The paper discusses how we can leverage cloud infrastructure for efficient hyperparameter tuning of deep neural networks on high dimensional hyperparameter spaces using Bayesian Optimization. The paper experiments Bayesian optimization in the cloud at different levels of concurrency for the warmup runs for the Bayesian optimization. Two different distributed hyperparameter tuning approaches were experimented in the cloud - Training on multiple nodes with higher warm-up concurrency Vs Distributed Training on multiple nodes with Horovod and reduced number of warm-up runs. The results indicate that greater number of warm-up runs for Bayesian optimization results in better exploration of the search space. The hyper parameter tuning and distributed training with Horovod in the cloud was performed using the HyperDrive framework of Azure Machine Learning Service for Video Activity Recognition problem. The experiment used a Long-term Recurrent Convolutional Network (LRCN) with transfer learning from Resnet50 backbone.

Meeting out similarity demands of clients, selection of threshold and computation of inter-cluster distance (ICD) are difficult while clustering. Hierarchical agglomerative clustering based approach is proposed for service discovery including two similarity models viz., Output Similarity Model (OSM) and Total Similarity Model (TSM) with additional levels for Degree of Match (DoM). The OSM which computes similarity between services using solely the outputs of services is proposed while clustering services to eliminate irrelevancy completely. The TSM which computes similarity between services using both inputs and outputs of services is proposed while discovering matched services of a given query. The work justifies the `complete linkage' as suitable method for computing ICD. It selects threshold-ICD in terms of DoM without altering the similarity demands of clients and ensures rightness of clusters. The computation time of discovery using clusters is found to be faster (7.32 against 170.59 seconds) than that of sequential method.