Abstract: This report synthesises findings from 8 peer-reviewed papers addressing the following research question: What is the comparative effect of GraphSAGE's inductive sampling strategies versus GCN's transductive learning on F1-score stability during zero-shot graph anomaly detection across unseen domains. Traditional…

Abstract: This report synthesises findings from 11 peer-reviewed papers addressing the following research question: To what extent does the depth of Graph Convolutional Network layers influence the robustness and accuracy of graph anomaly detection systems under distributional shift in multi-domain settings. Deep convolutional…

Abstract: This report synthesises findings from 15 peer-reviewed papers addressing the following research question: What is the impact of Gaussian noise injection on the inference efficiency and prediction accuracy of spatiotemporal graph neural networks compared to standard diffusion-based approaches. Summary Continuously…

Abstract: This report synthesises findings from 14 peer-reviewed papers addressing the following research question: How do different attention mechanisms in Graph Attention Networks impact the cross-domain generalization of graph anomaly detection models when evaluated on heterogeneous node classification. Deep convolutional…

Abstract: This report synthesises findings from 12 peer-reviewed papers addressing the following research question: How does INT4 quantization of LLaVA-UHD impact accuracy on SEED-Bench visual reasoning subtasks compared to FP16 precision. Recent advancements in Chain of Thought (COT) generation have significantly improved the…

Abstract: This report synthesises findings from 15 peer-reviewed papers addressing the following research question: How does the inference efficiency of large language models compare when evaluated on adversarial reasoning benchmarks versus standard code generation tasks. We introduce self-invoking code generation, a new task…

Abstract: This report synthesises findings from 12 peer-reviewed papers addressing the following research question: How do inference latency and detection accuracy trade-offs differ between Mul-GAD and GCN-AE when subjected to increasing levels of graph structure noise. Anomaly detection is defined as discovering patterns that…

Abstract: This report synthesises findings from 14 peer-reviewed papers addressing the following research question: How does the robustness of graph-based anomaly detection models like Mul-GAD compare to GCN-AE under adversarial structural perturbations on the Reddit dataset. Real-time traffic prediction models play a pivotal…

Abstract: This report synthesises findings from 15 peer-reviewed papers addressing the following research question: How robust is GADT3's performance against adversarial perturbations in the ACM and DBLP benchmarks when combined with dimensionality reduction, as measured by changes in detection accuracy and. Anomaly detection…

Abstract: This report synthesises findings from 10 peer-reviewed papers addressing the following research question: What is the impact of adversarial training on the anomaly detection accuracy of GAS versus Mul-GAD when evaluated on perturbed graph structures. Deep neural networks (DNN) have achieved unprecedented success in…

Abstract: This report synthesises findings from 14 peer-reviewed papers addressing the following research question: How does the integration of different data augmentation techniques affect the performance of GADT3 in cross-domain anomaly detection on the ACM and DBLP benchmarks compared to baseline methods. Anomaly detection…

Abstract: This report synthesises findings from 12 peer-reviewed papers addressing the following research question: Can graph anomaly detection models trained with feature masking generalize effectively across heterogeneous datasets like Amazon and Yelp without domain-specific fine-tuning. Graph anomaly detection attracts…

Abstract: This report synthesises findings from 15 peer-reviewed papers addressing the following research question: How does homophily-guided self-supervision impact the reasoning accuracy of billion-parameter LLMs on perturbed social graph datasets compared to standard pretraining. High-level automation is increasingly…

Abstract: This report synthesises findings from 10 peer-reviewed papers addressing the following research question: What is the impact of test-time training frameworks versus static supervised models on inference latency and detection accuracy for graph-based anomaly detection. Deep convolutional neural networks have performed…

Abstract: This report synthesises findings from 4 peer-reviewed papers addressing the following research question: How do self-supervised graph anomaly detection methods compare to supervised baselines in robustness when 20\% of node features are masked on Amazon and Yelp datasets. Graph anomaly detection (GAD) suffers from…

Abstract: This report synthesises findings from 14 peer-reviewed papers addressing the following research question: How does knowledge distillation affect the zero-shot image-text retrieval accuracy and inference throughput of CLIP variants on the Flickr30k and MSCOCO datasets. Abstract The rapid evolution of large language…

Abstract: This report synthesises findings from 15 peer-reviewed papers addressing the following research question: How does the inference throughput of GNN-based anomaly detectors compare to tree ensemble baselines on large-scale synthetic graphs with varying sparsity levels. Abstract Tabular data, spreadsheets organized in…

Abstract: This report synthesises findings from 15 peer-reviewed papers addressing the following research question: What is the impact of graph size scaling on the detection accuracy of supervised GNN models versus traditional methods in standardized GAD benchmarks. As industries become automated and connectivity technologies…

Abstract: This report synthesises findings from 4 peer-reviewed papers addressing the following research question: How does text data augmentation impact the zero-shot image classification accuracy of CLIP compared to ALIGN on ImageNet variants. Contrastive Language-Image Pretraining (CLIP) has emerged as a powerful paradigm…

Abstract: This report synthesises findings from 3 peer-reviewed papers addressing the following research question: What is the effect of counterfactual text augmentation on the robustness of multimodal models against adversarial perturbations in VQA tasks. Deep neural networks (DNNs) are an indispensable machine learning tool…

Abstract: This report synthesises findings from 13 peer-reviewed papers addressing the following research question: How does the few-shot accuracy of Flamingo compare to domain-adapted models on multimodal code vulnerability detection benchmarks when subjected to adversarial perturbations in natural language. In this report,…

Abstract: This report synthesises findings from 14 peer-reviewed papers addressing the following research question: To what extent do Visual Language Models like Flamingo generalize to unseen CWE categories in zero-shot settings compared to fine-tuned code-specific multimodal models. A convolutional neural network (CNN) is one…

Abstract: This report synthesises findings from 10 peer-reviewed papers addressing the following research question: What is the comparative robustness of Llama3 versus GRU models against varying levels of injected noise in multivariate time series forecasting tasks measured by RMSE. Gradient Boosted Decision Trees (GBDT's) are…

Abstract: This report synthesises findings from 4 peer-reviewed papers addressing the following research question: How do non-negative activation constraints in multimodal evidential transformers impact inference throughput compared to standard ReLU baselines on image-text retrieval benchmarks. Abstract Plant diseases cause…

Abstract: This report synthesises findings from 9 peer-reviewed papers addressing the following research question: How does Multi-Objective Reinforcement Learning impact pass@k scores on HumanEval-Java and HumanEval-JavaScript compared to single-objective PPO under varying user preference distributions. Abstract The rapid…