Serverless edge computing (SEC) that integrates serverless and edge computing paradigms has facilitated the deployment of intelligent Internet-of-things (IoT) applications. In SEC systems, energy efficiency and serverless pricing are essential to maintain operational sustainability. Nevertheless, most existing energy-saving techniques focus only on stable energy scenarios and are therefore inapplicable to energy-harvesting SEC systems powered by intermittent renewable sources. On the other hand, serverless pricing policies generally neglect the personalized perceptions of user quality-of-experience (QoE) preferences, thereby resulting in holistic user QoE degradation from a system perspective. Moreover, these approaches cannot guarantee functional correctness of serverless applications due to the appearance of computation and communication errors in practical SEC systems. To tackle these challenges, we investigate the preference-aware fault-tolerant function embedding problem for enhancing the holistic user QoE in energy-harvesting SEC systems. We first design a personalized QoE preference predictor to characterize trade-offs between service completion time and resultant service fees of individual users. Subsequently, we develop a reinforcement learning method to decide static function embedding decisions at the offline phase. Considering the intermittency of renewable sources, we further provide an energy-adaptive function replica freezing strategy at the online phase. Evaluations demonstrate that our approach boosts the holistic user QoE by 32.2% over state-of-the-art algorithms.

Cross-modal retrieval is crucial for achieving accurate and efficient information retrieval by establishing semantic correlations between heterogeneous images and text. However, traditional image-text training sets suffer from information asymmetry, which includes short lengths and limited sentence structures. This phenomenon often results in insufficient representations of essential visual information. We introduce RichDataset, which offers extensive semantic information. It includes diverse real-life image-text pairs and AI-generated content across domains such as news, entertainment, education, and posters. Compared with classic benchmarks such as Flickr30k and MS-COCO, RichDataset exhibits a novel and balanced distribution. Existing cross-modal retrieval models face challenges in extracting distinct features from the emerging data, leading to low retrieval accuracy. We propose SGG-MVAR, a comprehensive retrieval model guided by multiview scene information and semantic relationships. Leveraging a scene knowledge database, our model parses scene graphs and identifies differences in attributes and relationships. We conduct extensive experiments to evaluate our proposed dataset and model. All experimental results consistently demonstrate a significant improvement in recall for cross-modal retrieval.

In recent years, federated learning (FL) has gained increasing attention in industrial Internet-of-Things (IIoT) domains due to its privacy-preserving advantages. However, prior works commonly adopt a one-size-fits-all strategy for FL computation resource management and reward allocation, disregarding the time-varying participant states across different FL training rounds. Consequently, these methods fail to ensure the sustainability and active participation of IIoT devices in realistic FL deployments. To bridge this gap, we propose a personalized FL methodology for green IIoT systems powered by renewable energy sources. We first establish an incentive model along with its preference parameter-solving scheme to accurately characterize the incentive preferences of individual FL participants. Subsequently, a personalized participant scheduling approach is developed to accommodate dynamic resource usage patterns and diverse incentive preferences among FL participants. Our technique integrates empirical insights into conventional proximal policy optimization methods to accelerate policy learning within reinforcement learning frameworks. Experimental results on an FL prototype system show that our methodology improves the FL model accuracy by 25.92% compared with representative baseline algorithms.

As one of the pillars in cluster computing frameworks, coflow scheduling algorithms can effectively shorten the network transmission time of cluster computing jobs, thus reducing the job completion times and improving the execution performance. However, most of existing coflow scheduling algorithms failed to consider the influences of concurrent flows, which can degrade their performance under a massive number of concurrent flows. To fill the gap, we propose a unified communication agent named Courier to minimize the number of concurrent flows in cluster computing applications, which is compatible with the mainstream coflow scheduling approaches. To maintain the scheduling order given by the scheduling algorithms, Courier merges multiple flows between each pair of hosts into a unified flow, and determines its order based on that of origin flows. In addition, in order to adapt to various types of topologies, Courier introduces a control mechanism to adjust the number of flows while maintaining the scheduling order. Extensive large-scale trace-driven simulations have shown that Courier is compatible with existing scheduling algorithms, and outperforms the state-of-the-art approaches by about 30% under a variety of workloads and topologies.

The spoken language understanding (SLU) is an important part of spoken dialogue system (SDS). In the paper, we focus on how to extract a set of act-slot-value tuples from users’ utterances in the 1st Chinese Audio-Textual Spoken Language Understanding Challenge (CATSLU). This paper adopts the pretrained BERT model to encode users’ utterances and builds multiple classifiers to get the required tuples. In our framework, finding acts and values of slots are recognized as classification tasks respectively. Such multi-task training is expected to help the encoder to get better understanding of the utterance. Since the system is built on the transcriptions given by automatic speech recognition (ASR), some tricks are applied to correct the errors of the tuples. We also found that using the minimum edit distance (MED) between results and candidates to rebuild the tuples was beneficial in our experiments.