Recent advances have enabled oracle classifiers that can classify across many classes and input distributions with high accuracy without retraining. However, these classifiers are relatively heavyweight, so that applying them to classify video is costly. We show that day-to-day video exhibits highly skewed class distributions over the short term, and that these distributions can be classified by much simpler models. We formulate the problem of detecting the short-term skews online and exploiting models based on it as a new sequential decision making problem dubbed the Online Bandit Problem, and present a new algorithm to solve it. When applied to recognizing faces in TV shows and movies, we realize end-to-end classification speedups of 2.4-7.8x/2.6-11.2x (on GPU/CPU) relative to a state-of-the-art convolutional neural network, at competitive accuracy.

We consider applying computer vision to video on cloud-backed mobile devices using Deep Neural Networks (DNNs). The computational demands of DNNs are high enough that, without careful resource management, such applications strain device battery, wireless data, and cloud cost budgets. We pose the corresponding resource management problem, which we call Approximate Model Scheduling, as one of serving a stream of heterogeneous (i.e., solving multiple classification problems) requests under resource constraints. We present the design and implementation of an optimizing compiler and runtime scheduler to address this problem. Going beyond traditional resource allocators, we allow each request to be served approximately, by systematically trading off DNN classification accuracy for resource use, and remotely, by reasoning about on-device/cloud execution trade-offs. To inform the resource allocator, we characterize how several common DNNs, when subjected to state-of-the art optimizations, trade off accuracy for resource use such as memory, computation, and energy. The heterogeneous streaming setting is a novel one for DNN execution, and we introduce two new and powerful DNN optimizations that exploit it. Using the challenging continuous mobile vision domain as a case study, we show that our techniques yield significant reductions in resource usage and perform effectively over a broad range of operating conditions.

In this work we explore cyberbullying and other toxic behavior in team competition online games. Using a dataset of over 10 million player reports on 1.46 million toxic players along with corresponding crowdsourced decisions, we test several hypotheses drawn from theories explaining toxic behavior. Besides providing large-scale, empirical based understanding of toxic behavior, our work can be used as a basis for building systems to detect, prevent, and counter-act toxic behavior.

Current app stores distribute some malware to unsuspecting users, even though the app approval process may be costly and time-consuming. High-integrity app stores must provide stronger guarantees that their apps are not malicious. We propose a verification model for use in such app stores to guarantee that the apps are free of malicious information flows. In our model, the software vendor and the app store auditor collaborate -- each does tasks that are easy for her/him, reducing overall verification cost. The software vendor provides a behavioral specification of information flow (at a finer granularity than used by current app stores) and source code annotated with information-flow type qualifiers. A flow-sensitive, context-sensitive information-flow type system checks the information flow type qualifiers in the source code and proves that only information flows in the specification can occur at run time. The app store auditor uses the vendor-provided source code to manually verify declassifications.

Today’s smartphone operating systems frequently fail to provide users with visibility into how third-party applications collect and share their private data. We address these shortcomings with TaintDroid, an efficient, system-wide dynamic taint tracking and analysis system capable of simultaneously tracking multiple sources of sensitive data. TaintDroid enables realtime analysis by leveraging Android’s virtualized execution environment. TaintDroid incurs only 32% performance overhead on a CPU-bound microbenchmark and imposes negligible overhead on interactive third-party applications. Using TaintDroid to monitor the behavior of 30 popular third-party Android applications, in our 2010 study we found 20 applications potentially misused users’ private information; so did a similar fraction of the tested applications in our 2012 study. Monitoring the flow of privacy-sensitive data with TaintDroid provides valuable input for smartphone users and security service firms seeking to identify misbehaving applications.

This paper presents the design and implementation of a programming system that enables third-party developers to add spoken natural language (SNL) interfaces to standalone mobile applications. The central challenge is to create statistical recognition models that are accurate and resource-efficient in the face of the variety of natural language, while requiring little specialized knowledge from developers. We show that given a few examples from the developer, it is possible to elicit comprehensive sets of paraphrases of the examples using internet crowds. The exhaustive nature of these paraphrases allows us to use relatively simple, automatically derived statistical models for speech and language understanding that perform well without per-application tuning. We have realized our design fully as an extension to the Visual Studio IDE. Based on a new benchmark dataset with 3500 spoken instances of 27 commands from 20 subjects and a small developer study, we establish the promise of our approach and the impact of various design choices.

As personal information increases in value, the incentives for remote services to collect as much of it as possible increase as well. In the current Internet, the default assumption is that all behavior can be correlated using a variety of identifying information, not the least of which is a user's IP address. Tools like Tor, Privoxy, and even NATs, are located at the opposite end of the spectrum and prevent any behavior from being linked. Instead, our goal is to provide users with more control over linkability---which activites of the user can be correlated at the remote services---not necessarily more anonymity.

We consider estimation of multiple high-dimensional Gaussian graphical models corresponding to a single set of nodes under several distinct conditions. We assume that most aspects of the networks are shared, but that there are some structured differences between them. Specifically, the network differences are generated from node perturbations: a few nodes are perturbed across networks, and most or all edges stemming from such nodes differ between networks. This corresponds to a simple model for the mechanism underlying many cancers, in which the gene regulatory network is disrupted due to the aberrant activity of a few specific genes. We propose to solve this problem using the structured joint graphical lasso, a convex optimization problem that is based upon the use of a novel symmetric overlap norm penalty, which we solve using an alternating directions method of multipliers algorithm. Our proposal is illustrated on synthetic data and on an application to brain cancer gene expression data.

As the Internet has become more popular, it has increasingly been a target and medium for monitoring, censorship, content discrimination, and denial of service. Although anonymizing overlays such as Tor [2] provide some help to end users in combating these trends, the overlays themselves have become targets in turn. In this paper, we take a fresh approach: instead of running Tor on top of IP, we propose to run Tor instead of IP. We ask: what might the Internet look like if privacy and censorship resistance had been designed in from scratch? To be practical, any proposal also needs to be robust to failures, achieve reasonable efficiency compared to today's Internet, and be consistent with ISP economic concerns. Although preliminary, we argue that our design achieves these goals.

SSL/TLS is a standard protocol for secure Internet communication. Despite its great success, today's SSL deployment is largely limited to security-critical domains. The low adoption rate of SSL is mainly due to high computation overhead on the server side.

Social networking services are a fast-growing business in the Internet. However, it is unknown if online relationships and their growth patterns are the same as in real-life social networks. In this paper, we compare the structures of three online social networking services: Cyworld, MySpace, and orkut, each with more than 10 million users, respectively. We have access to complete data of Cyworld's ilchon (friend) relationships and analyze its degree distribution, clustering property, degree correlation, and evolution over time. We also use Cyworld data to evaluate the validity of snowball sampling method, which we use to crawl and obtain partial network topologies of MySpace and orkut. Cyworld, the oldest of the three, demonstrates a changing scaling behavior over time in degree distribution. The latest Cyworld data's degree distribution exhibits a multi-scaling behavior, while those of MySpace and orkut have simple scaling behaviors with different exponents. Very interestingly, each of the two e ponents corresponds to the different segments in Cyworld's degree distribution. Certain online social networking services encourage online activities that cannot be easily copied in real life; we show that they deviate from close-knit online social networks which show a similar degree correlation pattern to real-life social networks.