In the wake of the COVID-19 pandemic, the World Bank's 2020 Global Economic Prospects forecasts a baseline global GDP contraction of 5.2 percent, making it the deepest global recession in decades. Between 71 to 100 million people are expected to be pushed into extreme poverty, almost half of them in South Asia and more than a third in Sub-Saharan Africa. As a result, since March 2020 over 215 countries and territories have implemented 1,414 social protection measures to respond to the pandemic and ensuing economic crisis. Social assistance programs account for 62 percent of all social protection response measures, half of them being cash-based transfers of some sort. This major shock has revealed the many challenges governments face when attempting to quickly respond to crises in order to protect the poor and vulnerable. Providing timely assistance and support to those households most in need can increase their resilience and reduce the negative impacts of the shock on their short and medium-term well-being. Nonetheless, the lack of readily available and up-to-date socioeconomic data necessary to prioritize shock-responsive social protection measures is an important binding constraint for many governments in developing countries. This paper presents a portion of our work on a larger project with the World Bank to identify the most vulnerable populations in these countries. Having timely access to such information, particularly in data-deprived contexts, can improve the capacity of governments to design and operationalize better and more shock-responsive social protection measures.

Accurate spatial-temporal traffic flow forecasting is essential for helping traffic managers to take control measures and drivers to choose the optimal travel routes. Recently, graph convolutional networks (GCNs) have been widely used in traffic flow prediction owing to their powerful ability to capture spatial-temporal dependencies. The design of the spatial-temporal graph adjacency matrix is a key to the success of GCNs, and it is still an open question. This paper proposes reconstructing the binary adjacency matrix via tensor decomposition, and a traffic flow forecasting method is proposed. First, we reformulate the spatial-temporal fusion graph adjacency matrix into a three-way adjacency tensor. Then, we reconstructed the adjacency tensor via Tucker decomposition, wherein more informative and global spatial-temporal dependencies are encoded. Finally, a Spatial-temporal Synchronous Graph Convolutional module for localized spatial-temporal correlations learning and a Dilated Convolution module for global correlations learning are assembled to aggregate and learn the comprehensive spatial-temporal dependencies of the road network. Experimental results on four open-access datasets demonstrate that the proposed model outperforms state-of-the-art approaches in terms of the prediction performance and computational cost.

Accurate and timely rain prediction is crucial for decision making and is also a challenging task. This paper presents a solution which won the 2 nd prize in the Weather4cast 2022 NeurIPS competition using 3D U-Nets and EarthFormers for 8-hour probabilistic rain prediction based on multi-band satellite images. The spatial context effect of the input satellite image has been deeply explored and optimal context range has been found. Based on the imbalanced rain distribution, we trained multiple models with different loss functions. To further improve the model performance, multi-model ensemble and threshold optimization were used to produce the final probabilistic rain prediction. Experiment results and leaderboard scores demonstrate that optimal spatial context, combined loss function, multi-model ensemble, and threshold optimization all provide modest model gain. A permutation test was used to analyze the effect of each satellite band on rain prediction, and results show that satellite bands signifying cloudtop phase (8.7 um) and cloud-top height (10.8 and 13.4 um) are the best predictors for rain prediction. The source code is available at https://github.com/bugsuse/weather4cast-2022-stage2.

Deep Learning has significantly impacted the application of data-to-decision throughout research and industry, however, they lack a rigorous mathematical foundation, which creates situations where algorithmic results fail to be practically invertible. In this paper we present a nearly invertible mapping between $\mathbb{R}^{2^n}$ and $\mathbb{R}^{n+1}$ via a topological connection between $S^{2^n-1}$ and $S^n$. Throughout the paper we utilize the algebra of Multicomplex rotation groups and polyspherical coordinates to define two maps: the first is a contraction from $S^{2^n-1}$ to $\displaystyle \otimes^n_{k=1} SO(2)$, and the second is a projection from $\displaystyle \otimes^n_{k=1} SO(2)$ to $S^{n}$. Together these form a composite map that we call the LG Fibration. In analogy to the generation of Hopf Fibration using Hypercomplex geometry from $S^{(2n-1)} \mapsto CP^n$, our fibration uses Multicomplex geometry to project $S^{2^n-1}$ onto $S^n$. We also investigate the algebraic properties of the LG Fibration, ultimately deriving a distance difference function to determine which pairs of vectors have an invariant inner product under the transformation. The LG Fibration has applications to Machine Learning and AI, in analogy to the current applications of Hopf Fibrations in adaptive UAV control. Furthermore, the ability to invert the LG Fibration for nearly all elements allows for the development of Machine Learning algorithms that may avoid the issues of uncertainty and reproducibility that currently plague contemporary methods. The primary result of this paper is a novel method of nearly invertible geometric dimensional reduction from $S^{2^n-1}$ to $S^n$, which has the capability to extend the research in both mathematics and AI, including but not limited to the fields of homotopy groups of spheres, algebraic topology, machine learning, and algebraic biology.

GPS trajectories are the essential foundations for many trajectory-based applications, such as travel time estimation, traffic prediction and trajectory similarity measurement. Most applications require a large amount of high sample rate trajectories to achieve a good performance. However, many real-life trajectories are collected with low sample rate due to energy concern or other constraints.We study the task of trajectory recovery in this paper as a means for increasing the sample rate of low sample trajectories. Currently, most existing works on trajectory recovery follow a sequence-to-sequence diagram, with an encoder to encode a trajectory and a decoder to recover real GPS points in the trajectory. However, these works ignore the topology of road network and only use grid information or raw GPS points as input. Therefore, the encoder model is not able to capture rich spatial information of the GPS points along the trajectory, making the prediction less accurate and lack spatial consistency. In this paper, we propose a road network enhanced transformer-based framework, namely RNTrajRec, for trajectory recovery. RNTrajRec first uses a graph model, namely GridGNN, to learn the embedding features of each road segment. It next develops a spatial-temporal transformer model, namely GPSFormer, to learn rich spatial and temporal features along with a Sub-Graph Generation module to capture the spatial features for each GPS point in the trajectory. It finally forwards the outputs of encoder model into a multi-task decoder model to recover the missing GPS points. Extensive experiments based on three large-scale real-life trajectory datasets confirm the effectiveness of our approach.

Spatial-temporal (ST) graph modeling, such as traffic speed forecasting and taxi demand prediction, is an important task in deep learning area. However, for the nodes in graph, their ST patterns can vary greatly in difficulties for modeling, owning to the heterogeneous nature of ST data. We argue that unveiling the nodes to the model in a meaningful order, from easy to complex, can provide performance improvements over traditional training procedure. The idea has its root in Curriculum Learning which suggests in the early stage of training models can be sensitive to noise and difficult samples. In this paper, we propose ST-Curriculum Dropout, a novel and easy-to-implement strategy for spatial-temporal graph modeling. Specifically, we evaluate the learning difficulty of each node in high-level feature space and drop those difficult ones out to ensure the model only needs to handle fundamental ST relations at the beginning, before gradually moving to hard ones. Our strategy can be applied to any canonical deep learning architecture without extra trainable parameters, and extensive experiments on a wide range of datasets are conducted to illustrate that, by controlling the difficulty level of ST relations as the training progresses, the model is able to capture better representation of the data and thus yields better generalization.

Camera and radar sensors have significant advantages in cost, reliability, and maintenance compared to LiDAR. Existing fusion methods often fuse the outputs of single modalities at the result-level, called the late fusion strategy. This can benefit from using off-the-shelf single sensor detection algorithms, but late fusion cannot fully exploit the complementary properties of sensors, thus having limited performance despite the huge potential of camera-radar fusion. Here we propose a novel proposal-level early fusion approach that effectively exploits both spatial and contextual properties of camera and radar for 3D object detection. Our fusion framework first associates image proposal with radar points in the polar coordinate system to efficiently handle the discrepancy between the coordinate system and spatial properties. Using this as a first stage, following consecutive cross-attention based feature fusion layers adaptively exchange spatio-contextual information between camera and radar, leading to a robust and attentive fusion. Our camera-radar fusion approach achieves the state-of-the-art 41.1% mAP and 52.3% NDS on the nuScenes test set, which is 8.7 and 10.8 points higher than the camera-only baseline, as well as yielding competitive performance on the LiDAR method.

In this paper, we attempt to detect an inflection or change-point resulting from the Covid-19 pandemic on supply chain data received from a large furniture company. To accomplish this, we utilize a modified CUSUM (Cumulative Sum) procedure on the company's spatial-temporal order data as well as a GLR (Generalized Likelihood Ratio) based method. We model the order data using the Hawkes Process Network, a multi-dimensional self and mutually exciting point process, by discretizing the spatial data and treating each order as an event that has a corresponding node and time. We apply the methodologies on the company's most ordered item on a national scale and perform a deep dive into a single state. Because the item was ordered infrequently in the state compared to the nation, this approach allows us to show efficacy upon different degrees of data sparsity. Furthermore, it showcases use potential across differing levels of spatial detail.

Big streams of Earth images from satellites or other platforms (e.g., drones and mobile phones) are becoming increasingly available at low or no cost and with enhanced spatial and temporal resolution. This thesis recognizes the unprecedented opportunities offered by the high quality and open access Earth observation data of our times and introduces novel machine learning and big data methods to properly exploit them towards developing applications for sustainable and resilient agriculture. The thesis addresses three distinct thematic areas, i.e., the monitoring of the Common Agricultural Policy (CAP), the monitoring of food security and applications for smart and resilient agriculture. The methodological innovations of the developments related to the three thematic areas address the following issues: i) the processing of big Earth Observation (EO) data, ii) the scarcity of annotated data for machine learning model training and iii) the gap between machine learning outputs and actionable advice. This thesis demonstrated how big data technologies such as data cubes, distributed learning, linked open data and semantic enrichment can be used to exploit the data deluge and extract knowledge to address real user needs. Furthermore, this thesis argues for the importance of semi-supervised and unsupervised machine learning models that circumvent the ever-present challenge of scarce annotations and thus allow for model generalization in space and time. Specifically, it is shown how merely few ground truth data are needed to generate high quality crop type maps and crop phenology estimations. Finally, this thesis argues there is considerable distance in value between model inferences and decision making in real-world scenarios and thereby showcases the power of causal and interpretable machine learning in bridging this gap.

Depending on what you work on, you may or may have not heard of Voronoi Diagrams. Voronoi diagrams are all about finding the closest point. We use this in our everyday lives, from trying to find the closest supermarket, train station, etc. A plane that is divided up into cells, covering a specific region that is close to a particular point -- that point is what we are trying to find. This plane is known as a Voronoi diagram and was named after Georgy Voronoi, a Ukrainian mathematician.