Recently, video diffusion models (VDMs) have garnered significant attention due to their notable advancements in generating coherent and realistic video content. However, processing multiple frame features concurrently, coupled with the considerable model size, results in high latency and extensive memory consumption, hindering their broader application. Post-training quantization (PTQ) is an effective technique to reduce memory footprint and improve computational efficiency. Unlike image diffusion, we observe that the temporal features, which are integrated into all frame features, exhibit pronounced skewness. Furthermore, we investigate significant inter-channel disparities and asymmetries in the activation of video diffusion models, resulting in low coverage of quantization levels by individual channels and increasing the challenge of quantization. To address these issues, we introduce the first PTQ strategy tailored for video diffusion models, dubbed QVD. Specifically, we propose the High Temporal Discriminability Quantization (HTDQ) method, designed for temporal features, which retains the high discriminability of quantized features, providing precise temporal guidance for all video frames. In addition, we present the Scattered Channel Range Integration (SCRI) method which aims to improve the coverage of quantization levels across individual channels. Experimental validations across various models, datasets, and bit-width settings demonstrate the effectiveness of our QVD in terms of diverse metrics. In particular, we achieve near-lossless performance degradation on W8A8, outperforming the current methods by 205.12 in FVD.

In order to improve trajectory tracking accuracy, a reinforcement learning method was employed to address the trajectory tracking task in autonomous driving. There are many conveniences and advantages in theories, methods and tools utilizing reinforcement learning to solve trajectory tracking control problems. To teach an intelligent agent effective driving skills, the structured road information in the urban environment was extracted to generate an accurate reference trajectory. Then real-world scenarios were modeled to build a simulation environment for training. To effectively train the intelligent driving agent, Imitation Learning was firstly employed to teach the agent primary driving skills. Afterwards, Reinforcement Learning was adopted to optimize the agent’s driving policy. After the intelligent driving agent was well trained in the simulator, the tracking experiments were conducted in the simulator and the real-world scenarios. The proposed method was compared with base-line methods of geometric tracking, optimization-based tracking and learning-based tracking. The experimental results demonstrated that Reinforcement-Tracking can achieve accurate trajectory tracking performance and even exceed the accuracy of most baseline methods.

Autonomous vehicles need to solve the road keeping problem and the existing solutions based on reinforcement learning are mainly implemented in the simulators. The key of transferring the well-trained models to the real world is bridging the gaps between the simulator scenarios and the real scenarios. In this paper, we propose a method called ReinforcementDriving which explores navigation skills and trajectories from simulator for full-sized road keeping. Based on the real scenario, a driving simulator is firstly established to train an intelligent driving agent. The well-trained ReinforcementDriving agent is evaluated in a real-world scenario. We compare our work with human driving, optimal control-based tracking methods and other reinforcement learning-based lane following methods. The results demonstrate that the ReinforcementDriving system can effectively achieve lane keeping in a realistic scenario with satisfactory running time and lateral accuracy.

The Internet of Things (IoT) has set off a new information technology revolution due to its convenience and efficiency. An IoT enables sharing economy, as more people are willing to share their own things (mostly mobile devices) to leverage the under-used value. In such a situation where owners and users are often not familiar with each other, an efficient access control mechanism is needed to deal with the trust issue and support service accountability to help owners accurately get their deserved profits. Besides, in such a sharing economy environment, the mobility of most shared IoT devices and their privacy preserving should also be taken into account. Regrettably, the existing schemes cannot achieve all of the aforementioned goals simultaneously and only few schemes were implemented to evaluate the claimed performance. In this article, we propose an efficient, accountable, and privacy-preserving access control solution for IoT in a sharing economy environment. In our scheme, we utilize the one-time signature to achieve anonymous authentication and let gateways store the signatures as service credentials for accountability. Meanwhile, we adopt the identity-based authentication to exclude malicious gateways and shared devices from the system and design a specialized protocol for those devices moving with the users. We conduct a detailed security analysis to show that our scheme can effectively defend against potential attacks, and also implement a prototype system to demonstrate that our design is indeed an efficient one.

Due to easy access to smartphones, recent years have witnessed an increasing interest in using the mobile phone as a sensing and computation platform for vehicle steering detection. However, relatively lower accuracy of smartphone sensors than on-board diagnostic (OBD)-based systems often leads to lower accuracy. We propose an ensemble learning-based model combined with the heuristic algorithm for smartphone-based vehicle steering detection in this paper. Ensemble learning has been widely recognized for its powerful generalization capability, high accuracy, and rapid convergence. However, applying the ensemble learning approach to steering detection of the smartphone-based vehicle entails many challenges due to the limitation of smartphone storage, the constraint on power consumption, and the requirement of being real-time. To address these challenges, we propose a series of techniques to reduce the complexity of the model and energy consumption, while at the same time maintaining high detection accuracy. The performance of the proposed system has been demonstrated using a real dataset and can achieve an accuracy of 97.37%. We also conduct two case studies on real road environment in Beijing with different smartphones.

Due to the unavailability of Vehicle-to-Infrastructure (V2I) communication in current transportation systems, Traffic Light Detection (TLD) is still considered an important module in autonomous vehicles and Driver Assistance Systems (DAS). To overcome low flexibility and accuracy of vision-based heuristic algorithms and high power consumption of deep learning-based methods, we propose a lightweight and real-time traffic light detector for the autonomous vehicle platform. Our model consists of a heuristic candidate region selection module to identify all possible traffic lights, and a lightweight Convolution Neural Network (CNN) classifier to classify the results obtained. Offline simulations on the GPU server with the collected dataset and several public datasets show that our model achieves higher average accuracy and less time consumption. By integrating our detector module on NVidia Jetson TX1/TX2, we conduct on-road tests on two full-scale self-driving vehicle platforms (a car and a bus) in normal traffic conditions. Our model can achieve an average detection accuracy of 99.3 percent (mRttld) and 99.7 percent (Rttld) at 10Hz on TX1 and TX2, respectively. The on-road tests also show that our traffic light detection module can achieve <; + 1:5m errors at stop lines when working with other selfdriving modules.

Aggressive driving is the main cause of traffic accidents all over the world. Aggressive steering is the second leading cause of traffic accidents, just behind speeding. To detect aggressive steering and encourage drivers to cultivate better driving behaviors, automatic recognition of different vehicle steering modes is required so a decision can be taken on whether the behavior is aggressive or not. This paper investigates various patterns of turning, changing lanes and U-turns, and then design and implement a system termed MultiWave that utilizes the gyroscope of a smartphone to automatically detect vehicle steering patterns. MultiWave divides driving behaviors into different categories by analyzing the data collected by gyroscope sensors. Since it analyzes the gyroscopic sensor data of turning, changing lanes and U-turns, MultiWave is flexible and robust for most urban situations. The classification accuracy of MultiWave can reach averages of 92%, 76.5%, and 87% for vehicle turning, changing lanes and U-turn, respectively.

With the wide use of computers, how to make students understand the special way of thinking of computer science and find the appropriate methods to solve problems in their own fields like computer scientists are the challenges for educators. This paper proposed a computational thinking model based on blended learning Mctbbland a descending dimension method for problem-space transformation. The Mctbblmodel analyzes and researches the application of computational thinking in computer science by the organic combination of computational thinking and blended learning in a life-cycle perspective. On one hand, the model can summarize the core concepts and important principles in computer science and extract the typical thoughts and general methods of solving problems; on the other hand, the model can implement the teaching and learning of computational thinking by means of blended learning. Based on the model, the teachers can teach the problem-solving process by using computer more intuitively and clearly and the students can understand the application of computer knowledge more easily and simply. The Mctbbl prolongs the study life cycle which takes preview as beginning, teacher-student discussion as body and practical application as core to ensure integrity, endurance and consistency of teaching.

The recommender systems community is paying great attention to diversity as key qualities beyond accuracy in real recommendation scenarios. Multifarious diversity-increasing approaches have been developed to enhance recommendation diversity in the related literature while making personalized recommendations to users. In this work, we present Gaussian Cloud Recommendation Algorithm (GCRA), a novel method designed to balance accuracy and diversity personalized top-N recommendation lists in order to capture the user's complete spectrum of tastes. Our proposed algorithm does not require semantic information. Meanwhile we propose a unified framework to extend the traditional CF algorithms via utilizing GCRA for improving the recommendation system performance. Our work builds upon prior research on recommender systems. Though being detrimental to average accuracy, we show that our method can capture the user's complete spectrum of interests. Systematic experiments on three real-world data sets have demonstrated the effectiveness of our proposed approach in learning both accuracy and diversity.

The number of Non-communicable disease (NCD) patients in China is growing rapidly, which is far beyond the capacity of the national health and social security system. Community health stations do not have enough doctors to take care of their patients in traditional ways. In order to establish a bridge between doctors and patients, we propose eBPlatform, which is an information system based on the Internet of Things (IoT) technology for homecare of the NCD patients. The eBox is a sensor which can be deployed in the patient's home for blood pressure measurement, blood sugar measurement and ECG signals collection. Some services are running on the remote server, which can receive the samples, filter and analyze the ECG signals. The uploaded data will be pushed to a web portal, with which doctors provide treatments online. The system requirements, design and implementation of hardware and software are discussed respectively. Finally, we investigate a case study with 50 NCD patients for half a year in Beijing. The results show that eBPlatform can increase the efficiency of the doctor and make a big progress to eliminate the numerical imbalance between community medical practitioners and NCD patients.

In this paper, we aim to propose an new approach to automatically generate text and photo overviews for landmarks, using the vast quantity of geo-tagged photos from photo sharing websites such as Flickr. To that end we first cluster a large-scale geo-tagged web photo collections into groups by location. And next, we make our focus on analyzing tags(user-entered text labels associated with a photo) and develop an TF-IDF like approach to evaluate the representativeness of a tag within group. To provide a more friendly and comprehensive overview, we try to mine topic patterns in each group. So we propose a clustering method among tags to discover some meaningful topics. The text overviews are generated based on tag's representativeness and topics discovered, and the photo overviews generating is simply a retrieval work according to the overview tags generated. We perform evaluations over 530,123 real geo-tagged photos taken in Beijing from Flickr and The experiment shows our approach can obtain a better result than state-of-the-art method.

The power law distribution of social tag frequency is one of the most important statistic characteristics in social tagging system. Researchers have proposed several models to predict this distribution, but these models depend on specific system. What's more, data fitting methods were failed to describe the raw data precisely compared with a mathematical model. Random Group Formation (RGF) model is a newly proposed mathematical model which could predict the power law distribution precisely. This model does not depend on the specific system. It could predict the power law distribution precisely using only three values: total number of elements, groups and the number of elements in the largest group. In this paper, RGF model is employed to predict the power law distribution of tag frequency. The experiments show that the power law distribution of tag frequency can be predicted precisely by three values, and the distribution is independent on specific system. The analysis of the tag dataset using RGF model shows that the exponent of power law decreases with the increasing of the dataset's size.