Search

Recent progress in large language code models (LLCMs) has led to a dramatic surge in the use of software development. Nevertheless, it is widely known that training a well-performed LLCM requires a plethora of workforce for collecting the data and high quality annotation. Additionally, the training dataset may be proprietary (or partially open source to the public), and the training process is often conducted on a large-scale cluster of GPUs with high costs. Inspired by the recent success of imitation attacks in extracting computer vision and natural language models, this work launches the first imitation attack on LLCMs: by querying a target LLCM with carefully-designed queries and collecting the outputs, the adversary can train an imitation model that manifests close behavior with the target LLCM. We systematically investigate the effectiveness of launching imitation attacks under different query schemes and different LLCM tasks. We also design novel methods to polish the LLCM outputs, resulting in an effective imitation training process. We summarize our findings and provide lessons harvested in this study that can help better depict the attack surface of LLCMs. Our research contributes to the growing body of knowledge on imitation attacks and defenses in deep neural models, particularly in the domain of code related tasks., Comment: 11 pages

Crowdsourcing mobile user's network performance has become an effective way of understanding and improving mobile network performance and user quality-of-experience. However, the current measurement method is still based on the landline measurement paradigm in which a measurement app measures the path to fixed (measurement or web) servers. In this work, we introduce a new paradigm of measuring per-app mobile network performance. We design and implement MopEye, an Android app to measure network round-trip delay for each app whenever there is app traffic. This opportunistic measurement can be conducted automatically without users intervention. Therefore, it can facilitate a large-scale and long-term crowdsourcing of mobile network performance. In the course of implementing MopEye, we have overcome a suite of challenges to make the continuous latency monitoring lightweight and accurate. We have deployed MopEye to Google Play for an IRB-approved crowdsourcing study in a period of ten months, which obtains over five million measurements from 6,266 Android apps on 2,351 smartphones. The analysis reveals a number of new findings on the per-app network performance and mobile DNS performance., Comment: This paper has been accepted by 2017 USENIX Annual Technical Conference, ATC'17 (https://www.usenix.org/conference/atc17/technical-sessions/presentation/wu)

Cross-app collaboration via inter-component communication is a fundamental mechanism on Android. Although it brings the benefits such as functionality reuse and data sharing, a threat called component hijacking is also introduced. By hijacking a vulnerable component in victim apps, an attack app can escalate its privilege for originally prohibited operations. Many prior studies have been performed to understand and mitigate this issue, but component hijacking remains a serious open problem in the Android ecosystem due to no effective defense deployed in the wild. In this paper, we present our vision on practically defending against component hijacking in Android apps. First, we argue that to fundamentally prevent component hijacking, we need to switch from the previous mindset (i.e., performing system-level control or repackaging vulnerable apps after they are already released) to a more proactive version that aims to help security-inexperienced developers make secure components in the first place. To this end, we propose to embed into apps a secure component library (SecComp), which performs in-app mandatory access control on behalf of app components. An important factor for SecComp to be effective is that we find it is possible to devise a set of practical in-app policies to stop component hijacking. Furthermore, we allow developers design custom policies, beyond our by-default generic policies, to support more fine-grained access control. We have overcome challenges to implement a preliminary SecComp prototype, which stops component hijacking with very low performance overhead. We hope the future research that fully implements our vision can eventually help real-world apps get rid of component hijacking., Comment: This is a technical report on Android security from Singapore Management University

Mobile network performance measurement is important for understanding mobile user experience, problem diagnosis, and service comparison. A number of crowdsourcing measurement apps (e.g., MobiPerf [4, 6] and Netalyzr [5, 7]) have been embarked for the last few years. Unlike existing apps that use active measurement methods, we employ a novel passive-active approach to continuously monitor per-app network performance on unrooted smartphones without injecting additional network traffic. By leveraging the VpnService API on Android, MopEye, our measurement app, intercepts all network traffic and then relays them to their destinations using socket APIs. Therefore, not only MopEye can measure the round-trip time accurately, it can do so without injecting additional traffic. As a result, the bandwidth cost (and monetary cost of data usage) for conducting such a measurement is eliminated, and the measurement can be conducted free of user intervention. Our evaluation shows that MopEye's RTT measurement is very close to result of tcpdump and is more accurate than MobiPerf. We have used MopEye to conduct a one-week measurement revealing multiple interesting findings on different apps' performance., Comment: This is the copy of our CoNEXT'15 poster, published in December 1, 2015. MopEye is available on Google Play at https://play.google.com/store/apps/details?id=com.mopeye