Graph Neural Networks (GNNs) achieve strong performance on node classification tasks, but their effectiveness often depends on the quality of the supervision, and real-world labels are often noisy. Learning curves—which describe how test performance scales with the number of labelled training nodes—have been extensively studied in classical machine learning, but their behaviour under realistic annotation noise in GNNs remains poorly explored.

We present a systematic empirical study of how three label noise protocols—symmetric random flipping, feature-dependent asymmetric flipping, and structure-dependent flipping—affect the learning curve shape of ChebNet across four benchmark graphs spanning homophilic and heterophilic structure, at noise rates η ∈ {0.1, 0.3, 0.5}.

The central finding is that noise does not simply shift the learning curve downward: above a moderate noise rate it reduces the effective slope, so the gap between clean and noisy performance widens as the label budget grows. Feature-dependent asymmetric noise is consistently the most harmful protocol across all datasets and budgets for η ≥ 0.3, while structure-dependent noise is the least harmful on homophilic graphs. On graphs where the model already operates near its performance limit, noise type has little practical effect.

These findings suggest that beyond a moderate noise rate, cleaning existing labels yields greater returns than acquiring more noisy ones, and that the nature of annotation error interacts with graph structure in ways that single-budget evaluations cannot detect. ... Read more

Learning curves describe how model performance changes as more labeled data becomes available and can help estimate whether collecting additional labels is worthwhile. However, it remains unclear which mathematical functions best represent and extrapolate learning curves for graph neural networks. This study compares power-law and exponential models for learning curves generated by a graph neural network on node-classification datasets with different graph characteristics. The models are evaluated separately on how well they describe observed performance and how accurately they predict performance at larger, unseen labeling budgets. The results show that neither model family is universally preferable. Exponential models provide better descriptive fit on some datasets, while power-law models provide better descriptive fit on others. In the extrapolation experiments, power-law models often give more accurate predictions at larger labeled-node budgets, although the preferred model still depends on the dataset and fitting range. These findings indicate that descriptive fit and extrapolation accuracy should be treated as separate objectives. Overall, power-law behaviour appears to be a useful modelling assumption for some GNN learning curves, especially for extrapolation, but it should not be assumed to hold universally.
... Read more

As renewable energy penetration grows across Europe, transmission networks face increasingly frequent and unpredictable congestion, a problem costing an estimated 4.2 billion euros per year in Europe alone. Yet most data-driven forecasting tools either focus on generation or demand in isolation, or require access to proprietary grid models unavailable to market participants. This thesis addresses the question of whether interzonal power flows, and by extension congestion, can be forecast purely from publicly available market and weather data.

The proposed approach is a spatio-temporal graph neural network that operates directly on transmission edges rather than nodes. It combines an LSTM encoder for temporal dynamics, a Transformer-based graph message-passing module with updatable edge representations, and a future-aware decoder that ingests day-ahead prices and weather forecasts. The model is trained and evaluated on Italy's seven-zone electricity market over a 2025 test year.

Against baselines ranging from naive persistence to gradient-boosted trees and LSTM, the model achieves the best normalized absolute error, directional accuracy, and congestion detection F1 score, with advantages that persist across all six forecast horizons. Critically, the proposed model achieves the best AUROC, demonstrating its ability to rank truly congested hours above non-congested ones regardless of where the threshold defining congestion is placed.

Through comprehensive experiments and analyses, the model proves to be more accurate than standard industry methods and domain-specific models. Further experiments demonstrate the quality of the forecast per edge and horizon, and break down the contribution of the different choices regarding its design. Moreover, the impact of future features is assessed, showing a significant performance increase in congestion detection.

The central contribution to the renewable energy field is a reproducible, open-data framework that transforms observable market and weather signals into physically grounded flow forecasts, without access to network topology or TSO-proprietary models. ... Read more

Recent text-to-image (T2I) diffusion models can generate highly realistic images, but they often struggle to correctly arrange multiple objects according to specified spatial relationships. This limitation reduces their usefulness as controllable design tools. The problem is particularly challenging for modern multi-modal diffusion transformers (MM-DiTs), such as Stable Diffusion 3.5 and FLUX, whose architecture prevents the direct application of earlier layout-control techniques. Existing solutions either require costly model retraining or use training-free methods that provide limited and often unreliable control. This thesis introduces FOCAL, a training-free layout controller that formulates spatial guidance as a stochastic optimal control problem during diffusion sampling. By applying a closed-form correction derived from the model’s attention maps, FOCAL simultaneously enforces object placement and attention separation without modifying model weights. The method improves compositional accuracy across multiple backbones and achieves performance competitive with much larger state-of-the-art systems. ... Read more

Node classification on text-attributed graphs requires both structural reasoning and rich semantic understanding.
While Graph Neural Networks (GNNs) have become the dominant solution by leveraging graph topology, they often rely on limited textual representations.
Recent advances, therefore, explore integrating large language models (LLMs) to provide stronger semantic encoding for graph learning. However, existing work often evaluates different LLM integration paradigms under individually designed experimental settings, making it difficult to assess their relative strengths for node classification tasks. In this work, we present a controlled empirical comparison between classical message-passing graph neural networks, parameter-efficient LLM-GNN integration (ENGINE), prompt-based generative reasoning (LLaGA), and a lightweight hybrid model that combines structural message passing with label-aware semantic alignment. We evaluate all models on two widely used textual graph benchmarks, \textsc{Cora} and \textsc{WikiCS}, under a unified transductive evaluation protocol and varying levels of training supervision. Our results show that the performance of ENGINE consistently outperforms strong GNN baselines, whereas LLaGA is more sensitive to inference constraints and evaluation protocols. The additional hybrid model we propose in this thesis further demonstrates complementary benefits, particularly in low-supervision regimes. These findings clarify practical trade-offs between discriminative and generative LLM-based graph models and highlight hybrid designs as a promising direction for efficient and robust node classification on textual graphs. ... Read more

Flooding is one of the most frequent and destructive natural hazards, accounting for significant human and economic losses every year. Flood hazard mapping allows to identify vulnerable areas by estimating water depth, extent, and intensity under specific scenarios. These maps are created via numerical models as they provide accurate flood simulations. However, they are computationally expensive, particularly for overland flow at high resolution. Data-driven methods based on neural networks offer a promising alternative, delivering faster predictions while maintaining high accuracy.

To provide a comprehensive perspective on deep learning for flood mapping, we first reviewed the state of the art across different applications, considering a range of flood types, spatial scales, and deep learning architectures. Our analysis shows that deep learning methods generally outperform both traditional numerical approaches and conventional machine learning in terms of speed and accuracy. However, most existing models are tailored to individual case studies, neglect the dynamic evolution of flood waves, and cannot transfer to new topographic settings and boundary conditions not seen during training. Furthermore, current approaches struggle to incorporate physical principles, represent hydraulic structures, and provide physically consistent methods for validating outputs, particularly in the context of uncertainty quantification. Finally, we highlight that dike-breach floods remain largely under-represented in the literature, despite their high uncertainty stemming from flood defence failures.

In this thesis, we introduce Graph Neural Networks (GNNs) as hydraulics-inspired Surrogate models for simulating the spatio-temporal evolution of floods. While applicable to different flood types, our focus is on dike-breach floods due to their high uncertainty and particular relevance in the Netherlands and other low-lying areas. We build GNNs that are conceptually analogous to finite-volume methods used to solve the shallow water equations: finite-volume cells are treated as graph nodes, and flux exchanges are learned between adjacent cells by the model. The flood propagation in the proposed SWE-GNN model resembles hydraulics principles and enforces water to propagate only from cells with water. The model works in the same fashion as numerical solvers, auto-regressively predicting the evolution of the hydraulic states over time. By stacking multiple GNN layers, the model captures wider spatial dependencies without requiring small numerical time steps, theoretically needed for stability conditions. We also develop a multi-step ahead loss function combined with curriculum learning that further stabilizes long-term predictions.

We propose a multi-scale GNN formulation that models flood dynamics across different spatial resolutions, enabling the capture of both local and large-scale propagation processes. Time-varying boundary conditions are incorporated through ghost cells, removing the need for separate numerical solvers to initialize simulations. To enhance generalization across unseen unstructured meshes and reduce training data demands, we enforce invariance principles, ensuring the model is independent of coordinate rotations. This multi-scale approach proves both faster and more accurate than its single-scale counterpart. Our methods are validated on a suite of two-dimensional synthetic dike-breach simulations generated with a high-fidelity numerical solver. These datasets progressively increase in complexity by varying initial conditions, location of boundary conditions, size of the domain, computational mesh, and time-varying hydrographs used as boundary condition. Results demonstrate that the GNN generalizes well to unseen topographies, boundary configurations, and mesh configurations, without relying on inputs from numerical simulations. The models achieve a testing critical success index consistently higher than 70% in all datasets. The model also shows generalization to a real case study, dike ring 15 in the Netherlands, with only one fine-tuning simulation.

Finally, we extend the model to explicitly include hydraulic structures and to quantify uncertainty in flood hazard mapping of another real-world case study, dike ring 41 in the Netherlands. The framework is tested for large-scale uncertainty analysis with 10,000 scenarios. All simulations are completed in under 10 hours on a single GPU, corresponding to a speed-up of approximately 10,000 times with respect to the numerical solver, with over half of the scenarios maintaining plausible mass conservation. The combined scenarios are then used to produce probabilistic flood hazard maps, which assume equal likelihood of occurrence for each event. We also analyse the flood uncertainty for a given breach location and return period, showing that the model ensemble can provide better flooding estimates than the deterministic scenario.

This thesis highlights the potential of GNN-based surrogates for time-sensitive flood risk assessments under uncertainty. By demonstrating both accuracy and computational efficiency, it contributes to bridging the gap between complex hydraulic modelling and large-scale flood risk analysis. Despite being applied on dike-breach floods, the framework introduced in this thesis can readily be applied to fluvial and coastal floods without modifications. This open pathways for integrating surrogates into operational decision making and future flood resilience planning.
... Read more

Multivariate time series modeling requires capturing complex dependencies both within individual time series and across different variables. Existing graph-based approaches are limited to pairwise interactions, while recent product cell complex methods assume homogeneous higher-order relationships. This thesis proposes the Simplicial Product Complex (SPC), a topological framework that constructs simplicial complexes in the product space to capture heterogeneous higher-order interactions across space and time. The key innovation is the ability to distinguish between different relationship types and learn their relative importance from data. We develop the Simplicial Product Complex Convolutional Neural Network (SPCCNN) to perform data-adaptive learning over these structures. Experimental evaluation shows SPCCNN achieves competitive performance with state-of-the-art methods while offering enhanced flexibility through parameterized structures. The model adapts to dataset-specific patterns and maintains computational efficiency through sparsity regularization. Our findings demonstrate the effectiveness of higher-order simplicial modeling for capturing complex temporal dynamics in multivariate time series data. ... Read more

Accurate residential property valuation is essential for mortgage lending, taxation, and urban planning, yet remains challenging due to complex spatial and temporal dynamics. Traditional econometric models are interpretable but rely on restrictive assumptions, while machine learning–based automated valuation models (AVMs) improve predictive accuracy but treat transactions as independent, overlooking spatial spillovers and evolving trends. Recent graph-based approaches partially address spatial autocorrelation, but often rely on static or unidirectional structures that limit their expressiveness.

We introduce a Multi-Scale Bidirectional Spatio-Temporal Graph Neural Network (MBSTGNN) that models transactions and neighbourhoods as dynamic graphs linked through bidirectional message passing. A temporal memory mechanism maintains consistency across time, enabling the model to capture evolving market conditions. Evaluated on Rotterdam housing transactions, MBSTGNN outperforms strong baselines, particularly in sparse-data settings, and produces embeddings that reveal domain-consistent socio-spatial and temporal patterns. These results demonstrate its potential for advancing automated valuation and related spatio-temporal prediction tasks. ... Read more

Estimating bike trip times is becoming more and more important in many different areas such as urban mobility and route planning. However, especially in real-world, the GPS data used to generate these estimations is frequently noisy, irregularly sampled, or incomplete. With an emphasis on how these strategies interact with trip length and speed variance, this study intends to examine the effects of various data resampling techniques on the precision of bicycle travel time estimations. To analyze the impact of different preprocessing methods, we apply and assess a graph neural network model using various resampling techniques. Instinctively, the assumption that we expect to be concluded from this research is that there is no single resampling technique works well for every kind of trip. Rather, trip parameters like duration and speed fluctuation have a significant impact on accuracy. ... Read more

Accuracy‐driven recommender systems risk confining users to "filter‐bubbles'' of familiar content. Recent work on coVariance Neural Networks (VNNs) provides a scalable alternative to Principal Component Analysis (PCA) for modelling high-order correlations, but their impact on beyond-accuracy metrics (BAMs), such as Novelty and Diversity, remains unexplored.
We use the user–user covariance (or its inverse, the precision matrix) as a graph shift operator (GSO) and train SelectionGNN-based VNNs on the MovieLens-100K dataset.
Two training regimes are evaluated: (i) RMSE-only (No-BAM-SVNN) and (ii) a compound loss that also includes novelty and diversity terms (BAM-SVNN).
For each regime we sweep six graph configurations: covariance/precision crossed with {dense, hard-threshold, soft-threshold} sparsification, under five random seeds, yielding 30 runs per regime.
Baseline comparisons include PCA, a naive mean–std model, and a random predictor.

The best SVNN configuration increases recommendation Novelty by 2.8 percentage points and matches PCA’s Diversity while incurring only a 0.03 RMSE penalty.
Hard-thresholded precision graphs provide the lowest SVNN RMSE (0.952), whereas dense covariance graphs maximise diversity (0.868).
Integrating novelty/diversity directly into the loss offers no additional benefit yet multiplies runtime by x33.
One-way ANOVA indicates that model family explains 97.6% of RMSE variance (\(\eta^2=0.976\)) and 77.8% of novelty variance.

This work is the first to benchmark (sparsified) VNNs on beyond-accuracy metrics, demonstrating a favourable accuracy–novelty trade-off and clarifying when sparsification and BAM-weighted training pay off.
All code, data splits and statistical notebooks are released for full reproducibility. ... Read more

This research investigates the effectiveness of graph-based data augmentation techniques in improving the performance of DG4b, a deep learning model designed to estimate bicycle travel times in urban environments. Given the limitations of real-world cycling datasets, particularly data scarcity and trip-length imbalance, we propose two augmentation methods: Graph Stitching (GS), which combines segments of existing trips to form new trajectories, and Graphon-Inspired Trip Generation (GITG), which uses an empirically estimated transition kernel to simulate realistic trip patterns through probabilistic sampling.
Despite limited improvements, this study establishes a foundation for future research in graph-based trajectory augmentation. Integrating richer trip-level features, such as dynamic environmental conditions or behavioral data, with structural augmentation could lead to more effective training data and improved model generalization. ... Read more

Aligning diffusion model outputs with downstream objectives is essential for improving task-specific performance. Broadly, inference-time training-free approaches for aligning diffusion models can be categorized into two main strategies: sampling-based methods, which explore multiple candidate outputs and select those with higher reward signals, and gradient-guided methods, which use differentiable reward approximations to directly steer the generation process. In this work, we propose a universal algorithm, CoDeX, which brings together the strengths of blockwise sampling and gradient-based guidance into a unified framework. Building on the blockwise sampling paradigm of CoDe, CoDeX integrates local gradient signals during sampling, thereby addressing the sampling inefficiency inherent in complex reward-based sampling approaches like CoDe. At the same time, it overcomes the limited applicability of traditional gradient-guided methods, which often struggle with non-differentiable rewards. By cohesively combining these two paradigms, CoDeX enables more efficient sampling while offering better trade-offs between reward alignment and divergence from the diffusion unconditional prior. Empirical results demonstrate that CoDeX consistently outperforms CoDe and remains competitive with state-of-the-art baselines across a range of tasks. ... Read more

Recommender systems help users navigate vast catalogs of content through recommendations, of which rating prediction remains an important task. Traditional methods such as collaborative filtering often struggle to model higher-order relationships between users and items, as well as suffer from the cold start problem when the number of users and items is still low. Graph Neural Networks (GNNs) have shown promise in this area, although they are often limited by their focus on local graph structures. This study explores the application of Covariance Neural Networks (VNNs) for rating prediction, leveraging covariance matrices to leverage global statistical dependencies and model higher-order relationships. Using the MovieLens-100k dataset, we evaluate the performance of VNNs against baselines and other models, using RMSE as the metric of evaluation. Our results demonstrate that VNNs outperform simple matrix completion techniques, but are limited by their susceptibility to oversmoothing. This work highlights the potential of VNNs for recommender systems while underscoring the need for careful architectural design to balance performance and stability. ... Read more

Accurate prediction of bicycle travel time is critical for efficient urban mobility and sustainable transport planning. However, real-world datasets are noisy, imbalanced and lack rich contextual features. This limits the effectiveness of current graph-based neural network models. This research aims to explore how feature engineering and model enhancements can improve the performance of a Graph Convolutional Neural Network (GCNN) in the context of travel time prediction. Building on a currently existing DG4b architecture, the input data is enriched with temporal, spatial and traffic-related features. Architectural enhancements are integrated by employing techniques such as graph data augmentation, and multi-scale graph learning. Using a dataset from Berlin, the improvements are evaluated primarily in terms of prediction accuracy across varying trip lengths, which implicitly reflect speed variability and route diversity. The goal is to explore how targeted feature engineering and graph-based modeling techniques influence the accuracy of bicycle travel time estimation, especially across different trip durations that reflect real-world cycling variability. The results show that optimal feature engineering improved the model up to 6% and a combination of the model enhancement techniques improved the model up to 2%. ... Read more

Graph Neural Networks (GNNs) are an effective architecture for implementing collaborative filtering-based recommender systems. This paper evaluates the performance and computational complexity of precision matrix-based VNNs as a collaborative filter on the MovieLens-100K dataset. Results show the estimated precision matrix contains a high amount of noise when calculated from sparse data, which impacts the performance of the model. After sparsifying the precision matrix, the performance and computational complexity improved significantly. ... Read more

This research investigates the application of Graph Neural Networks (GNNs) for rating prediction in recommender systems, utilizing precision matrices as graph filters. The focus is on movie recommendation, where graph-based structures are especially relevant due to the importance of user and item relationships. A leave-one-out masking strategy is employed during training to ensure the model learns from all available training data. The proposed model achieves a test root mean squared error (RMSE) of approximately 0.95 on the MovieLens-100k dataset, performing reasonably well compared to existing graph-based and matrix factorization methods. While the model accurately predicts average ratings, it tends to overestimate lower ratings. These results demonstrate the potential of precision-based graph filters in GNNs but also reveal significant room for improvement before reaching state-of-the-art performance. Future work may include output calibration and sparsification of the precision matrix to enhance both efficiency and predictive accuracy. ... Read more

Multivariate time series arise in a wide range of domains, such as weather forecasting and financial modeling, where multiple interdependent variables evolve simultaneously over time. For instance, temperature readings at one location may have a delayed influence on nearby regions, while currency exchange rates exhibit complex, lagged interactions across global financial markets Effectively modeling these spatiotemporal interactions, particularly in streaming and non-stationary settings, remains a fundamental challenge. Traditional approaches such as Temporal PCA operate on sample covariance matrices but often suffer from stability issues in the estimated eigenvectors, especially in low-data regimes or when the corresponding eigenvalues are close. These covariance estimation errors can propagate into the learned representation and degrade performance in downstream tasks. Recent graph-based learning methods address this limitation by constructing graphs from the sample covariance matrix and learning from its structure. However, these approaches typically consider only lag-zero correlations, thereby limiting their ability to model cross-temporal dependencies and fully capture the spatiotemporal structure inherent in multivariate time series. To overcome this, this thesis proposes the Lagged spatiotemporal coVariance Neural Network (LVNN), a neural network architecture that leverages lagged covariance information to learn representations from multivariate time series in a streaming setting. LVNN constructs a spatiotemporal graph by concatenating consecutive temporal samples, computing their extended sample covariance matrix, and using it as a structural prior for graph convolutions. This design enables the model to capture variable interactions not only within time steps but also across temporal lags. However, the use of a larger spatiotemporal covariance matrix introduces additional computational overhead and spurious correlations. To address this, we introduce two structural modifications to our proposed model. First, we retain only spatial and backward temporal connections, corresponding to the block upper triangular part of the extended covariance matrix. Second, we apply thresholding-based sparsification techniques to prune weak correlations and improve scalability. We begin by proving that LVNN is robust to perturbations in the online estimation of the extended covariance matrix in stationary settings, improving over the stability issues of temporal PCA-based methods. These findings are empirically validated on synthetic stationary datasets. Then, to assess the effectiveness of the learned embeddings, we evaluate LVNN on single-step forecasting tasks using three real-world datasets across different forecasting horizons. The standard LVNN model performs comparably to our baselines, while the LVNN variant with the block upper triangular matrix demonstrates the most consistent performance. Furthermore, applying hard and soft thresholding sparsification techniques to the extended covariance matrix substantially reduces the computational overhead, with only a minor impact on forecasting performance. These results support our hypothesis that cross-temporal covariance terms are a valuable source of inductive bias for representation learning in multivariate time series. ... Read more

Extending the concepts of classical signal processing to graphs, a wide array of methods have come to the fore, including filtering, reconstruction, classification, and sampling. Existing approaches in graph signal processing consider a known and static topology, i.e., fixed number of nodes and a fixed edge support. Two types of tasks stand out, namely, topology inference, where the edge support along with their weights are estimated from signals; and data processing, where existing data and the known topology are used to perform different tasks. However, such tasks become quite challenging when the network size and support changes over time. Particularly, these challenges involve adapting to the changing topology, data distributions and dealing with unknown topological information. The latter manifests for example, when new nodes are available to attach to the graph but their connectivity is uncertain as is the case in cold start graph-based recommender systems.
The key contribution of this dissertation is proposing methodologies for signal processing over dynamic networks which are aimed at the two aforementioned tasks. For dynamic networks with incoming nodes, we process signals by introducing a parametric stochastic attachment model. In this model, the incoming nodes connect with probability to existing nodes with certain weights. This uncertainty allows us to model input output relations and allows us to cast them in the context of different graph signal processing tasks. We learn the model attachment parameters in a task-aware setting, allowing us to interpret topology identification in task-aware settings. Separately, we also propose filter design strategies for processing signals both at the incoming and existing nodes using stochastic attachment models.
Another contribution of this dissertation is to extend graph signal processing with graph filters to the scenario where the graph keeps growing in size with streaming data. We propose online graph filter design which updates the filter online, based on incoming nodes. We design this both for scenarios where the incoming node connectivity is known and unknown. In the unknown connectivity case, we study the performance difference between knowing and not knowing the topology and how the stochastic attachment influences it. We also show that by adapting the stochastic attachment, we can learn faster from the data stream.
Finally,we consider the task of topology decomposition and identification for dynamic networks with fixed nodes but changing edge support. We build a tensor of partially observed adjacency matrices corresponding to such a dynamic topology and express this in terms of underlying latent graphs and their temporal signatures. Furthermore, we account for the time-varying graph signals as a prior to aid identifying these latent graphs and missing components of the topology. These latent graphs are individually and collectively expressive and provide interpretable decompositions along with outperforming traditional structure agnostic low-rank decompositions. ... Read more

Machine learning has been growing beyond data living on Euclidean spaces (e.g., texts, audios, images). Graph machine learning models, e.g., graph neural networks (GNNs), succeed in learning from graph-structured data using the graph topological information. In this thesis, we focus on a new domain, simplicial complexes. Not only are they a popular higher-order network model that generalizes graph models encoding pairwise relations between nodes, but they also allow us to support signals on various network objects (nodes, edges, and faces). For example, edge flows defined on a simplicial complex can be studied in terms of both divergent and rotational properties, providing a better model for real-world flows like traffic flows, water flows, money flows, etc.
The main theme of the thesis is to develop principled machine learning models for signals on simplicial complexes. By principled, we mean that the models leverage the intrinsic priors of the domain and the signals, namely, the topological structure of simplicial compelxes and the Hodge decomposition of simplicial signals. The latter states that, for example, edge flows can be decomposed into a divergence-free part and a curl-free part, each modeling the distinct properties of real-world flows — the conservation of flows (e.g., water flows) and the rotational properties of flows (e.g., electric currents).... ... Read more

Graph Neural Networks (GNNs) have achieved state-of-the-art performance in various applications due to their ability to capture complex structural relationships within graph data. However, their inherent black-box nature poses significant challenges to model interpretability, particularly in the context of link prediction tasks. This thesis investigates the explainability of GNNs for link prediction, a critical task in domains such as social network analysis and recommendation systems. We propose an efficient extension of the Zorro explainer, originally designed for node classification, to address the unique challenges of explaining link predictions. Our method leverages top similar nodes for initialization and incorporates techniques to accelerate the greedy search process, ultimately producing binary node and feature masks as explanations. Additionally, we introduce a novel Edge-level RDT-Fidelity metric to better evaluate explainers that generate edge masks, complementing existing metrics like RDT-Fidelity and Sparsity. Experimental results on the Cora and PubMed datasets demonstrate the superior performance and efficiency of our Zorro extension. This work contributes to the growing field of GNN interpretability by providing new methodologies and evaluation frameworks specifically tailored for link prediction. ... Read more