To improve Quality of Experience (QoE) for short video applications, most commercial companies adopt preloading and adaptive streaming technologies concurrently. Though preloading can reduce rebuffering, it may greatly waste bandwidth if the downloaded video chunks are not played. Also, each short video's downloading competes against others, which makes the existing adaptive streaming technologies fail to optimize the QoE for all videos. In this paper, we propose PDAS, a Probability-Driven Adaptive Streaming framework, to minimize the bandwidth waste while guaranteeing QoE simultaneously. We formulate PDAS into an optimization problem, where a probabilistic model is designed to describe the swiping events. Then, the maximum preload size is controlled by the proposed probability-driven max-buffer model, which reduces the bandwidth waste by proactively sleeping. At last, the optimization problem is solved by jointly deciding the preload order and preload bitrate. Extensive experimental results demonstrate that PDAS achieves almost 22.34% gains on QoE and 22.80% reductions on bandwidth usage against the existing methods. As for online evaluation, PDAS ranks first in the ACM MM 2022 Grand Challenge: Short Video Streaming.

As an indispensable part of modern critical infrastructures, cameras deployed at strategic places and prime junctions in an intelligent transportation system can help operators in observing traffic flow, identifying any emergency situation, or making decisions regarding road congestion without arriving on the scene. However, these cameras are usually equipped with heterogeneous and turbulent networks, making the real-time smooth playback of traffic monitoring videos with high quality a grand challenge. In this article, we propose a lightweight Deep Reinforcement Learning-based approach, namely, sRC-C (smart bitRate Control with a Continuous action space) , to enhance the quality of real-time traffic monitoring by adjusting the video bitrate adaptively. Distinguished from the existing bitrate adjusting approaches, sRC-C can overcome the bias incurred by deterministic discretization of candidate bitrates by adjusting the video bitrate with more fine-grained control from a continuous action space, thus significantly improving the Quality-of-Service (QoS). With carefully designed state space and neural network model, sRC-C can be implemented on cameras with scarce resources to support real-time live video streaming with low inference time. Extensive experiments show that sRC-C can reduce the frame loss counts and hold time by 24% and 15.5%, respectively, even with comparable bandwidth utilization. Meanwhile, compared to the-state-of-art approaches, sRC-C can improve the QoS by 30.4%.