Titlebook: Artificial Intelligence for Edge Computing; Mudhakar Srivatsa,Tarek Abdelzaher,Ting He Book 2023 The Editor(s) (if applicable) and The Aut

helper-T-cells

Self-Supervised Learning from Unlabeled IoT Dataupervised fashion (i.e., without an explicit need to label the data). While many approaches were proposed for training foundation models, two of the most important ones are (1) contrastive learning, and (2) masking. Briefly, contrastive learning teaches the model a notion of semantic similarity by p

弓箭

Out of Distribution Detectionr fingerprints of neural network models. The key idea of NeuralFP is to exploit the different behavior in how the neural network model responds to normal data versus OOD data. Specifically, NeuralFP builds autoencoders for each layer of the neural network model and then analyzes the error distributi

陈腐思想

Irrepressible

Communication Efficient Distributed Learningcal-SGD: Distributed SGD with quantization, sparsification and local computations. In: NeurIPS, 2019), Basu et al. (IEEE J Sel Areas Inf Theory 1(1):217–226, 2020), Singh et al. (SPARQ-SGD: Event-triggered and compressed communication in decentralized optimization. In: IEEE Control and Decision Conf

小官

Chauvinistic

Arrhythmia

Book 2023cs, such as personal wearables and smart household appliances, to city infrastructure and industrial process control systems. One common trait across these applications is that the data that the application operates on come directly (typically via sensors) from the physical world. Thus, from the per

DEFT

Introduction: Internal Folklore,f training neural networks for edge AI applications. Training is a pre-requisite of all that comes next. The key training bottleneck in AI is traditionally the cost of data labeling. Unlabeled data are widely available in the IoT space, but labeling is expensive. This leads to the central question c

蜿蜒而流

Fairies and Humans between Possible Worlds,neural network with ReLU activation that has no bias term. We show that, depending on the ground-truth function, the test error of overfitted NTK models exhibits characteristics that are different from the “double-descent” of other overparameterized linear models with simple Fourier or Gaussian feat

呼吸

The Method of Explicit Scheduler,able to out of distribution detection (OOD) due to some unique characteristics of anomalies. OOD records are rare, heterogeneous, boundless, and prohibitively high costs for collecting large-scale OOD data. OOD records leads to false predictions for AI models. It reduces user confidence in AI produc