Search for: All records

Deploying complex machine learning models on resource-constrained devices is challenging due to limited computational power, memory, and model retrainability. To address these limitations, a hybrid system can be established by augmenting the local model with a server-side model, where samples are selectively deferred by a rejector and then sent to the server for processing. The hybrid system enables efficient use of computational resources while minimizing the overhead associated with server usage. The recently proposed Learning to Help (L2H) model proposed training a server model given a fixed local (client) model. This differs from the Learning to Defer (L2D) framework which trains the client for a fixed (expert) server. In both L2D and L2H, the training includes learning a rejector at the client to determine when to query the server. In this work, we extend the L2H model from binary to multi-class classification problems and demonstrate its applicability in a number of different scenarios of practical interest in which access to the server may be limited by cost, availability, or policy. We derive a stage-switching surrogate loss function that is differentiable, convex, and consistent with the Bayes rule corresponding to the 0-1 loss for the L2H model. Experiments show that our proposed methods offer an efficient and practical solution for multi-class classification in resource-constrained environments. 

Future real-time applications like smart cities will use complex Machine Learning (ML) models for a variety of tasks. Timely status information is required for these applications to be reliable. Offloading computation to a mobile edge cloud (MEC) can reduce the completion time of these tasks. However, using the MEC may come at a cost such as related to use of a cloud service or privacy. In this paper, we consider a source that generates time-stamped status updates for delivery to a monitor after processing by the mobile device or MEC. We study how a scheduler must forward these updates to achieve timely updates at the monitor but also limit MEC usage. We measure timeliness at the monitor using the age of information (AoI) metric. We formulate this problem as an infinite horizon Markov decision process (MDP) with an average cost criterion. We prove that an optimal scheduling policy has an age-threshold structure that depends on how long an update has been in service. 

This work explores systems that deliver source updates requiring multiple sequential processing steps. We model and analyze the Age of Information (AoI) performance of various system designs under both parallel and series server setups. In parallel setups, each processor executes all computation steps with multiple processors working in parallel, while in series setups, each processor performs a specific step in sequence. In practice, processing faster is better in terms of age but it also consumes more power. To address this age-power trade-off, we formulate and solve an optimization problem to determine the optimal service rates for each processing step under a given power budget. Our analysis focuses on a special case where updates require two computational steps. The results show that the service rate of the second step should generally be faster than that of the first step to achieve minimum AoI and reduce power wastage. Furthermore, parallel processing is found to offer a better age-power trade-off compared to series processing. 

Time-critical applications, such as virtual reality and cyber-physical systems, require not only low end-to-end latency, but also the timely delivery of information. While high-speed Ethernet adoption has reduced interconnect fabric latency, bottlenecks persist in data storage, retrieval, and processing. This work examines status updating systems where sources generate time-stamped updates that are stored in memory, and readers fulfill client requests by accessing these stored updates. Clients then utilize the retrieved updates for further computations. The asynchronous interaction between writers and readers presents challenges, including: (i) the potential for readers to encounter stale updates due to temporal disparities between the writing and reading processes, (ii) the necessity to synchronize writers and readers to prevent race conditions, and (iii) the imperative for clients to process and deliver updates within strict temporal constraints. In the first part, we study optimal reading policies in both discrete and continuous time domains to minimize the Age of Information (AoI) of source updates at the client. One of the main contributions of this part includes showing that lazy reading is timely. In the second part, we analyze the impact of synchronization primitives on update timeliness in a packet forwarding scenario, where location updates are written to a shared routing table, and application updates read from it to ensure correct delivery. Our theoretical and experimental results show that using a lock-based primitive is suitable for timely application update delivery at higher location update rates, while a lock-free mechanism is more effective at lower rates. The final part focuses on optimizing update processing when updates require multiple sequential computational steps. We compare the age performance across a multitude of pipelined and parallel server models and characterize the age-power trade-off in these models. Additionally, our analysis reveals that synchronous sequential processing is more conducive to timely update processing than asynchronous methods, and that parallel processing outperforms pipeline services in terms of AoI. 

We consider a system where the updates from independent sources are disseminated via a publish-subscribe mechanism. The sources are the publishers and a decision process (DP), acting as a subscriber, derives decision updates from the source data. We derive the stationary expected age of information (AoI) of decision updates delivered to a monitor. We show that a lazy computation policy in which the DP may sit idle before computing its next decision update can reduce the average AoI at the monitor even though the DP exerts no control over the generation of source updates. This AoI reduction is shown to occur because lazy computation can offset the negative effect of high variance in the computation time. 

Large corporations, government entities and institutions such as hospitals and census bureaus routinely collect our personal and sensitive information for providing services. A key technological challenge is designing algorithms for these services that provide useful results, while simultaneously maintaining the privacy of the individuals whose data are being shared. Differential privacy (DP) is a cryptographically motivated and mathematically rigorous approach for addressing this challenge. Under DP, a randomized algorithm provides privacy guarantees by approximating the desired functionality, leading to a privacy–utility trade-off. Strong (pure DP) privacy guarantees are often costly in terms of utility. Motivated by the need for a more efficient mechanism with better privacy–utility trade-off, we propose Gaussian FM, an improvement to the functional mechanism (FM) that offers higher utility at the expense of a weakened (approximate) DP guarantee. We analytically show that the proposed Gaussian FM algorithm can offer orders of magnitude smaller noise compared to the existing FM algorithms. We further extend our Gaussian FM algorithm to decentralized-data settings by incorporating the CAPE protocol and propose capeFM. Our method can offer the same level of utility as its centralized counterparts for a range of parameter choices. We empirically show that our proposed algorithms outperform existing state-of-the-art approaches on synthetic and real datasets.