Xijunlee.github.io Website Review

Although there are currently several developed machine learning suites,such as Keras, Pytorch, etc, facilitating the prevalence of machine learning techniques, data scientists or machine learning engineers still need to face the difficulty of hyperparameter choice of machine learning models. It has been empirically proved that for a given model, different choices of hyperparameter also result in very different performances (accuracy, recall, etc). There is a trend of automating the hyperparameter selection for machine learning model, existing methods to solve the hyperparameter optimization of machine learning model, and discusses several open problems as well as future research direction of this subfield. I wrote the post after reading this chapter. It is my honor to take a presentaion in KDD 2018, which introduces my recent work at Noah’s Ark Lab of Huawei. What a great experience! My work during internship at Noah’s Ark Lab of Huawei, A Data-Driven Three-Layer Algorithm for Split Delivery Vehicle Routing Problem with 3D Container Loading Constraint, has been recently accepted by Proceedings of the 24th ACM SIGKDD International Conference on Split Delivery Vehicle Routing Problem with 3D Loading Constraints (3L-SDVRP) can be seen as the most important problem in large-scale manufacturing logistics. The goal is to devise a strategy consisting of three NP-hard planning components: vehicle routing, cargo splitting and container loading, which shall be jointly optimized for cost savings. The problem is an enhanced variant of the clas*ical logistics problem 3L-CVRP, and its complexity leaps beyond current studies of solvability. Our solution employs a novel data-driven three-layer search algorithm (DTSA), which we designed to improve both the efficiency and effectiveness of traditional meta-heuristic approaches, A detailed experimental evaluation on real data shows our algorithm is versatile in solving this practical complex constrained multi-objective optimization problem, and our framework may be of general interest. DTSA performs much better than the state-of-the-art