Lin Zhong

The remarkable success of machine learning, especially deep learning, has produced a variety of cloud-based services for mobile users. Such services require an end user to send data to the service provider, which presents a serious challenge to end-user privacy. To address this concern, prior works either add noise to the data or send features extracted from the raw data. They struggle to balance between the utility and privacy because added noise reduces utility and raw data can be reconstructed from extracted features. This work represents a methodical departure from prior works: we balance between a measure of privacy and another of utility by leveraging adversarial learning to find a sweeter tradeoff. We design an encoder that optimizes against the reconstruction error (a measure of privacy), adversarially by a Decoder, and the inference accuracy (a measure of utility) by a Classifier. The result is RAN, a novel deep model with a new training algorithm that automatically extracts features for classification that are both private and useful. It turns out that adversarially forcing the extracted features to only conveys the intended information required by classification leads to an implicit regularization leading to better classification accuracy than the original model which completely ignores privacy. Thus, we achieve better privacy with better utility, a surprising possibility in machine learning! We conducted extensive experiments on five popular datasets over four training schemes, and demonstrate the superiority of RAN compared with existing alternatives.

In prior work, we have shown that the underdiagnosed problem of state spill remains a barrier to realizing complex systems that are easy to maintain, evolve, and run reliably. This paper shares our early experience building Theseus from scratch, an OS with the guiding principle of eliminating state spill. Theseus takes inspiration from distributed systems to rethink state management, and leverages Rust language features for maximum safety, code reuse, and efficient isolation. We intend to demonstrate Theseus as a runtime composable OS, in which entities are easily interchangeable and can evolve independently without reconfiguring or rebooting.

In this paper, we present and study a digital-controlled method, called SoftNull, to enable full-duplex in many-antenna systems. Unlike most designs that rely on analog cancelers to suppress self-interference, SoftNull relies on digital transmit beamforming to reduce self-interference. SoftNull does not attempt to perfectly null self-interference, but instead seeks to reduce self-interference sufficiently to prevent swamping the receiver's dynamic range. Residual self-interference is then cancelled digitally by the receiver. We evaluate the performance of SoftNull using measurements from a 72-element antenna array in both indoor and outdoor environments. We find that SoftNull can significantly outperform half-duplex for small cells operating in the many-antenna regime, where the number of antennas is many more than the number of users served simultaneously.

Context information brings new opportunities for efficient and effective applications and services on mobile devices. A wide range of research has exploited context dependency, i.e. the relations between context(s) and the outcome, to achieve significant, quantified, performance gains for a variety of applications and services. These works typically have to deal with the challenges of multiple context sources leading to a sparse training dataset, and the challenges of energy hungry context sensors. Often, they address these challenges in an application specific and ad-hoc manner. We liberate mobile application designers and researchers from these burdens by providing a methodical approach to these challenges. In particular, we 1) define and measure the context-dependency of three principal types of mobile usage (visited websites, phone calls, and app usage) in an application agnostic yet practical manner, providing insight into the performance of potential application. 2) Address the challenge of data sparseness when dealing with multiple context sources in a systematic manner. 3) Present SmartContext to address the energy challenge by automatically selecting among context sources while ensuring a minimum accuracy for each estimation. Our analysis and findings are based on one year of usage and context traces collected in real-life settings from 24 iPhone users. We present findings regarding the context dependency of three types of mobile usage from 24 users, yet our methodology and the lessons we learn can be readily extended to other types of usage as well as system resources. Our findings guide the development of context aware systems, and highlight the challenges and expectations regarding the context dependency of mobile usage.

The nature of the conducting filaments in many resistive switching systems has been elusive. Through in situ transmission electron microscopy, we image the real-time formation and evolution of the filament in a silicon oxide resistive switch. The electroforming process is revealed to involve the local enrichment of silicon from the silicon oxide matrix. Semi-metallic silicon nanocrystals with structural variations from the conventional diamond cubic form of silicon are observed, which likely accounts for the conduction in the filament. The growth and shrinkage of the silicon nanocrystals in response to different electrical stimuli show energetically viable transition processes in the silicon forms, offering evidence to the switching mechanism. The study here also provides insights into the electrical breakdown process in silicon oxide layers, which are ubiquitous in a host of electronic devices.

Current and emerging mobile devices are omni directional in wireless communication. Such omni directionality not only limits device energy efficiency but also poses a significant challenge toward the capacity of wireless networks through inter-link interference. In this work, we seek to make mobile clients directional with beamsteering. We first demonstrate that beamsteering is already feasible to mobile devices such as Netbooks and eBook readers in terms of form factor, power efficiency, and device mobility. We further reveal that beamsteering mobile clients face a unique challenge to balance client efficiency and network capacity. There is an optimal operating point for a beamsteering mobile client in terms of the number of antennas and transmit power that achieve the required capacity with lowest power. Finally, we provide a distributed algorithm called BeamAdapt that allows each client to closely approach its optimal point iteratively without central coordination. We also offer a cellular system realization of BeamAdapt. Using Qualnet-based simulation, we show that BeamAdapt with four antennas can reduce client power consumption by 55% while maintaining a required network throughput for a large-scale network, compared to the same network with omni directional mobile clients.

Silicon oxide (SiOx) has been widely used in many electronic systems as a supportive and insulating medium. Here we demonstrate various electrical phenomena such as negative differential resistance, resistive switching and current hysteresis intrinsic to a thin layer of SiOx. These behaviors can largely mimic numerous electrical phenomena observed in molecules and other nanomaterials, suggesting that substantial caution should be paid when studying conduction in electronic systems with SiOx as a component. The actual switching can be the result of SiOx and not the presumed molecular or nanomaterial component. These electrical properties and the underlying mechanisms are discussed in detail.

Because of its excellent dielectric properties, silicon oxide (SiO_x) has long been used and considered as a passive, insulating component in the construction of electronic devices. In contrast, here we demonstrate resistive switches and memories that use SiO_x as the sole active material and can be implemented in entirely metal-free embodiments. Through cross-sectional transmission electron microscopy, we determine that the switching takes place through the voltage-driven formation and modification of silicon (Si) nanocrystals (NCs) embedded in the SiO_x matrix, with SiO_x itself also serving as the source of the formation of this Si pathway. The small sizes of the Si NCs (d ∼ 5 nm) suggest that scaling to ultrasmall domains could be feasible. Meanwhile, the switch also shows robust nonvolatile properties, high ON/OFF ratios (>10⁵), fast switching (sub-100-ns), and good endurance (10⁴ write-erase cycles). These properties in a SiO_x-based material composition showcase its potentials in constructing memory or logic devices that are fully CMOS compatible.

Voltage‐controlled resistive switching in various gap systems on SiO₂ substrates is reported. The nanoscale‐sized gaps are made by several means using different materials including metals, semiconductors, and amorphous carbon. The switching site is further reduced in size by using multiwalled carbon nanotubes and single‐walled carbon nanotubes. The switching in all the gap systems shares the same characteristics. This independence of switching on the material compositions of the electrodes, accompanied by observable damage to the SiO₂ substrate at the gap region, bespeaks the intrinsic switching from post‐breakdown SiO₂. It calls for caution when studying resistive switching in nanosystems on oxide substrates, since oxide breakdown extrinsic to the nanosystem can mimic resistive switching. Meanwhile, the high ON/OFF ratio (≈10⁵), fast switching time (2 µs, tested limit), and durable cycles show promising memory properties. The observed intermediate states reveal the filamentary nature of the switching.