Filters








309,907 Hits in 4.7 sec

Predicting Periodicity with Temporal Difference Learning [article]

Kristopher De Asis, Brendan Bennett, Richard S. Sutton
2018 arXiv   pre-print
Temporal difference (TD) learning is an important approach in reinforcement learning, as it combines ideas from dynamic programming and Monte Carlo methods in a way that allows for online and incremental  ...  A key idea of TD learning is that it is learning predictive knowledge about the environment in the form of value functions, from which it can derive its behavior to address long-term sequential decision  ...  Acknowledgments The authors thank Roshan Shariff for insights and discussions contributing to the results presented in this paper, and the entire Reinforcement Learning and Artificial Intelligence research  ... 
arXiv:1809.07435v1 fatcat:uvmltmayynfd3puqbnwoiqz43e

Attentive Crowd Flow Machines [article]

Lingbo Liu, Ruimao Zhang, Jiefeng Peng, Guanbin Li, Bowen Du, and Liang Lin
2018 arXiv   pre-print
representations of temporally-varying data with an attention mechanism.  ...  Based on the ACFM, we further build a deep architecture with the application to citywide crowd flow prediction, which naturally incorporates the sequential and periodic data as well as other external influences  ...  As shown on the right of Fig. 2 , we employ ACFM to learn periodic representation with the periodic temporal features P as input.  ... 
arXiv:1809.00101v1 fatcat:axouhska3zg4hjhdey4foldvou

Sequential Prediction of Social Media Popularity with Deep Temporal Context Networks [article]

Bo Wu, Wen-Huang Cheng, Yongdong Zhang, Qiushi Huang, Jintao Li, Tao Mei
2017 arXiv   pre-print
Then, based on the embedded data sequence over time, temporal context learning attempts to recurrently learn two adaptive temporal contexts for sequential popularity.  ...  Finally, a novel temporal attention is designed to predict new popularity (the popularity of a new user-post pair) with temporal coherence across multiple time-scales.  ...  Based on the above idea, we propose two types of sequential temporal contexts for learning two different types of temporal coherence of popularity: Neighboring Temporal Context (NTC) and Periodic Temporal  ... 
arXiv:1712.04443v1 fatcat:uumtbych65cebfdiydp675fjii

PRNet: A Periodic Residual Learning Network for Crowd Flow Forecasting [article]

Chengxin Wang, Yuxuan Liang, Gary Tan
2021 arXiv   pre-print
Differing from existing methods, PRNet frames the crowd flow forecasting as a periodic residual learning problem by modeling the deviation between the input (the previous time period) and the output (the  ...  better predictions.  ...  It shows that with a powerful periodicity learning mechanism, one can generate favorable predictions even with a lightweight ST module.  ... 
arXiv:2112.06132v1 fatcat:l3g5sxg65bhr5n6zsq6rri2yc4

Spatial-Temporal Hypergraph Self-Supervised Learning for Crime Prediction [article]

Zhonghang Li and Chao Huang and Lianghao Xia and Yong Xu and Jian Pei
2022 arXiv   pre-print
While many efforts have been devoted to proposing various spatial-temporal forecasting techniques to explore dependence across locations and time periods, most of them follow a supervised learning manner  ...  Inspired by the recent success in self-supervised learning, this work proposes a Spatial-Temporal Hypergraph Self-Supervised Learning framework (ST-HSL) to tackle the label scarcity issue in crime prediction  ...  shifted attention is proposed to learn temporal patterns among different time periods.  ... 
arXiv:2204.08587v2 fatcat:fdjwlnuqjfhjxhpwxj5r7sysva

Learning temporal context shapes prestimulus alpha oscillations and improves visual discrimination performance

Tahereh Toosi, Ehsan K. Tousi, Hossein Esteky
2017 Journal of Neurophysiology  
Moreover, learning the temporal context shaped the prestimulus alpha power: modulation of prestimulus alpha power grew during the predictable block and correlated with performance enhancement.  ...  This modulation only occurred in the longest delay period, 800 ms, in which predictability also improved the behavioral performance of the subjects.  ...  This study aimed to understand how learning of the temporal context affects the cortical oscillations at different frequencies.  ... 
doi:10.1152/jn.00969.2016 pmid:28515289 pmcid:PMC5539440 fatcat:grod2sinendj7b4nchmv5texjq

Periodic-CRN: A Convolutional Recurrent Model for Crowd Density Prediction with Recurring Periodic Patterns

Ali Zonoozi, Jung-jae Kim, Xiao-Li Li, Gao Cong
2018 Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence  
explicit periodic representations, and can be optimized with multi-step ahead prediction.  ...  To address this lack, we propose novel 'Periodic-CRN' (PCRN) method, which adapts convolutional recurrent network (CRN) to accurately capture spatial and temporal correlations, learns and incorporates  ...  We also do an experiment on multi-step ahead prediction with different interval durations.  ... 
doi:10.24963/ijcai.2018/519 dblp:conf/ijcai/ZonooziKLC18 fatcat:jithvapw3babtiqeym73znomxm

Dynamic Spatial-Temporal Representation Learning for Traffic Flow Prediction [article]

Lingbo Liu, Jiajie Zhen, Guanbin Li, Geng Zhan, Zhaocheng He, Bowen Du, Liang Lin
2020 arXiv   pre-print
Further, we develop two deep learning frameworks based on ATFM to predict citywide short-term/long-term traffic flow by adaptively incorporating the sequential and periodic data as well as other external  ...  To address this problem, we propose a unified neural network called Attentive Traffic Flow Machine (ATFM), which can effectively learn the spatial-temporal feature representations of traffic flow with  ...  traffic flow prediction; • PRNN-w/o-Attention: takes periodic features P in as input and learns periodic representation with a LSTM layer to predict future traffic flow; • PRNN: takes periodic features  ... 
arXiv:1909.02902v4 fatcat:w3fm72776nh7vg7toydupefdk4

Topical Behavior Prediction from Massive Logs [article]

Shih-Chieh Su
2017 arXiv   pre-print
Both the temporal and the spatial relationships of the behavior are explored with the deep learning architectures combing the recurrent neural network (RNN) and the convolutional neural network (CNN).  ...  A new learning framework called the spatially connected convolutional networks (SCCN) is introduced to predict the topical metrics more efficiently.  ...  Temporal Gain The temporal gain is observed by inspecting the prediction difference between the MLP and the TDRN.  ... 
arXiv:1708.03381v1 fatcat:iy33tpxnbraz3gt2fudius6cgy

Spatiotemporal Graph Convolutional Network for Multi-Scale Traffic Forecasting

Yi Wang, Changfeng Jing
2022 ISPRS International Journal of Geo-Information  
However, the existing work focused on the overall traffic network instead of traffic nodes, and the latter can be useful in learning different patterns among nodes.  ...  Tres2GCN was validated by comparing it with 10 baseline methods using two public traffic volume datasets.  ...  time periods on prediction (with PeMS08 as an example).  ... 
doi:10.3390/ijgi11020102 fatcat:vfbjdc7t3bfyre3cbu5df3m4mm

Sequential Prediction of Social Media Popularity with Deep Temporal Context Networks

Bo Wu, Wen-Huang Cheng, Yongdong Zhang, Qiushi Huang, Jintao Li, Tao Mei
2017 Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence  
Then, based on the embedded data sequence over time, temporal context learning attempts to recurrently learn two adaptive temporal contexts for sequential popularity.  ...  Finally, a novel temporal attention is designed to predict new popularity (the popularity of a new user-post pair) with temporal coherence across multiple time-scales.  ...  Based on the above idea, we propose two types of sequential temporal contexts for learning two different types of temporal coherence of popularity: Neighboring Temporal Context (NTC) and Periodic Temporal  ... 
doi:10.24963/ijcai.2017/427 dblp:conf/ijcai/WuCZHLM17 fatcat:75qhgtqbcvar3cd7yf2kxbjtka

Behavioral dynamics on the web

Kira Radinsky, Krysta M. Svore, Susan T. Dumais, Milad Shokouhi, Jaime Teevan, Alex Bocharov, Eric Horvitz
2013 ACM Transactions on Information Systems  
We develop a temporal modeling framework adapted from physics and signal processing and harness it to predict temporal patterns in search behavior using smoothing, trends, periodicities and surprises.  ...  We also develop a novel methodology that learns to select the best prediction model from a family of predictive models for a given query or a class of queries.  ...  Queries with many different distinct search results clicked at different times benefit from temporal modeling.  ... 
doi:10.1145/2493175.2493181 fatcat:uuchmtxa5zcd5aw2tyrihrq2li

AdaRNN: Adaptive Learning and Forecasting of Time Series [article]

Yuntao Du, Jindong Wang, Wenjie Feng, Sinno Pan, Tao Qin, Renjun Xu, Chongjun Wang
2021 arXiv   pre-print
Second, we propose Temporal Distribution Matching to reduce the distribution mismatch in TS to learn the adaptive TS model.  ...  AdaRNN is a general framework with flexible distribution distances integrated.  ...  To learn a prediction model with good generalization performance under TCS, a crucial research issue is to capture the common knowledge shared among different periods of D [22] .  ... 
arXiv:2108.04443v2 fatcat:ragucd7j7jb2zi4la2khiiw36a

Cross-City Transfer Learning for Deep Spatio-Temporal Prediction [article]

Leye Wang, Xu Geng, Xiaojuan Ma, Feng Liu, Qiang Yang
2018 arXiv   pre-print
To address the problem, we propose a novel cross-city transfer learning method for deep spatio-temporal prediction tasks, called RegionTrans.  ...  Spatio-temporal prediction is a key type of tasks in urban computing, e.g., traffic flow and air quality. Adequate data is usually a prerequisite, especially when deep learning is adopted.  ...  Recently, with the development of big data techniques, deep learning becomes popular in spatio-temporal prediction, e.g. crowd flow, taxi demand, precipitation predictions, and achieves state-of-the-art  ... 
arXiv:1802.00386v2 fatcat:54wqpv4gzfbwbolnev653fxyia

Predicting Temporal Sets with Deep Neural Networks [article]

Le Yu, Leilei Sun, Bowen Du, Chuanren Liu, Hui Xiong, Weifeng Lv
2020 arXiv   pre-print
a predictive model with the latent representations.  ...  Many possible existing methods, if adapted for the problem of temporal sets prediction, usually follow a two-step strategy by first projecting temporal sets into latent representations and then learning  ...  RELATED WORK This section reviews the existing literature related to our work, and also points out the differences of previous studies with our research. Next-period Set Prediction.  ... 
arXiv:2006.11483v3 fatcat:5bciwl4lszhphbtnokegvprxii
« Previous Showing results 1 — 15 out of 309,907 results