More Like this

1. Seasonal Variations and Spatial Patterns of Arctic Cloud Changes in Association with Sea Ice Loss during 1950–2019 in ERA5

   https://doi.org/10.1175/JCLI-D-23-0117.1  

   Jenkins, Matthew T.; Dai, Aiguo; Deser, Clara (January 2024, Journal of Climate)

   Abstract The dynamic and thermodynamic mechanisms that link retreating sea ice to increased Arctic cloud amount and cloud water content are unclear. Using the fifth generation of the ECMWF Reanalysis (ERA5), the long-term changes between years 1950–79 and 1990–2019 in Arctic clouds are estimated along with their relationship to sea ice loss. A comparison of ERA5 to CERES satellite cloud fractions reveals that ERA5 simulates the seasonal cycle, variations, and changes of cloud fraction well over water surfaces during 2001–20. This suggests that ERA5 may reliably represent the cloud response to sea ice loss because melting sea ice exposes more water surfaces in the Arctic. Increases in ERA5 Arctic cloud fraction and water content are largest during October–March from ∼950 to 700 hPa over areas with significant (≥15%) sea ice loss. Further, regions with significant sea ice loss experience higher convective available potential energy (∼2–2.75 J kg⁻¹), planetary boundary layer height (∼120–200 m), and near-surface specific humidity (∼0.25–0.40 g kg⁻¹) and a greater reduction of the lower-tropospheric temperature inversion (∼3°–4°C) than regions with small (<15%) sea ice loss in autumn and winter. Areas with significant sea ice loss also show strengthened upward motion between 1000 and 700 hPa, enhanced horizontal convergence (divergence) of air, and decreased (increased) relative humidity from 1000 to 950 hPa (950–700 hPa) during the cold season. Analyses of moisture divergence, evaporation minus precipitation, and meridional moisture flux fields suggest that increased local surface water fluxes, rather than atmospheric motions, provide a key source of moisture for increased Arctic clouds over newly exposed water surfaces during October–March. Significance StatementSea ice loss has been shown to be a primary contributor to Arctic warming. Despite the evidence linking large sea ice retreat to Arctic warming, some studies have suggested that enhanced downwelling longwave radiation associated with increased clouds and water vapor is the primary reason for Arctic amplification. However, it is unclear how sea ice loss is linked to changes in clouds and water vapor in the Arctic. Here, we investigate the relationship between Arctic sea ice loss and changes in clouds using the ERA5 dataset. Improved knowledge of the relationship between Arctic sea ice loss and changes in clouds will help further our understanding of the role of the cloud feedback in Arctic warming. 

   more » « less
2. Direct Synthesis of Iterative Algorithms With Bounds on Achievable Worst-Case Convergence Rate

   https://doi.org/10.23919/ACC45564.2020.9147401  

   Lessard, Laurent; Seiler, Peter (July 2020, American Control Conference)

   Iterative first-order methods such as gradient descent and its variants are widely used for solving optimization and machine learning problems. There has been recent interest in analytic or numerically efficient methods for computing worst-case performance bounds for such algorithms, for example over the class of strongly convex loss functions. A popular approach is to assume the algorithm has a fixed size (fixed dimension, or memory) and that its structure is parameterized by one or two hyperparameters, for example a learning rate and a momentum parameter. Then, a Lyapunov function is sought to certify robust stability and subsequent optimization can be performed to find optimal hyperparameter tunings. In the present work, we instead fix the constraints that characterize the loss function and apply techniques from robust control synthesis to directly search over algorithms. This approach yields stronger results than those previously available, since the bounds produced hold over algorithms with an arbitrary, but finite, amount of memory rather than just holding for algorithms with a prescribed structure. 

   more » « less
3. Multivariate compound events drive historical floods and associated losses along the U.S. East and Gulf coasts

   https://doi.org/10.1038/s44304-025-00076-5  

   Ali, Javed; Wahl, Thomas; Morim, Joao; Enriquez, Alejandra; Gall, Melanie; Emrich, Christopher T (February 2025, npj Natural Hazards)

   Compound flooding events are a threat to many coastal regions and can have widespread socio-economic implications. However, their frequency of occurrence, underlying flood drivers, and direct link to past socio-economic losses are largely unknown despite being key to supporting risk and adaptation assessments. Here, we present an impact-based analysis of compound flooding for 203 coastal counties along the U.S. Gulf and East coasts by combining data from multiple flood drivers and socio-economic loss information from 1980 to 2018. We find that ~80% of all flood events recorded in our study area were compound rather than univariate. In addition, we show that historical compound flooding events in most counties were driven by more than two flood drivers (hydrological, meteorological, and/or oceanographic) and distinct spatial clusters exist that exhibit variability in the underlying driver of compound flood events. Furthermore, we find that in more than 80% of the counties, over 80% of recorded property and crop losses were linked to compound flooding. Nearly 80% of counties have a higher median loss from compound than univariate events. For these counties, the median property loss is over 26 times greater, and the median crop loss is over 76 times greater for compound events on average. Our analysis overcomes some of the limitations of previous compound-event studies based on pre-defined flood drivers and offers new insights into the complex relationship between hazards and associated socio-economic impacts. 

   more » « less
4. Deep Clustering with Incomplete Noisy Pairwise Annotations: A Geometric Regularization Approach

   Nguyen, Tri; Ibrahim, Shahana; Fu, Xiao (July 2023, Proceedings of the 40th International Conference on Machine Learning)

   The recent integration of deep learning and pairwise similarity annotation-based constrained clustering—i.e., deep constrained clustering (DCC)—has proven effective for incorporating weak supervision into massive data clustering: Less than 1% of pair similarity annotations can often substantially enhance the clustering accuracy. However, beyond empirical successes, there is a lack of understanding of DCC. In addition, many DCC paradigms are sensitive to annotation noise, but performance-guaranteed noisy DCC methods have been largely elusive. This work first takes a deep look into a recently emerged logistic loss function of DCC, and characterizes its theoretical properties. Our result shows that the logistic DCC loss ensures the identifiability of data membership under reasonable conditions, which may shed light on its effectiveness in practice. Building upon this understanding, a new loss function based on geometric factor analysis is proposed to fend against noisy annotations. It is shown that even under unknown annotation confusions, the data membership can still be provably identified under our proposed learning criterion. The proposed approach is tested over multiple datasets to validate our claims.} 

   more » « less
5. FACS-GCN: Fairness-Aware Cost-Sensitive Boosting of Graph Convolutional Networks

   https://doi.org/10.1109/IJCNN55064.2022.9892919  

   Santos, Francisco; Ye, Junke; Masrour, Farzan; Tan, Pang-Ning; Esfahanian, Abdol-Hossein (July 2022, Proceedings of the 2022 International Joint Conference on Neural Networks (IJCNN))

   Graph neural networks (GNNs) have emerged as a powerful tool for modeling graph data due to their ability to learn a concise representation of the data by integrating the node attributes and link information in a principled fashion. However, despite their promise, there are several practical challenges that must be overcome to effectively use them for node classification problems. In particular, current approaches are vulnerable to different kinds of biases inherent in the graph data. First, if the class distribution is imbalanced, then the GNNs' loss function is biased towards classifying the majority class correctly rather than the minority class, which hurts the performance of the latter class. Second, due to homophily effect, the learned representation and subsequent downstream tasks may favor certain demographic groups over others when applied to social network data. To mitigate such biases, we propose a novel framework called Fairness-Aware Cost Sensitive Graph Convolutional Network (FACS-GCN) for classifying nodes in networks with skewed class distributions. Our approach combines a cost-sensitive exponential loss with an adversarial learning component to alleviate the ill-effects of both biases. The framework employs a stagewise additive modeling approach to ensure there is no significant loss in accuracy when imparting fairness into the GNN. Experimental results on 6 benchmark graph data demonstrate the effectiveness of FACS-GCN against comparable baseline methods in terms of promoting fairness while maintaining a high model accuracy on the majority of the datasets. 

   more » « less