Jing Huang

TL;DR: This paper proposes to exploit the long-range dependencies among the multiple time steps for bus arrival prediction via recurrent neural network (RNN) through RNN with long short-term memory block to correct the prediction for a station by the correlated multiple passed stations.

TL;DR: In this paper, the authors proposed a method for predicting the arrival time of a bus, and the method comprises the steps: collecting bus data and bus route data in a certain period, and obtaining a time interval data pair of the bus at each data point; inputting the bus operation sequence to an LSTM recurrent neural network; and obtaining the predicted bus arrival time based on a link-type prediction method through the arrival-time prediction model.

TL;DR: This work focuses on Pick-Up (PU)/Drop-Off (DO) points from taxi trips, and proposes a fine-grained approach to unveil a set of low spatio-temporal patterns from the regularity-discovered intensity, which enables domain experts to discover patterns that were previously unattainable for them.

TL;DR: This paper proposes a simple yet powerful descriptor for object detection and recognition, called Rotative Maximal Pattern (RMP), and shows that this approach leads to a promising performance on Caltech 101, Scene 15, UIUCsport and Stanford 40 action data sets.

TL;DR: In this paper, an in depth analysis of two types of noises (Poisson and Gaussian) for cross-media topic detection is provided and it is observed that the combination of Poisson noise and topic sizes performs best while Gaussian noise has a faster optimization speed than that ofPoisson one.