In recent years, deep learning (DL) models have demonstrated remarkable achievements on non-trivial tasks such as speech recognition, image processing, and natural language understanding. One of the significant contributors to its success is the proliferation of end devices that acted as a catalyst to provide data for data-hungry DL models. However, computing DL training and inference is the main challenge. Usually, central cloud servers are used for the computation, but it opens up other significant challenges, such as high latency, increased communication costs, and privacy concerns. To mitigate these drawbacks, considerable efforts have been made to push the processing of DL models to edge servers (a mesh of computing devices near end devices). Moreover, the confluence point of DL and edge has given rise to edge intelligence (EI). International Electrotechnical Commission (IEC) defines EI as the concept where the data is acquired, stored, and processed utilizing edge computing with DL and advanced networking capabilities. Broadly, EI has six levels of categories based on where the training and inference of DL take place, e.g ., cloud server, edge server and end devices. This survey paper focuses primarily on the fifth level of EI, called all in-edge level, where DL training and inference (deployment) are performed solely by edge servers. All in-edge is suitable when the end devices have low computing resources, e.g ., Internet-of-Things, and other requirements such as latency and communication cost are important such as in mission-critical applications ( e.g ., health care). Besides, 5G/6G networks are envisioned to use all in-edge. Firstly, this paper presents all in-edge computing architectures, including centralized, decentralized, and distributed. Secondly, this paper presents enabling technologies, such as model parallelism, data parallelism, and split learning, which facilitates DL training and deployment at edge servers. Thirdly, model adaptation techniques based on model compression and conditional computation are described because the standard cloud-based DL deployment cannot be directly applied to all in-edge due to its limited computational resources. Fourthly, this paper discusses eleven key performance metrics to evaluate the performance of DL at all in-edge efficiently. Finally, several open research challenges in the area of all in-edge are presented.

Machine learning (ML) in healthcare data analytics is attracting much attention because of the unprecedented power of ML to extract knowledge that improves the decision-making process. At the same time, laws and ethics codes drafted by countries to govern healthcare data are becoming stringent. Although healthcare practitioners are struggling with an enforced governance framework, we see the emergence of distributed learning-based frameworks disrupting traditional-ML-model development. Splitfed learning (SFL) is one of the recent developments in distributed machine learning that empowers healthcare practitioners to preserve the privacy of input data and enables them to train ML models. However, SFL has some extra communication and computation overheads at the client side due to the requirement of client-side model synchronization. For a resource-constrained client side (hospitals with limited computational powers), removing such conditions is required to gain efficiency in the learning. In this regard, this paper studies SFL without client-side model synchronization. The resulting architecture is known as multi-head split learning (MHSL). At the same time, it is important to investigate information leakage, which indicates how much information is gained by the server related to the raw data directly out of the smashed data—the output of the client-side model portion—passed to it by the client. Our empirical studies examine the Resnet-18 and Conv1-D architecture model on the ECG and HAM-10000 datasets under IID data distribution. The results find that SFL provides 1.81% and 2.36% better accuracy than MHSL on the ECG and HAM-10000 datasets, respectively (for cut-layer value set to 1). Analysis of experimentation with various client-side model portions demonstrates that it has an impact on the overall performance. With an increase in layers in the client-side model portion, SFL performance improves while MHSL performance degrades. Experiment results also demonstrate that information leakage provided by mutual information score values in SFL is more than MHSL for ECG and HAM-10000 datasets by $2\times {10}^{-5}$ and $4\times {10}^{-3}$, respectively.

Distributed collaborative machine learning (DCML) is a promising method in the Internet of Things (IoT) domain for training deep learning models, as data is distributed across multiple devices. A key advantage of this approach is that it improves data privacy by removing the necessity for the centralized aggregation of raw data but also empowers IoT devices with low computational power. Among various techniques in a DCML framework, federated split learning, known as splitfed learning (SFL), is the most suitable for efficient training and testing when devices have limited computational capabilities. Nevertheless, when resource-constrained IoT devices have only positive labeled data, multiclass classification deep learning models in SFL fail to converge or provide suboptimal results. To overcome these challenges, we propose splitfed learning with positive labels (SFPL). SFPL applies a random shuffling function to the smashed data received from clients before supplying it to the server for model training. Additionally, SFPL incorporates the local batch normalization for the client-side model portion during the inference phase. Our results demonstrate that SFPL outperforms SFL: (i) by factors of 51.54 and 32.57 for ResNet-56 and ResNet-32, respectively, with the CIFAR-100 dataset, and (ii) by factors of 9.23 and 8.52 for ResNet-32 and ResNet-8, respectively, with CIFAR-10 dataset. Overall, this investigation underscores the efficacy of the proposed SFPL framework in DCML.

In recent years, deep learning (DL) models have demonstrated remarkable achievements on non-trivial tasks such as speech recognition and natural language understanding. One of the significant contributors to its success is the proliferation of end devices that acted as a catalyst to provide data for data-hungry DL models. However, computing DL training and inference is the main challenge. Usually, central cloud servers are used for the computation, but it opens up other significant challenges, such as high latency, increased communication costs, and privacy concerns. To mitigate these drawbacks, considerable efforts have been made to push the processing of DL models to edge servers. Moreover, the confluence point of DL and edge has given rise to edge intelligence (EI). This survey paper focuses primarily on the fifth level of EI, called all in-edge level, where DL training and inference (deployment) are performed solely by edge servers. All in-edge is suitable when the end devices have low computing resources, e.g., Internet-of-Things, and other requirements such as latency and communication cost are important in mission-critical applications, e.g., health care. Firstly, this paper presents all in-edge computing architectures, including centralized, decentralized, and distributed. Secondly, this paper presents enabling technologies, such as model parallelism and split learning, which facilitate DL training and deployment at edge servers. Thirdly, model adaptation techniques based on model compression and conditional computation are described because the standard cloud-based DL deployment cannot be directly applied to all in-edge due to its limited computational resources. Fourthly, this paper discusses eleven key performance metrics to evaluate the performance of DL at all in-edge efficiently. Finally, several open research challenges in the area of all in-edge are presented.