Traffic prediction has drawn increasing attention in AI research field due to the increasing availability of large-scale traffic data and its importance in the real world. For example, an accurate taxi demand prediction can assist taxi companies in pre-allocating taxis. The key challenge of traffic prediction lies in how to model the complex spatial dependencies and temporal dynamics. Although both factors have been considered in modeling, existing works make strong assumptions about spatial dependence and temporal dynamics, i.e., spatial dependence is stationary in time, and temporal dynamics is strictly periodical. However, in practice the spatial dependence could be dynamic (i.e., changing from time to time), and the temporal dynamics could have some perturbation from one period to another period. In this paper, we make two important observations: (1) the spatial dependencies between locations are dynamic; and (2) the temporal dependency follows daily and weekly pattern but it is not strictly periodic for its dynamic temporal shifting. To address these two issues, we propose a novel Spatial-Temporal Dynamic Network (STDN), in which a flow gating mechanism is introduced to learn the dynamic similarity between locations, and a periodically shifted attention mechanism is designed to handle long-term periodic temporal shifting. To the best of our knowledge, this is the first work that tackle both issues in a unified framework. Our experimental results on real-world traffic datasets verify the effectiveness of the proposed method.

Periodicity is a frequently happening phenomenon for moving objects. Finding periodic behaviors is essential to understanding object movements. However, periodic behaviors could be complicated, involving multiple interleaving periods, partial time span, and spatiotemporal noises and outliers.In this paper, we address the problem of mining periodic behaviors for moving objects. It involves two sub-problems: how to detect the periods in complex movement, and how to mine periodic movement behaviors. Our main assumption is that the observed movement is generated from multiple interleaved periodic behaviors associated with certain reference locations. Based on this assumption, we propose a twostage algorithm, Periodica, to solve the problem. At the first stage, the notion of reference spot is proposed to capture the reference locations. Through reference spots, multiple periods in the movement can be retrieved using a method that combines Fourier transform and autocorrelation. At the second stage, a probabilistic model is proposed to characterize the periodic behaviors. For a specific period, periodic behaviors are statistically generalized from partial movement sequences through hierarchical clustering. Empirical studies on both synthetic and real data sets demonstrate the effectiveness of our method.

Eleven thousand groundwater samples collected in the 2010s in an area of Marcellus shale-gas development are analyzed to assess spatial and temporal patterns of water quality. Using a new data mining technique, we confirm previous observations that methane concentrations in groundwater tend to be naturally elevated in valleys and near faults, but we also show that methane is also more concentrated near an anticline. Data mining also highlights waters with elevated methane that are not otherwise explained by geologic features. These slightly elevated concentrations occur near 7 out of the 1,385 shale-gas wells and near some conventional gas wells in the study area. For ten analytes for which uncensored data are abundant in this 3,000 km rural region, concentrations are unchanged or improved as compared to samples analyzed prior to 1990. Specifically, TDS, Fe, Mn, sulfate, and pH show small but statistically significant improvement, and As, Pb, Ba, Cl, and Na show no change. Evidence from this rural area could document improved groundwater quality caused by decreased acid rain (pH, sulfate) since the imposition of the Clean Air Act or decreased steel production (Fe, Mn). Such improvements have not been reported in groundwater in more developed areas of the U.S.

In this study, we have utilized electrophysiological single unit recordings to evaluate the effects of nonpeptidergic angiotensin II (ANG II) antagonists on neural responses of hypothalamic paraventricular nucleus (PVN) neurons to either electrical stimulation in subfornical organ (SFO) or direct application of ANG II. Electrical stimulation (200-400 microA; 0.1 ms) in the SFO resulted in excitatory responses in 36 of 50 PVN neurons tested. Peristimulus histogram analysis of such excitatory effects demonstrated latencies of < 30 ms and variability of response times of approximately 50 ms in 14 of these 36 neurons. In view of previous anatomic and electrophysiological studies such inputs were therefore considered to be monosynaptically mediated by direct neural inputs from the SFO. The remaining 22 cells excited by such SFO stimulation showed responses of longer latency and duration suggestive of a different underlying synaptic mechanism. Local pressure ejection of ANG II into the PVN resulted in increased neural activity in 50% (9 of 18) of the neurons tested. After systemic (3 mg/kg iv) or local (2 x 10(-2) M; 1-25 s; 2-40 psi) microinjection of the nonpeptidergic angiotensin II1 (AT1) receptor antagonist losartan, SFO excitations were attenuated in 63.9% (23 of 36) of the PVN neurons tested, such pharmacologically blocked excitatory responses being reduced by 68.3 +/- 5.2% from control stimulation effects (P < 0.001). Similar losartan-induced attenuations of both short latency (presumed monosynaptic) (50.0%) and longer latency (72.7%) responses were observed. In addition, losartan also abolished the excitatory effects of local administration of ANG II on 77.8% (7 of 9) of ANG II-sensitive neurons in PVN tested.(ABSTRACT TRUNCATED AT 250 WORDS)

Background Recent studies indicate important roles for long noncoding RNAs (lncRNAs) in the regulation of gene expression by acting as competing endogenous RNAs (ceRNAs). However, the specific role of lncRNAs in skeletal muscle atrophy is still unclear. Our study aimed to identify the function of lncRNAs that control skeletal muscle myogenesis and atrophy. Methods RNA sequencing was performed to identify the skeletal muscle transcriptome (lncRNA and messenger RNA) between hypertrophic broilers and leaner broilers. To study the ‘sponge’ function of lncRNA, we constructed a lncRNA‐microRNA (miRNA)‐gene interaction network by integrated our previous submitted skeletal muscle miRNA sequencing data. The primary myoblast cells and animal model were used to assess the biological function of the lncIRS1 in vitro or in vivo . Results We constructed a myogenesis‐associated lncRNA‐miRNA‐gene network and identified a novel ceRNA lncRNA named lncIRS1 that is specifically enriched in skeletal muscle. LncIRS1 could regulate myoblast proliferation and differentiation in vitro , and muscle mass and mean muscle fibre in vivo . LncIRS1 increases gradually during myogenic differentiation. Mechanistically, lncIRS1 acts as a ceRNA for miR‐15a, miR‐15b‐5p, and miR‐15c‐5p to regulate IRS1 expression, which is the downstream of the IGF1 receptor. Overexpression of lncIRS1 not only increased the protein abundance of IRS1 but also promoted phosphorylation level of AKT (p‐AKT) a central component of insulin‐like growth factor‐1 pathway. Furthermore, lncIRS1 regulates the expression of atrophy‐related genes and can rescue muscle atrophy. Conclusions The newly identified lncIRS1 acts as a sponge for miR‐15 family to regulate IRS1 expression, resulting in promoting skeletal muscle myogenesis and controlling atrophy.

Taxi demand prediction is an important building block to enabling intelligent transportation systems in a smart city. An accurate prediction model can help the city pre-allocate resources to meet travel demand and to reduce empty taxis on streets which waste energy and worsen the traffic congestion. With the increasing popularity of taxi requesting services such as Uber and Didi Chuxing (in China), we are able to collect large-scale taxi demand data continuously. How to utilize such big data to improve the demand prediction is an interesting and critical real-world problem. Traditional demand prediction methods mostly rely on time series forecasting techniques, which fail to model the complex non-linear spatial and temporal relations. Recent advances in deep learning have shown superior performance on traditionally challenging tasks such as image classification by learning the complex features and correlations from large-scale data. This breakthrough has inspired researchers to explore deep learning techniques on traffic prediction problems. However, existing methods on traffic prediction have only considered spatial relation (e.g., using CNN) or temporal relation (e.g., using LSTM) independently. We propose a Deep Multi-View Spatial-Temporal Network (DMVST-Net) framework to model both spatial and temporal relations. Specifically, our proposed model consists of three views: temporal view (modeling correlations between future demand values with near time points via LSTM), spatial view (modeling local spatial correlation via local CNN), and semantic view (modeling correlations among regions sharing similar temporal patterns). Experiments on large-scale real taxi demand data demonstrate effectiveness of our approach over state-of-the-art methods.

The increased availability of large-scale trajectory data provides rich information for the study of urban dynamics. For example, New York City Taxi & Limousine Commission regularly releases source/destination information of taxi trips, where 173 million taxi trips released for Year 2013 [1]. Such a big dataset provides us potential new perspectives to address the traditional traffic problems. In this paper, we study the travel time estimation problem. Instead of following the traditional route-based travel time estimation, we propose to simply use a large amount of taxi trips without using the intermediate trajectory points to estimate the travel time between source and destination. Our experiments show very promising results. The proposed big data-driven approach significantly outperforms both state-of-the-art route-based method and online map services. Our study indicates that novel simple approaches could be empowered by the big data and these approaches could serve as new baselines for some traditional computational problems.

Traffic congestion plagues cities around the world. Recent years have witnessed an unprecedented trend in applying reinforcement learning for traffic signal control. However, the primary challenge is to control and coordinate traffic lights in large-scale urban networks. No one has ever tested RL models on a network of more than a thousand traffic lights. In this paper, we tackle the problem of multi-intersection traffic signal control, especially for large-scale networks, based on RL techniques and transportation theories. This problem is quite difficult because there are challenges such as scalability, signal coordination, data feasibility, etc. To address these challenges, we (1) design our RL agents utilizing ‘pressure’ concept to achieve signal coordination in region-level; (2) show that implicit coordination could be achieved by individual control agents with well-crafted reward design thus reducing the dimensionality; and (3) conduct extensive experiments on multiple scenarios, including a real-world scenario with 2510 traffic lights in Manhattan, New York City 1 2.