The Large-scale 3D reconstruction, texturing and semantic mapping are nowadays widely used for automated driving vehicles, virtual reality and automatic data generation. However, most approaches are developed for RGB-D cameras with colored dense point clouds and not suitable for large-scale outdoor environments using sparse LiDAR point clouds. Since a 3D surface can be usually observed from multiple camera images with different view poses, an optimal image patch selection for the texturing and an optimal semantic class estimation for the semantic mapping are still challenging.To address these problems, we propose a novel 3D reconstruction, texturing and semantic mapping system using LiDAR and camera sensors. An Adaptive Truncated Signed Distance Function is introduced to describe surfaces implicitly, which can deal with different LiDAR point sparsities and improve model quality. The from this implicit function extracted triangle mesh map is then textured from a series of registered camera images by applying an optimal image patch selection strategy. Besides that, a Markov Random Field-based data fusion approach is proposed to estimate the optimal semantic class for each triangle mesh. Our approach is evaluated on a synthetic dataset, the KITTI dataset and a dataset recorded with our experimental vehicle. The results show that the 3D models generated using our approach are more accurate in comparison to using other state-of-the-art approaches. The texturing and semantic mapping achieve also very promising results.

With the popularity of computers and mobile devices and the development of the Internet, browsers (applications used to retrieve and display information resources on the World Wide Web) are often included by default and have become an indispensable software. Therefore, research on browser security issues is essential for protecting information assets. Among many browsers in the industry, Chrome, as a cross-platform web browser developed by Google, occupies a large market share in desktop browsers, and its security risks are further amplified as its kernel is used by many other browsers. Therefore, the research on the security issues of Chrome browser is critical for browser security.This paper focuses on the vulnerability detection of the process communication interface in Chrome browser, and designs and implements a fuzzing framework, auto-mojo-fuzz (AMF). The fuzzing process mainly designs a sample optimization technique to ensure the effectiveness of input samples and improve the efficiency of fuzzing. After implementing the AMF solution, we evaluate the generated test samples to demonstrate the effectiveness of the sample optimization technique. We also prove the possibility of discovering more vulnerabilities with AMF, and tests it with the latest version of Chrome browser, finding five unique crashes, four of which are verified as security vulnerabilities, effectively proving the automatic and efficient ability of this framework to discover vulnerabilities in the process communication interfaces in browsers.

Managing player engagement is an important problem in the video game industry, as many games generate revenue via subscription models and microtransactions. We consider a class of online video games whereby players are repeatedly matched by the game to compete against one another. Players have different skill levels which affect the outcomes of matches, and the win-loss record influence their willingness to remain engaged. The goal is to maximize the overall player engagement over time by optimizing the dynamic matchmaking strategy. We propose a general but tractable framework to solve this problem, which can be formulated as an infinite linear program. We then focus on a stylized model where there are two skill levels and players churn only when they experience a losing streak. The optimal policy always matches as many low-skilled players who are not at risk of churning to high-skilled players who are one loss away from churning. In some scenarios when there are too many low-skilled players, high-skilled players are also matched to low-skilled players that are at risk of churning.

Multi-modal dense video captioning is a task using multiple information to detect all meaningful events and generate a textual description for each event. The existing works mainly rely on single visual or dual audio-visual modals in dense video captioning, while completely ignoring the text modal (subtitle). The text modal has a similar data structure as the video captions, which provides immediate semantic information to the content description for a video. In this paper, we propose a novel framework, called Two-Stage Cross-Modal Encoding Transformer Network (TS-CMETN), to realize the multi-modal dense video captioning task by fusing multiple features, including audio, visual, and text. First, we design a two-stage feature fusion encoder that hierarchically achieves the intra- and inter-modal information interaction. Second, we propose an anchor-free temporal event proposal module, which efficiently generates event proposals at each time step without the complex anchor calculation. Extensive experiments on the ActivityNet Captions dataset show that our proposed framework achieves high performance. Moreover, our approach can adaptively handle cases of the missing text modal. Our code and data are available at https://github.com/xieyulai/TM-CMETN .

In recent years Deep Neural Networks (DNNs) have achieved remarkable results and even showed superhuman capabilities in a broad range of domains. This led people to trust in DNN classifications even in security-sensitive environments like autonomous driving. Despite their impressive achievements, DNNs are known to be vulnerable to adversarial examples. Such inputs contain small perturbations to intentionally fool the attacked model. In this paper, we present a novel end-to-end framework to detect such attacks without influencing the target model's performance. Inspired by research in neuron-coverage guided testing we show that dense layers of DNNs carry security-sensitive information. With a secondary DNN we analyze the activation patterns of the dense layers during classification run-time, which enables effective and real-time detection of adversarial examples. Our prototype implementation successfully detects adversarial examples in image, natural language, and audio processing. Thereby, we cover a variety of target DNN architectures. In addition to effectively defending against state-of-the-art attacks, our approach generalizes between different sets of adversarial examples. Our experiments indicate that we are able to detect future, yet unknown, attacks. Finally, during white-box adaptive attacks, we show our method cannot be easily bypassed.

In the online video game industry, a significant portion of the revenue is generated from microtransactions, where a small amount of real-world currency is exchanged for virtual items to be used in the game. One popular way to conduct microtransactions is via a loot box, which is a random bundle of virtual items whose contents are not revealed until after purchase. In this work, we consider how to optimally price and design loot boxes from the perspective of a revenue-maximizing video game company, and analyze customer surplus under such selling strategies. Our paper provides the first formal treatment of loot boxes, with the aim to provide customers, companies, and regulatory bodies with insights into this popular selling strategy. We consider two types of loot boxes: a traditional one where customers can receive (unwanted) duplicates, and a unique one where customers are guaranteed to never receive duplicates. We show that as the number of virtual items grows large, the unique box strategy is asymptotically optimal, while the traditional box strategy only garners 36.7% of the optimal revenue. On the other hand, unique box strategies leaves almost zero customer surplus, while traditional box strategies leaves positive surplus. Further, when designing traditional and unique loot boxes, we show it is asymptotically optimal to allocate the items uniformly, even when the item valuation distributions are highly heterogeneous. We also show that when the seller purposely misrepresents the allocation probabilities, then their revenue may increase significantly and thus strict regulation is needed. Finally, we show that even if the seller allows customers to salvage unwanted items, then the customer surplus can only increase by at most 1.4%.

Multisensory lab based in Peking University, has carried out basic studies in multisensory space and time processing, intersensory binding and haptic / tactile perception. We exploited a typical paradigm of multisensory illusion-temporal ventriloquist effect and applied it in a wide range of multisensory interactions (mainly focused on temporal processing). In this work, we summarized how the tactile stimuli were exploited to compose tactile cues and as tactile apparent motion to interface with other sensory stimuli (visual and auditory stimuli) to examine the underlying perceptual organization in a multisensory context. Moreover, we introduced two examples of wearable haptic/tactile perception in our lab, by using two customized tactile devices and discussed the potential applications in this field.

It has been recognized that cellular network interfaces are not energy efficient because of tail energy after each transmission. Although many research efforts have been made to reduce tail energy, they ignore the dynamic of link quality caused by user mobility or network congestion, which would lead to limited improvement without quality-of-experience guarantee. In this paper, we study to minimize energy consumption of the cellular network interface with a sequence of download/upload requests. Given accurate estimation of achievable link rate, we design a dynamic-programming (DP) based algorithm to obtain the optimal solution. Without the knowledge of dynamic link quality and future requests, an online algorithm is proposed to approximate the optimal solution. Finally, we conduct extensive simulations using real traces to evaluate the performance of our proposals, and the results show that 29% energy can be saved by using our algorithm under typical network settings.

The key technology of next generation network, IP Multimedia Subsystem (IMS), lacks of scalability, reliability and load balancing ability because of its traditional centralized control architecture. Using P2P (peer-to-peer) technology to transform architecture of the traditional IMS core network is a feasible solution, while the function separation and independent operation requirements of IMS network elements in telecommunication career level can not be met. In this paper, we present a multi-domain and multi-overlay framework of P2P IMS core network based on cloud infrastructure. In this framework, various network elements (NEs) are divided into different domains and overlay of the core network. Experimental results indicate that this framework can be deployed in the actual operating core network and provides core functions. Besides, the service capacity grows linearly as the number of serving nodes increases. The load can be automatically distributed on each element in a balanced way and the maximum load difference between elements is no more than 16%. Moreover, the response time of redesigned procedures is much smaller than the original procedures.

IP Multimedia Subsystem (IMS) has been accepted as the core control platform by 3GPP. It has been recognized as the vision beyond GSM for the Next Generation Network (NGN). The IMS framework delivers IP multimedia to mobile users through Session Initiation Protocol (SIP) and supports heterogeneous networks access. In this paper, we propose a virtualization-based cloud platform for the IMS core network, with a novel load-balance and disaster recovery policy. Experimental results indicate that the proposed mechanism improves system performance by dynamic allocating resources according to current load. The proposed cloud infrastructure is able to recover from a disaster in seconds by using live migration of virtual machines.

We consider a multiuser MIMO Mobile Satellite System (MSS) and model its channel as a cascade of a slow varying component, directivity vector, and a fast fading component, propagation component. We study the estimation of the slow varying part of the satellite channel at the gateway. Since the channel model is nonlinear, we propose a nonlinear parametric least squares approach. This optimization problem is shown to be equivalent to an eigenvalue complementary problem. The equivalent problem does not require an intermediate estimation of the nuisance (fast fading component) with relevant benefits in terms of computational complexity. The performance of the proposed algorithm is assessed by simulations based on realistic satellite channels.

We consider a block fading interference channels with partial channel state information and we address the issue of joint power and rate allocation in a game theoretic framework. The system is intrinsically affected by outage events. Resource allocation algorithms based on Bayesian games are proposed. The existence, uniqueness, and some stability properties of Nash equilibriums (NE) are analyzed. For some asymptotic setting, closed form expressions of NEs are also provided.

This paper presents an idea about parallel interference weighted canceller (PIWC) to mitigate the degrading effects of unreliable interference estimation that is the key shortcoming of a total interference canceller. Moreover, the computational complexity of the scheme is linear in the number of users and only a little more than that of total PIC. So, the PIWC has great cost of practicability. Based on the Gaussian approximation, the determinate weighting and fuzzy weighting have been studied. Simulation results confirm that the PIWC is better than the total PIC and the fuzzy weighting is better than determinate weighting.

This paper presents a novel multicarrier-2-dimension-code division multiple access (MC-2D-CDMA) system for the forward link. Because of the full utilization of the 2-dimension spreading characteristic, the ability to reject fading and multiple access interference (MAI) is enhanced. For this system, we examine some diversity combination methods. Among them, weighted least square combination (WLSC) has better performance but is not very suitable for the forward link; auxiliary vector combination (AVC), the combination of maximum ratio combination (MRC) and orthogonality restoring correlation (ORC) is very simple and robust, and very suitable for the forward link because it does not require any prior knowledge about other users and noise. The simulation results show their validity to anti-fading and anti-MAI.