An official website of the United States government
Abstract Many cities are experiencing increases in extreme heat because of global temperature rise combined with the urban heat island effect. The heterogeneity of urban morphology also leads to fine-scale variability in potential for heat exposure. Yet, how this rise in temperature and local variability together impacts urban residents differently at exposure-relevant scales is still not clear. Here we map the Universal Thermal Climate Index, a more complete indicator of human heat stress at an unprecedentedly fine spatial resolution (1 m), for 14 major cities in the United States using urban microclimate modeling. We examined the different heat exposure levels across different socioeconomic and racial/ethnic groups in these cities, finding that income level is most consistently associated with heat stress. We further conducted scenario simulations for a hypothetical 1 °C increase of air temperature in all cities. Results show that a 1 °C increase would have a substantial impact on human heat stress, with impacts that differ across cities. The results of this study can help us better evaluate the impact of extreme heat on urban residents at decision-relevant scales.
more »
« less
Point-to-Region Co-learning for Poverty Mapping at High Resolution Using Satellite Imagery
Despite improvements in safe water and sanitation services in low-income countries, a substantial proportion of the population in Africa still does not have access to these essential services. Up-to-date fine-scale maps of low-income settlements are urgently needed by authorities to improve service provision. We aim to develop a cost-effective solution to generate fine-scale maps of these vulnerable populations using multi-source public information. The problem is challenging as ground-truth maps are available at only a limited number of cities, and the patterns are heterogeneous across cities. Recent attempts tackling the spatial heterogeneity issue focus on scenarios where true labels partially exist for each input region, which are unavailable for the present problem. We propose a dynamic point-to-region co-learning framework to learn heterogeneity patterns that cannot be reflected by point-level information and generalize deep learners to new areas with no labels. We also propose an attention-based correction layer to remove spurious signatures, and a region-gate to capture both region-invariant and variant patterns. Experiment results on real-world fine-scale data in three cities of Kenya show that the proposed approach can largely improve model performance on various base network architectures.
more »
« less
- Award ID(s):
- 2147195
- PAR ID:
- 10506518
- Publisher / Repository:
- AAAI
- Date Published:
- Journal Name:
- Proceedings of the AAAI Conference on Artificial Intelligence
- Volume:
- 37
- Issue:
- 12
- ISSN:
- 2159-5399
- Page Range / eLocation ID:
- 14321 to 14328
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This paper investigates the long-term impacts of the federal Home Owners’ Loan Corporation (HOLC) mortgage risk assessment maps on the spatial dynamics of recent income and racial distributions in California metropolitan areas over the 1990-2010 period. We combine historical HOLC boundaries with modern Census tract data and apply recently developed methods of spatial distribution dynamics to examine if legacy impacts are reflected in recent urban dynamics. Cities with HOLC assessments are found to have higher levels of isolation segregation than the non-HOLC group, but no difference in unevenness segregation between the two groups of cities are found. We find no difference in income or racial and ethnic distributional dynamics between the two groups of cities over the period. At the intra-urban scale, we find that the intersectionality of residing in a C or D graded tract that is also a low-income tract falls predominately upon the minority populations in these eight HOLC cities. Our findings indicate that neighborhoods with poor housing markets and high minority concentrations rarely experience a dramatic change in either their racial and ethnic or socioeconomic compositions—and that negative externalities (e.g. lower home prices and greater segregation levels) emanate from these neighborhoods, with inertia spilling over into nearby zones.more » « less
-
Neighborhood ethnicity is related to mammal occupancy and activity across a desert metropolitan areaAbstract Cities support abundant human and wildlife populations that are shaped indirectly and directly by human decisions, often resulting in unequal access to environmental services and accessible open spaces. Urban land cover drives biodiversity patterns across metropolitan areas, but at smaller scales that matter to local residents, neighborhood socio‐cultural factors can influence the presence and abundance of wildlife. Neighborhood income is associated with plant and animal diversity in some cities, but the influence of other social variables is less well understood, especially across desert ecosystems. We explored wildlife distribution across gradients of neighborhood ethnicity in addition to income and landscape characteristics within residential areas of metropolitan Phoenix, Arizona, USA. Utilizing data from 38 wildlife cameras deployed in public parks and undeveloped open spaces within or near suburban neighborhoods, we estimated occupancy and activity patterns of common mammal species, including species native to the Sonoran Desert (coyote [Canis latrans] and desert cottontail rabbit [Sylvilagus audubonii]), and non‐native domestic cat (Felis catus). Neighborhood ethnicity (percentage of Latino residents) appeared to exhibit a negative relationship with occupancy for coyotes and cottontail rabbits. Additionally, daily activity patterns of coyotes occurred later in the evenings and mornings in neighborhoods with higher proportions of Latino residents, but activity was unaffected by differences in neighborhood income. This study is one of the first to show that social‐ecological mechanisms associated with patterns of neighborhood ethnicity as well as income may help to shape wildlife distribution in cities. These findings have implications for equitable management and provisioning of ecosystem services for urban residents and highlight the importance of considering a range of social covariates to better understand biodiversity outcomes in urban and urbanizing areas.more » « less
-
Rapid urbanization has posed significant burden on urban transportation infrastructures. In today's cities, both private and public transits have clear limitations to fulfill passengers' needs for quality of experience (QoE): Public transits operate along fixed routes with long wait time and total transit time; Private transits, such as taxis, private shuttles and ride-hailing services, provide point-to-point transits with high trip fare. In this paper, we propose CityLines, a transformative urban transit system, employing hybrid hub-and-spoke transit model with shared shuttles. Analogous to Airlines services, the proposed CityLines system routes urban trips among spokes through a few hubs or direct paths, with travel time as short as private transits and fare as low as public transits. CityLines allows both point-to-point connection to improve the passenger QoE, and hub-and-spoke connection to reduce the system operation cost. Our evaluation results show that CityLines framework can achieve both short travel time and high ride-sharing ratio.more » « less
-
Recent relation extraction (RE) works have shown encouraging improvements by conducting contrastive learning on silver labels generated by distant supervision before fine-tuning on gold labels. Existing methods typically assume all these silver labels are accurate and treat them equally; however, distant supervision is inevitably noisy–some silver labels are more reliable than others. In this paper, we propose fine-grained contrastive learning (FineCL) for RE, which leverages fine-grained information about which silver labels are and are not noisy to improve the quality of learned relationship representations for RE. We first assess the quality of silver labels via a simple and automatic approach we call “learning order denoising,” where we train a language model to learn these relations and record the order of learned training instances. We show that learning order largely corresponds to label accuracy–early-learned silver labels have, on average, more accurate labels than later-learned silver labels. Then, during pre-training, we increase the weights of accurate labels within a novel contrastive learning objective. Experiments on several RE benchmarks show that FineCL makes consistent and significant performance gains over state-of-the-art methods.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1609/aaai.v37i12.26675
