on the effectiveness of alleviating negative transfer.

Robotic arm grasping is a fundamental operation in robotic control task goals. Most current methods for robotic grasping focus on RGB-D policy in the table surface scenario or 3D point cloud analysis and inference in the 3D space. Comparing to these methods, we propose a novel real-time multimodal hierarchical encoder-decoder neural network that fuses RGB and depth data to realize robotic humanoid grasping in 3D space with only partial observation. The quantification of raw depth data's uncertainty and depth estimation fusing RGB is considered. We develop a general labeling method to label ground-truth on common RGB-D datasets. We evaluate the effectiveness and performance of our method on a physical robot setup and our method achieves over 90% success rate in both table surface and 3D space scenarios. The video is available in https://youtu.be/_iRyLcfbTfg.

In multi-hop ad hoc networks, sources may pump more traffic into the networks than that can be supported, resulting in high end-to-end delay and packet-loss rate. Controlling the offered load at the sources can eliminate this problem. To conduct traffic control, the throughput of the network is necessary. In addition, the authors propose a concept of bandwidth mapping to analyse how much traffic in application can be injected to make the network supporting the maximum throughput. A multi-dimensional Markov model is built to analyse the performance of IEEE 802.11 DCF in unsaturated condition, in which M/G/1/K queuing model is used to analyse the service condition of the packet queue inside the node. The effects of carrier sensing property, hidden-node problem and signal capture property are considered together to analyse the contention experienced by the nodes in the network. Furthermore, the bandwidth mapping situation of voice and video traffic is analysed in string topology of a multi-hop ad hoc network. This analysis can provide a theoretical guide for the traffic admission control.

A P-layer can be formed on a SiC wafer surface by using multiple Al ion implantations and postimplantation annealing in a low pressure CVD reactor. The Al depth profile was almost box shaped with a height of 1 10 19 cm 3 and a depth of 550 nm. Three different annealing processes were developed to protect the wafer surface. Variations in RMS roughness have been measured and compared with each other. The implanted SiC, annealed with a carbon cap, maintains a high-quality surface with an RMS roughness as low as 3.8 nm. Macrosteps and terraces were found in the SiC surface, which annealed by the other two processes (protect in Ar/protect with SiC capped wafer in Ar). The RMS roughness is 12.2 nm and 6.6 nm, respectively.

No abstract

No abstract