With the success of deep neural networks, Neural Architecture Search (NAS) as a way of automatic model design has attracted wide attention. As training every child model from scratch is very time-consuming, recent works leverage weight-sharing to speed up the model evaluation procedure. These approaches greatly reduce computation by maintaining a single copy of weights on the super-net and share the weights among every child model. However, weight-sharing has no theoretical guarantee and its impact has not been well studied before. In this paper, we conduct comprehensive experiments to reveal the impact of weight-sharing: (1) The bestperforming models from different runs or even from consecutive epochs within the same run have significant variance; (2) Even with high variance, we can extract valuable information from training the super-net with shared weights; (3) The interference between child models is a main factor that induces high variance; (4) Properly reducing the degree of weight sharing could effectively reduce variance and improve performance. 1

Most of existing embedding based recommendation models use embeddings (vectors) corresponding to a single fixed point in low-dimensional space, to represent users and items. Such embeddings fail to precisely represent the users/items with uncertainty often observed in recommender systems. Addressing this problem, we propose a unified deep recommendation framework employing Gaussian embeddings, which are proven adaptive to uncertain preferences exhibited by some users, resulting in better user representations and recommendation performance. Furthermore, our framework adopts Monte-Carlo sampling and convolutional neural networks to compute the correlation between the objective user and the candidate item, based on which precise recommendations are achieved. Our extensive experiments on two benchmark datasets not only justify that our proposed Gaussian embeddings capture the uncertainty of users very well, but also demonstrate its superior performance over the state-of-the-art recommendation models.

No abstract

Light-trail (LT) has been proposed as an attractive solution for optical networks to support emerging services such as video-on-demand, pseudo-wires, datacenters, etc. Due to the fact that light-trail (LT) is a shared-medium network, the media access control (MAC) protocol is essential to be implemented in the whole network. In this paper, we propose a novel MAC protocol called DAART (Delay-aware Adaptive Round Time) to improve the throughput in LT network. DAART MAC adopts the delay-aware mechanism to calculate the actual token holding time for every transmission node. More specifically, the token holding time can be evaluated from queue delay, transmission delay and slot-synchronization delay. To estimate the throughput performance of DAART MAC protocol, we give three different patterns of traffic and compare the throughput between DAART MAC and ART MAC. Simulation results show that DAART has more stable throughput performance than ART, which in return enhances the throughput of LT networks.

No abstract

No abstract

No abstract

Forecasting water consumption plays a great important role in water resource utilization and management. Grey model (GM) with differential evolution (DE) algorithm has obtained much great success in practical forecasting applications, especially for the forecasting problems with little historical information. In this paper, an enhanced DE based GM which namedStep-DE-GM is proposed to forecast urban water consumption. Simulation results show that Step-DE-GM(1,1) can reduce the value of mean absolute percentage error (MAPE) by 0.764% and 0.733% compared with GM(1,1) and DE-GM(1,1), which means Step-DE-GM achieves higher prediction accuracy.