Encoding geospatial data is crucial for enabling machine learning (ML) models to perform tasks that require spatial reasoning, such as identifying the topological relationships between two different geospatial objects. However, existing encoding methods are limited as they are typically customized to handle only specific types of spatial data, which impedes their applicability across different downstream tasks where multiple data types coexist. To address this, we introduce Poly2Vec, an encoding framework that unifies the modeling of different geospatial objects, including 2D points, polylines, and polygons, irrespective of the downstream task. We leverage the power of the 2D Fourier transform to encode useful spatial properties, such as shape and location, from geospatial objects into fixed-length vectors. These vectors are then inputted into neural network models for spatial reasoning tasks.This unified approach eliminates the need to develop and train separate models for each distinct spatial type. We evaluate Poly2Vec on both synthetic and real datasets of mixed geometry types and verify its consistent performance across several downstream spatial reasoning tasks.

Learning effective geospatial embeddings is crucial for a series of geospatial applications such as city analytics and earth monitoring. However, learning comprehensive region representations presents two significant challenges: first, the deficiency of effective intra-region feature representation; and second, the difficulty of learning from intricate inter-region dependencies. In this paper, we present GeoHG, an effective heterogeneous graph structure for learning comprehensive region embeddings for various downstream tasks. Specifically, we tailor satellite image representation learning through geo-entity segmentation and point-of-interest (POI) integration for expressive intra-regional features. Furthermore, GeoHG unifies informative spatial interdependencies and socio-environmental attributes into a powerful heterogeneous graph to encourage explicit modeling of higher-order inter-regional relationships. The intra-regional features and inter-regional correlations are seamlessly integrated by a model-agnostic graph learning framework for diverse downstream tasks. Extensive experiments demonstrate the effectiveness of GeoHG in geo-prediction tasks compared to existing methods, even under extreme data scarcity (with just 5% of training data). With interpretable region representations, GeoHG exhibits strong generalization capabilities across regions. We will release code and data upon paper notification.

With the rise of electronic data, particularly Earth observation data, data-based geospatial modelling using machine learning (ML) has gained popularity in environmental research. Accurate geospatial predictions are vital for domain research based on ecosystem monitoring and quality assessment and for policy-making and action planning, considering effective management of natural resources. The accuracy and computation speed of ML has generally proved efficient. However, many questions have yet to be addressed to obtain precise and reproducible results suitable for further use in both research and practice. A better understanding of the ML concepts applicable to geospatial problems enhances the development of data science tools providing transparent information crucial for making decisions on global challenges such as biosphere degradation and climate change. This survey reviews common nuances in geospatial modelling, such as imbalanced data, spatial autocorrelation, prediction errors, model generalisation, domain specificity, and uncertainty estimation. We provide an overview of techniques and popular programming tools to overcome or account for the challenges. We also discuss prospects for geospatial Artificial Intelligence in environmental applications. To date, obtaining spatial predictions is an essential step in the monitoring, assessment, and prognosis tasks applicable to all kinds of Earth systems on both local and global scales (Figure 1).

Visual object tracking has seen significant progress in recent years. However, the vast majority of this work focuses on tracking objects within the image plane of a single camera and ignores the uncertainty associated with predicted object locations. In this work, we focus on the geospatial object tracking problem using data from a distributed camera network. The goal is to predict an object's track in geospatial coordinates along with uncertainty over the object's location while respecting communication constraints that prohibit centralizing raw image data. We present a novel single-object geospatial tracking data set that includes high-accuracy ground truth object locations and video data from a network of four cameras. We present a modeling framework for addressing this task including a novel backbone model and explore how uncertainty calibration and fine-tuning through a differentiable tracker affect performance.

SAN MATEO, Calif., July 29, 2021 (GLOBE NEWSWIRE) -- Chooch AI, the leading computer vision AI platform, has been cited for accelerating adoption of computer vision–powered solutions across industry verticals by leading research company IDC. Chooch AI models are ready to deploy now both in the cloud and on edge devices. Clients include Fortune 500 companies and the US Government. Partners include NVIDIA, Intel, Dell, Deloitte, Convergint and Vantiq. IDC states that, "Chooch AI's horizontal- and vertical-agnostic platform supports rapid data set generation capabilities using machine labeling techniques such as smart annotation, data augmentation, and use of synthetic data, along with pretrained ready-to-use models. They believe this will accelerate adoption and time to value computer vision–powered solutions across industry verticals."

In the near future artificial intelligence is going to be as fundamental for business success as cloud is becoming to running a company's information technology. And AI is where cloud was five years ago. People are realizing there's value here, and business leaders see that AI can help them fundamentally change how their company works, how they get work done, and how they serve customers. In a recent survey of 1200 companies, Cognizant and ESI ThoughtLab found that 64% of executives believe AI will be important to the future of their business. For the largest organizations, the figure was a stunning 85 percent!

There is increased recognition for better methods to measure, predict and adjust for social risk factors in healthcare and population health. Current performance and quality measures generally do not take SRFs into account in the bonus/penalty structure, nor are SRFs generally included in most risk adjustment formulas. This can lead to unintended consequences, including the potential to perpetuate bias and disparities in health outcomes. To mitigate these issues, RTI International is developing an "artificially intelligent" approach to risk adjustment for SRFs using random forests to understand life expectancy variances at the census tract level. So how can AI help recognize how local area factors are independently associated with many health outcomes and may be informative either in conjunction with individual-level data or on their own?

The fourth industrial revolution and oncoming of the second machine -- as it is being often hailed -- has led to new disruptive technologies emerging, even as many bit the dust in the last few years. So how do CEOs and their top teams even begin to make sense of the swirl of technological breakthroughs affecting business today? How do they gauge the impact of artificial intelligence on their companies' future compared with, say, the Internet of Things or virtual reality? A PwC study some time back identified top eight disruptive technologies that matter now, and will have far-reaching impact in the days to come. The team tracked more than 150 discrete technologies and analyzed for technologies with the most cross-industry and global impact over the coming years.