What Graph Machine Learning got wrong, from 63 theses

Tried and failed

GNNs on static population graphs applied to tabular and imaging phenotype regression. Outcome: worse than baseline. Reason: Static heuristic graph construction introduces suboptimal connectivity compared to standard MLP baselines

Deep learning for interpretable brain age estimation · Imperial · PhD

Tried and failed

combining language models with graph structure embeddings applied to link prediction on sparse graphs. Outcome: worse than baseline. Reason: graph structure information introduced noise on sparse graphs, degrading prediction performance

Temporal link prediction in the wild · Imperial · PhD

Tried and failed

Graph convolutional networks with synthetic node features applied to non-attributed graph community detection. Outcome: worse than baseline. Reason: synthetic features like landing probabilities or random vectors fail to provide meaningful signals for GCN aggregation

On Statistical Learning for Structural Data: Data Fusion and Semi-supervised Learning · Harvard · PhD

Tried and failed

graph convolutional networks with neighborhood feature aggregation applied to influencer detection in social networks. Outcome: worse than baseline. Reason: aggregating neighborhood node features introduced noise and degraded predictive performance

Analyzing Networks with Hypergraphs: Detection, Classification, and Prediction · Virginia Tech · Doctorate

Considered and rejected

Considered and rejected: Rejected container-level node modeling in GNN graphs because generalization suffered compared to host-level node modeling with isolated application signatures.

Self-adaptive serverless edge computing for edge intelligence · TU Wien · PhD

Considered and rejected

Considered and rejected: Using full dependency graphs without pruning to entity-mention paths hurt representation learning in GraphTransfer.

Low-Resource Event Extraction · University of Oregon · PhD

Considered and rejected

Considered and rejected: Normalizing graph adjacency by the whole-matrix sum (P = A / sum(A)) instead of per-node row-sum normalization, rejected because it pushed low-degree nodes to the periphery and degraded recall.

Visualization beyond Riemannian Manifolds · Tübingen · PhD

Tried and failed

expanding knowledge graph embeddings to multi-hop neighborhoods applied to multimodal cross-attention feature enrichment. Outcome: worse than baseline. Reason: expanded neighborhood graphs introduced irrelevant concept noise that degraded cross-modal alignment

Commonsense for Zero-Shot Natural Language Video Localization · Virginia Tech · Master's

Tried and failed

random walk graph embedding on full hierarchy applied to unfiltered large-scale medical ontology. Outcome: worse than baseline. Reason: generating embeddings without domain-presence filtering introduced noise from irrelevant or overly granular concepts

Artificial Intelligence-Driven Clinical Decision Support for Antibiotic Optimisation · Imperial · PhD

Tried and failed

inserting multiple knowledge graph triplets into text applied to entity linking language models. Outcome: worse than baseline. Reason: inserting more candidate triplets introduced excessive noise compared to using a single triplet

Leveraging Structured Knowledge to Enable Efficient AI Assistants · Georgia Tech · PhD

Tried and failed

knowledge graph augmented retrieval generation applied to document question answering. Outcome: worse than baseline. Reason: vector search retrieved incorrect documents, compounding retrieval errors downstream

Loud Yet Invisible: A Humanist-Designed Pipeline for Unlocking the Early Modern Archive · Saskatchewan · Master's

Tried and failed

random graph connectivity in graph neural networks applied to named entity recognition. Outcome: worse than baseline. Reason: destroying structural layout priors removes informative relational context

From Structured Document To Structured Knowledge · MIT · PhD

Tried and failed

high correlation threshold for graph edge construction applied to node embedding correlation graph reconstruction. Outcome: worse than baseline. Reason: overly restrictive threshold caused excessive false negatives by omitting true dependencies

Learning and Reconstructing Conflicts in O-RAN · Virginia Tech · Master's

Tried and failed

k-nearest neighbor graph construction from features applied to adversarial graph purification and learning. Outcome: worse than baseline. Reason: Feature-only kNN connections discard spectral and topological structural information needed for accurate graph recovery

Accurate and Efficient Representation Learning on Large-Scale Graphs · Cornell · PhD

Tried and failed

graph convolutional network without pretraining applied to molecular property prediction. Outcome: worse than baseline. Reason: failed to exceed sequence-based models and barely matched simple baselines

Photoionization Detection of Volatile Organic Compounds · Harvard · PhD

Tried and failed

graph neural network on atomic coordinates applied to predicting global scalar material properties. Outcome: worse than baseline. Reason: provided only marginal accuracy improvement over random forest baselines despite much higher complexity

Using Deep Learning to Understand and Design Heterogeneous Materials · MIT · PhD

Tried and failed

graph neural networks with language model embeddings applied to chemical reaction yield prediction. Outcome: worse than baseline. Reason: failed to outperform random forests trained on molecular fingerprints and physical descriptors

Machine Learning for Chemical Reactivity Prediction: Paradigms, Challenges, and Applications · MIT · PhD

Tried and failed

increasing graph neural network architecture complexity applied to predicting electronic structure properties of crystals. Outcome: worse than baseline. Reason: advanced architectures provided only marginal accuracy gains while sacrificing physical interpretability

Computational and Data-Driven Design of Perturbed Metal Sites for Catalytic Transformations · Virginia Tech · Doctorate

Tried and failed

learned molecular graph embeddings applied to small molecule representation for drug-target interaction. Outcome: worse than baseline. Reason: learned embeddings failed to outperform simple fingerprint baselines

Learning the language of biomolecular interactions · MIT · PhD

Lost to a baseline

On CoauthorCS, simple Dense feed-forward baseline without graph structure (94.88±0.21%) outperformed GCN (93.85±0.23%), GAT (93.80±0.38%), and L-CAT (93.65±0.23%)

Mitigating Practical Limitations of GNNs through Explicit Inductive Biases · Tübingen · PhD

Lost to a baseline

Baseline graph-agnostic MLP (MAE 3.73 years) beat static population graph GCNs (MAE 3.89 to 5.19 years) and GATs (MAE 4.07 to 5.38 years).

Deep learning for interpretable brain age estimation · Imperial · PhD

Considered and rejected

Considered and rejected: Rejected Graph Neural Networks for node-level epidemic status inference in Chapter 1 in favor of standard MLPs for simplicity and generalizability to non-network models.

Approximate Bayesian Inference for Network Processes · Harvard · PhD

Considered and rejected

Considered and rejected: Rejected using complex non-linear deep learning or graph neural networks (GNNs) requiring structural input, choosing single-parameter-per-element compositional linear/ReLU models to retain human-interpretable chemical heuristics.

Learning Simple Chemical Heuristics to Model and Discover Materials · MIT · PhD

Considered and rejected

Considered and rejected: Rejected standalone deep graph neural networks (GNNs/CNNs) as primary operational classifiers due to black-box opacity, lack of auditability for financial regulators, and high computational inference costs.

AI-Driven Financial Market Surveillance: Detecting and Preventing Misconduct in Decentralized Systems · Politecnico di Torino · PhD

Lost to a baseline

Feature-based graph embedding methods MUSAE (AUC 0.5261 w/o POI, 0.4940 w/ POI), TENE (AUC 0.4928 w/o POI, 0.5316 w/ POI), and ASNE (AUC 0.4904 w/o POI, 0.5203 w/ POI) were outperformed by the raw feature baseline (AUC 0.5603 w/o POI, 0.7984 w/ POI).

Individual Human Mobility Models for sustainable cities applications · Scuola Normale Superiore · PhD

Tried and failed

graph neural networks as surrogate models applied to optimal power flow proxies. Outcome: worse than baseline. Reason: numerical instability and lower prediction accuracy compared to standard deep neural networks

Synergizing Machine Learning and Optimization: Scalable Real-Time Risk Assessment in Power Systems · Georgia Tech · PhD

Tried and failed

augmenting graph neural networks with handcrafted molecular descriptors applied to molecular property prediction from semi-empirical calculations. Outcome: worse than baseline. Reason: hyperparameter tuning and manual descriptor concatenation yielded minimal performance gains or degraded predictive accuracy

High-throughput virtual screening of molecules for photon conversion · Imperial · Master's

Tried and failed

heuristic subgraph decomposition algorithms applied to molecular property prediction with graph networks. Outcome: worse than baseline. Reason: heuristic partitioning rules captured less meaningful substructures than standard greedy decomposition across benchmarks

Motif-based Graph Neural Networks: Representation, interpretation, and extraction · Iowa State · PhD

Tried and failed

attentive fingerprint graph neural network applied to molecular property and activity prediction. Outcome: too slow. Reason: increased computational time without consistent performance gains over graph attention networks

Exploring Graph Neural Networks for Molecular Activity Prediction · Harvard · Master's

Tried and failed

graph attention networks applied to molecular property prediction. Outcome: overfit. Reason: attention mechanisms showed large train-validation performance gaps compared to standard graph convolutions

Exploring Graph Neural Networks for Molecular Activity Prediction · Harvard · Master's

Tried and failed

replacing graph neural network with multi-layer perceptron applied to 3D part grouping feature encoding. Outcome: worse than baseline. Reason: loss of pairwise relational information between components

Novel learning-based techniques and benchmarks for 3D part grouping · Imperial · PhD

Tried and failed

pruning peripheral nodes to isolate local features applied to graph neural network property prediction. Outcome: worse than baseline. Reason: removing context nodes degrades representations that rely on broader structural context

Multiscale Models Of Interfacial Mechanics In Low Dimensional Systems · Penn · PhD

Considered and rejected

Considered and rejected: Rejected Message Passing Neural Network (MPNN) architecture in favor of Graph Attention Network because it gave similar performance (53.5% vs 53.0%) but MPNN was less interpretable and more complex.

Evaluating privacy and robustness in modern data processing systems · Imperial · PhD

Considered and rejected

Considered and rejected: Rejected nonlinear activations and dense weight aggregation layers in graph convolution, as linear feature propagation/smoothing was proven more effective and less complex

Advances in Multi-Stage Recommender Systems for Large Scale Applications · University College Dublin · PhD

Considered and rejected

Considered and rejected: Complex GNN architectures (GAT, GraphSAGE) rejected in favor of simple 2-layer GCN because graph sparsity rendered complex message-passing unnecessary and computationally heavier.

Towards Network-Guided Large-Scale Foundation Models on Single-Cell Transcriptomics · Virginia Tech · Master's

Tried and failed

sequential layers after graph neural network applied to source code function classification. Outcome: worse than baseline. Reason: CNN and LSTM layers after GG-NN performed worse than standard global pooling layers

Recommending TEE-based Functions Using a Deep Learning Model · Virginia Tech · Master's

Tried and failed

random landmark node selection for embedding applied to graph representation learning. Outcome: worse than baseline. Reason: random selection captured less structural importance than degree-based landmark selection across datasets

Graph Embedding for Retrieval · EPFL · PhD

Tried and failed

min pooling across edge-embedded graph dimensions applied to text graph neural networks. Outcome: worse than baseline. Reason: failed to retain discriminative features compared to max or average pooling

Improving Text Classification Using Graph-based Methods · Virginia Tech · Doctorate

Tried and failed

higher-expressivity substructure features in graph neural networks applied to molecular property prediction. Outcome: did not generalise. Reason: substructures providing higher vertex disambiguation (paths/trees) did not capture domain-relevant inductive biases like cycles

Neural function approximation on graphs: shape modelling, graph discrimination & compression · Imperial · PhD

Tried and failed

broadly trained graph neural network property prediction applied to ternary crystal magnetic property screening. Outcome: did not generalise. Reason: generalized crystal graph model exhibited systematic errors on multi-component compositions without domain-specific fine-tuning

The discovery and design of rare-earth-free magnets using machine learning · UT Austin · PhD

Tried and failed

discrete generative diffusion on graph adjacency matrices applied to knowledge graph link correction. Outcome: overfit. Reason: severe test set overfitting and failure to generalize beyond training graph structures

Scalable Methods for Knowledge Graph Reasoning and Generation · EPFL · PhD

Tried and failed

Magnetic Laplacian on continuous directed graphs applied to fuzzy directed graph representation learning. Outcome: did not generalise. Reason: Fails to separate in- and out-neighbor messages when generalized to continuous edge directions, reducing expressiveness

Developing Differentiable Toolkits for Computational Biology · Harvard · PhD

Tried and failed

heat diffusion priors on graph time series applied to spatial network time series forecasting. Outcome: did not generalise. Reason: Strict volume conservation in diffusion models degrades long-term prediction accuracy.

Multivariate time series forecasting for freeway networks · EPFL · PhD

Tried and failed

weight decay and dropout regularization applied to GNNs on heterophilic directed graphs. Outcome: did not generalise. Reason: regularization techniques failed to improve performance on heterophilic graph data

Deep learning on real-world graphs · Imperial · PhD

Tried and failed

label and residual propagation on graphs applied to heterophilic and negatively correlated networks. Outcome: worse than baseline. Reason: standard smoothing assumes positive homophily, failing when connected nodes have dissimilar labels

Modeling and Inferring Attributed Graphs · Cornell · PhD

Tried and failed

feature propagation for missing node attributes applied to heterophilous or low-homophily graphs. Outcome: worse than baseline. Reason: smoothing features across connected nodes fails when neighbors do not share similar attributes or classes

Deep learning on real-world graphs · Imperial · PhD

Considered and rejected

Considered and rejected: Rejected Graph Attention Networks (GAT) because they operate in a transductive setting requiring the entire graph structure during training, hindering generalization to unseen expansion stations.

Data-Driven Bike-Share Ridership Prediction and Network Optimization · York University · PhD

Tried and failed

learned optimizers parameterized by graph recurrent networks applied to distributed optimization. Outcome: did not generalise. Reason: performance degraded or saturated when evaluated on time horizons longer than seen during training

Control and Optimization over Large-Scale Networks · Penn · PhD

Tried and failed

relational graph neural networks with static fully-connected topologies applied to distributed multi-sensor time-series dynamics. Outcome: did not generalise. Reason: fixed fully connected graphs fail to reflect realistic sparse and dynamic interactions

Data analytics for real-time event identification using real phasor measurement unit data · Iowa State · PhD

Tried and failed

zero-shot policy transfer across network topologies applied to optimization algorithm parameter tuning. Outcome: did not generalise. Reason: policies trained on one graph structure fail on differently sized or structured graphs without retraining

Designing policy optimization algorithms for multi-agent reinforcement learning · Georgia Tech · PhD

Tried and failed

deep message passing in graph neural networks applied to molecular property prediction. Outcome: did not generalise. Reason: Excessive message passing steps caused feature collapse and over-smoothing of node embeddings.

Advances in Artificial Intelligence for Accelerated Discovery of Energy Storage Polymers · Georgia Tech · PhD

Tried and failed

graph attention networks for relational modeling applied to inter-agent interaction modeling. Outcome: worse than baseline. Reason: Suffered from oversmoothing and failed to outperform a simple multilayer perceptron.

Unified and Multimodal Learning for Gaze Prediction in Naturalistic Settings · EPFL · PhD

Tried and failed

deepening spatio-temporal graph convolutional networks applied to motion and action prediction. Outcome: worse than baseline. Reason: increasing graph convolution depth caused over-smoothing of node representations

On the motion and action prediction using deep graph models · UT Austin · PhD

Tried and failed

stacking deeper continuous graph ODE blocks applied to spatio-temporal traffic forecasting. Outcome: overfit. Reason: increasing neural ODE-GNN depth from 2 to 6 layers caused overfitting relative to a single block

Graph-based Multi-ODE Neural Networks for Spatio-Temporal Traffic Forecasting · Virginia Tech · Master's

Tried and failed

increasing graph convolutional neural network layer depth applied to multi-object tracking feature association. Outcome: worse than baseline. Reason: deeper graph convolutions cause feature over-smoothing and degrade node discriminability

A Graph Convolutional Neural Network Based Approach for Object Tracking Using Augmented Detections With Optical Flow · Virginia Tech · Master's

Tried and failed

increasing graph convolutional network depth applied to small graph path prediction. Reason: over-smoothing and diminishing returns on small 5-node graphs beyond four layers

Space Layout Optimization For Natural Ventilation Using Machine Learning Techniques · Harvard · PhD

Tried and failed

standard graph convolutional networks applied to fully connected graphs in routing problems. Outcome: no signal. Reason: Low-pass filtering caused oversmoothing on complete graphs, producing uniform edge predictions that failed to guide search.

DEEP UNSUPERVISED MODELS LEVERAGING LEARNING AND REASONING · Cornell · PhD

Tried and failed

deep single-stage graph neural network applied to flow simulation on network graphs. Outcome: worse than baseline. Reason: increasing network depth caused severe oversmoothing of node representations

Towards quasi real-time simulations of district heating networks for an optimal sustainable design and control · EPFL · PhD

Considered and rejected

Considered and rejected: Restricting graph convolution depth to two hops to prevent oversmoothing and overfitting.

Graph Neural Networks for Interpretable Biomedical Data Analysis in Genomics and Structural Biology · Penn · PhD

Tried and failed

negative sampling without false negative filtering applied to knowledge graph embedding training. Outcome: worse than baseline. Reason: unfiltered corrupted triples inadvertently included true relations from the training set as negative examples

Using Multi-Relational Embeddings as Knowledge Graph Representations for Robotics Applications · Georgia Tech · PhD

Tried and failed

progressive architecture expansion for continual learning applied to knowledge graph relational embeddings. Outcome: worse than baseline. Reason: lower inference performance than finetuning and corrupted consolidated latent representations

Using Multi-Relational Embeddings as Knowledge Graph Representations for Robotics Applications · Georgia Tech · PhD

Tried and failed

relational graph convolutional networks applied to knowledge graph enrichment. Outcome: worse than baseline. Reason: None

Commonsense for Zero-Shot Natural Language Video Localization · Virginia Tech · Master's

Tried and failed

graph neural network embeddings for clustering applied to pairwise difference latent type identification. Outcome: worse than baseline. Reason: Node embeddings produced false positives and false negatives, underperforming baseline hierarchical clustering.

Essays On Machine Learning And Labor Economics · Penn · PhD

Tried and failed

uniform edge sampling from knowledge graphs applied to generating coherent multi-relational subgraphs. Reason: sampled disconnected or disjoint triplets that lacked semantic coherence for natural text expression

Collaborative AI Agents in the Era of Large Language Models · EPFL · PhD

Tried and failed

LLM reranking with web search retrieval applied to knowledge graph completion. Outcome: worse than baseline. Reason: Noisy external search results and mismatch with closed-world benchmark evaluation criteria.

From graphs to truth: towards efficient knowledge graph fusion for factual verification · Imperial · PhD

Lost to a baseline

DistMult and ComplEx embedding models were outperformed by TransE across all evaluated triple types in the threat knowledge graph.

Facilitating decision-making in large distributed systems with selfish and adversarial actors · Boston University · PhD

Considered and rejected

Considered and rejected: Rejected using purely data-driven GNN or translation-based embeddings without symbolic reasoning due to extreme data sparsity in clinical/standardization graphs.

Computational and human-based methods for knowledge discovery over knowledge graphs · Leibniz Universität Hannover · PhD

Considered and rejected

Considered and rejected: Rejected high-dimensional tensor GNN variants for MLN inference due to computational intractability on large knowledge graphs, choosing GNN with tunable embeddings.

Knowledge Reasoning with Graph Neural Networks · Georgia Tech · PhD

Left open

Evaluate 3D graph neural network architectures incorporating geometric features like distances, angles, and torsions for molecular activity prediction. Blocker: None

Exploring Graph Neural Networks for Molecular Activity Prediction · Harvard · Master's

Left open

Implement graph neural networks to encode drug compound chemical structures as molecular graphs instead of one-hot encodings for learning-to-rank models. Blocker: None

Robust learning to rank models and their biomedical applications · Boston University · PhD

Left open

Develop graph neural network embeddings or structure-agnostic representations for chemical structures in Bayesian optimization. Blocker: Lacks specific target problem, chemical dataset, and concrete baseline metrics

Bayesian optimisation in chemical problems · Imperial · PhD

Left open

Develop graph neural network architectures to predict edges and their directionality over higher-order representations such as graph-products and tensorial adjacencies. Blocker: None

Structure-aware graph representation learning using graph neural networks · Iowa State · PhD

Left open

Implement multitask learning neural networks with parallel regression and classification heads to predict physicochemical properties from GC chromatographic data. Blocker: None

Machine Learning for Structure-Agnostic Chemical Analysis from Chromatographic Data · Virginia Tech · Master's

Left open

Pre-train or augment graph neural network molecular activity predictors using large synthetic chemical datasets to improve performance on low-data empirical benchmarks. Blocker: None

Exploring Graph Neural Networks for Molecular Activity Prediction · Harvard · Master's

Left open

Develop graph neural network and transformer surrogate models trained on expanded MOF adsorption datasets to improve prediction and extrapolation. Blocker: None

Efficient and Accurate Incorporation of Flexibility and Defects Into the Modeling of Adsorption in Metal-Organic Frameworks · Georgia Tech · PhD

Left open

Implement a Graph Neural Network to replace the embedding prediction network in KD-EMD for spatial and relational knowledge distillation. Blocker: None

From Symbolic Reasoning to Object Embeddings: Advanced Approaches of Knowledge Distillation in Compacted Neural Networks · Texas Tech · PhD

Left open

Develop graph neural networks and graph variational autoencoders to optimize subgraph sampling and motif search on connectome graphs. Blocker: Lacks specific target architectures, quantitative efficiency benchmarks, or concrete graph sampling algorithms specified in the thesis

Deep Learning Tools for Next-Generation Connectomics · MIT · PhD