Search for: All records

Graph Neural Networks (GNNs) have become a building block in graph data processing, with wide applications in critical domains. The growing needs to deploy GNNs in high-stakes applications necessitate explainability for users in the decision-making processes. A popular paradigm for the explainability of GNNs is to identify explainable subgraphs by comparing their labels with the ones of original graphs. This task is challenging due to the substantial distributional shift from the original graphs in the training set to the set of explainable subgraphs, which prevents accurate prediction of labels with the subgraphs. To address it, in this paper, we propose a novel method that generates proxy graphs for explainable subgraphs that are in the distribution of training data. We introduce a parametric method that employs graph generators to produce proxy graphs. A new training objective based on information theory is designed to ensure that proxy graphs not only adhere to the distribution of training data but also preserve explanatory factors. Such generated proxy graphs can be reliably used to approximate the predictions of the labels of explainable subgraphs. Empirical evaluations across various datasets demonstrate our method achieves more accurate explanations for GNNs. 

Explaining deep learning models operating on time series data is crucial in various applications of interest which require interpretable and transparent insights from time series signals. In this work, we investigate this problem from an information theoretic perspective and show that most existing measures of explainability may suffer from trivial solutions and distributional shift issues. To address these issues, we introduce a simple yet practical objective function for time series explainable learning. The design of the objective function builds upon the principle of information bottleneck (IB), and modifies the IB objective function to avoid trivial solutions and distributional shift issues. We further present TimeX++, a novel explanation framework that leverages a parametric network to produce explanation-embedded instances that are both in-distributed and label-preserving. We evaluate TimeX++ on both synthetic and real-world datasets comparing its performance against leading baselines, and validate its practical efficacy through case studies in a real-world environmental application. Quantitative and qualitative evaluations show that TimeX++ outperforms baselines across all datasets, demonstrating a substantial improvement in explanation quality for time series data. 

Graph Neural Networks (GNNs) are neural models that leverage the dependency structure in graphical data via message passing among the graph nodes. GNNs have emerged as pivotal architectures in analyzing graph-structured data, and their expansive application in sensitive domains requires a comprehensive understanding of their decision-making processes — necessitating a framework for GNN explainability. An explanation function for GNNs takes a pre-trained GNN along with a graph as input, to produce a ‘sufficient statistic’ subgraph with respect to the graph label. A main challenge in studying GNN explainability is to provide fidelity measures that evaluate the performance of these explanation functions. This paper studies this foundational challenge, spotlighting the inherent limitations of prevailing fidelity metrics, including Fid+, Fid−, and Fid∆. Specifically, a formal, information-theoretic definition of explainability is introduced and it is shown that existing metrics often fail to align with this definition across various statistical scenarios. The reason is due to potential distribution shifts when subgraphs are removed in computing these fidelity measures. Subsequently, a robust class of fidelity measures are introduced, and it is shown analytically that they are resilient to distribution shift issues and are applicable in a wide range of scenarios. Extensive empirical analysis on both synthetic and real datasets are provided to illustrate that the proposed metrics are more coherent with gold standard metrics. The source code is available at https://trustai4s-lab.github.io/fidelity. 

Abstract The warm-to-cold densification of Atlantic Water (AW) around the perimeter of the Nordic Seas is a critical component of the Atlantic Meridional Overturning Circulation (AMOC). However, it remains unclear how ongoing changes in air-sea heat flux impact this transformation. Here we use observational data, and a one-dimensional mixing model following the flow, to investigate the role of air-sea heat flux on the cooling of AW. We focus on the Norwegian Atlantic Slope Current (NwASC) and Front Current (NwAFC), where the primary transformation of AW occurs. We find that air-sea heat flux accounts almost entirely for the net cooling of AW along the NwAFC, while oceanic lateral heat transfer appears to dominate the temperature change along the NwASC. Such differing impacts of air-sea interaction, which explain the contrasting long-term changes in the net cooling along two AW branches since the 1990s, need to be considered when understanding the AMOC variability. 

The Mid‐Atlantic Bight (MAB) Cold Pool is a bottom‐trapped, cold (temperature below 10°C) and fresh (practical salinity below 34) water mass that is isolated from the surface by the seasonal thermocline and is located over the midshelf and outer shelf of the MAB. The interannual variability of the Cold Pool with regard to its persistence time, volume, temperature, and seasonal along‐shelf propagation is investigated based on a long‐term (1958–2007) high‐resolution regional model of the northwest Atlantic Ocean. A Cold Pool Index is defined and computed in order to quantify the strength of the Cold Pool on the interannual timescale. Anomalous strong, weak, and normal years are categorized and compared based on the Cold Pool Index. A detailed quantitative study of the volume‐averaged heat budget of the Cold Pool region (CPR) has been examined on the interannual timescale. Results suggest that the initial temperature and abnormal warming/cooling due to advection are the primary drivers in the interannual variability of the near‐bottom CPR temperature anomaly during stratified seasons. The long persistence of temperature anomalies from winter to summer in the CPR also suggests a potential for seasonal predictability.