No abstract

We present Point-BERT, a new paradigm for learning Transformers to generalize the concept of BERT [8] to 3D point cloud. Inspired by BERT, we devise a Masked Point Modeling (MPM) task to pre-train point cloud Transformers. Specifically, we first divide a point cloud into several local point patches, and a point cloud Tokenizer with a discrete Variational AutoEncoder (dVAE) is designed to generate discrete point tokens containing meaningful local information. Then, we randomly mask out some patches of input point clouds and feed them into the backbone Transformers. The pre-training objective is to recover the original point tokens at the masked locations under the supervision of point tokens obtained by the Tokenizer. Extensive experiments demonstrate that the proposed BERT-style pretraining strategy significantly improves the performance of standard point cloud Transformers. Equipped with our pretraining strategy, we show that a pure Transformer architecture attains 93.8% accuracy on ModelNet40 and 83.1% accuracy on the hardest setting of ScanObjectNN, surpassing carefully designed point cloud models with much fewer hand-made designs. We also demonstrate that the representations learned by Point-BERT transfer well to new tasks and domains, where our models largely advance the state-of-the-art of few-shot point cloud classification task. The code and pre-trained models are available at https: //github.com/lulutang0608/Point-BERT.

No abstract

At present, most publish/subscribe middlewares suppose that there are equal Quality of Service (QoS) requirements for all users. However, in many real-world Internet of Things (IoT) service scenarios, different users may have different delay requirements. How to provide reliable differentiated services has become an urgent problem. The rise of Software-Defined Networking (SDN) provides endless possibilities to improve the QoS of publish/subscribe middlewares due to its greater programmability. We can encode event topics and priorities into flow entries of SDN switches directly to meet customized requirements. In this paper, we first propose an SDN-like publish/subscribe middleware architecture and describe how to use this architecture and priority queues supported by OpenFlow switches to realize differentiated services. Then we present a machine learning method using the eXtreme Gradient Boosting (XGBoost) model to solve the difficult issue of getting the queuing delay of switches accurately. Finally, we propose a reliable differentiated services guarantee mechanism according to the queuing delay and the programmability of SDN to improve QoS, namely, a two-layer queue management mechanism. Experimental evaluations show that the delay predicted by the XGBoost method is closer to the real value; our mechanism can save end-to-end delay, reduce packet loss rate, and allocate bandwidth more reasonably.

No abstract