Browsing by Author "Li, Ang"
Results Per Page
Sort Options
Item Open Access Enable Intelligence on Billion Devices with Deep Learning(2022) Li, AngWith the proliferation of edge computing and Internet of Things (IoT), billions of edge devices (e.g., smartphone, AR/VR headset, autonomous car, etc) are deployed in our daily life and constantly generating the gigantic amount of data at the network edge. Bringing deep learning to such huge volumes of data will boost many novel applications and services in edge ecosystem and fuel the continuous booming of artificial intelligence (AI). Driven by this motivation, there is an urgent need to push the AI frontier to the network edge in order to fully exploit big data residing on edge devices.
However, empowering edge intelligence with AI, especially deep learning, is technically challenging, due to the several critical challenges including privacy, efficiency, and performance.Conventional wisdom requires edge devices to transmit the data to cloud datacenters for training and inference. But moving a huge amount of data is prohibited by cost, high transmission delay, and privacy leakage. The emerging federated learning (FL) is a promising distributed learning paradigm that enables massive devices to collaboratively learn a machine learning model (e.g., deep neural network) without explicitly sharing data, and hence the privacy concerns caused by data sharing in the centralized learning can be mitigated. But FL is facing some critical challenges that hinder its deployments to edge devices, such as communication cost and data heterogeneity.
Once we obtain a learned machine learning model, the next step is to deploy the model for serving applications and services. One straightforward approach is to deploy the model on device to perform the inference locally. Unfortunately, on-device AI often suffers from poor performance because most AI applications requires high computational power, which is technically unaffordable for resource-constrained edge devices. Edge computing pushes the cloud services from the network core to the network edge, and hence bridging devices with edge servers can alleviate the computational cost of running AI models on device alone. However, such a collaborative deployment scheme will inevitably incur transmission delay and raise privacy concern due to data movement between devices and edge servers. For example, the device can send the features extracted from raw data (e.g., images) to the cloud where a pre-trained machine learning model is deployed, but these extracted features can still be exploited by attackers to recover raw data and to infer embedded private attributes (e.g., age, gender, etc.).
In this dissertation, I start with presenting a privacy-respecting data crowdsourcing framework for deep learning to address the privacy issue in centralized training. Then, I shift the setting from the centralized one to the decentralized environment, where three novel FL frameworks are proposed to jointly improve communication and computation efficiency while handling the heterogeneous data across devices. In addition to improving the learning on large-scale edge devices, I also design an efficient edge-assisted photorealistic video style transfer system for mobile phones by leveraging the collaboration between smartphones and the edge server. Besides, in order to mitigate the privacy concern caused by the data movement in the collaborative system, an adversarial training framework is proposed to prevent the adversary from reconstructing the raw data and inferring private attributes.
Item Open Access Towards Systematic and Accurate Environment Selection for Emerging Cloud Applications(2012) Li, AngAs cloud computing is gaining popularity, many application owners are migrating their
applications into the cloud. However, because of the diversity of the cloud environments
and the complexity of the modern applications, it is very challenging to find out which
cloud environment is best fitted for one's application.
In this dissertation, we design and build systems to help application owners select the
most suitable cloud environments for their applications. The first part of this thesis focuses
on how to compare the general fitness of the cloud environments. We present CloudCmp,
a novel comparator of public cloud providers. CloudCmp measures the elastic computing,
persistent storage, and networking services offered by a cloud along metrics that directly
reflect their impact on the performance of customer applications. CloudCmp strives to
ensure fairness, representativeness, and compliance of these measurements while limiting
measurement cost. Applying CloudCmp to four cloud providers that together account
for most of the cloud customers today, we find that their offered services vary widely in
performance and costs, underscoring the need for thoughtful cloud environment selection.
From case studies on three representative cloud applications, we show that CloudCmp can
guide customers in selecting the best-performing provider for their applications.
The second part focuses on how to let customers compare cloud environments in the
context of their own applications. We describe CloudProphet, a novel system that can
accurately estimate an application's performance inside a candidate cloud environment
without the need of migration. CloudProphet generates highly portable shadow programs
to mimic the behavior of a real application, and deploys them inside the cloud to estimate
the application's performance. We use the trace-and-replay technique to automatically
generate high-fidelity shadows, and leverage the popular dispatcher-worker pattern
to accurately extract and enforce the inter-component dependencies. Our evaluation in
three popular cloud platforms shows that CloudProphet can help customers pick the bestperforming
cloud environment, and can also accurately estimate the performance of a
variety of applications.