Abstract: This report synthesises findings from 3 peer-reviewed papers addressing the following research question: How does the robustness of Mul-GAD against adversarial perturbations in graph structures compare to methods like GraphSAINT or Cluster-GCN when evaluated using perturbation resilience metrics (e.g.,. Machine…

Abstract: This report synthesises findings from 3 peer-reviewed papers addressing the following research question: How does the training time scalability of Mul-GAD vary against GAT and GraphSAGE as the number of nodes increases in large-scale graph benchmarks. With a long history of traditional Graph Anomaly Detection (GAD)…

Abstract: This report synthesises findings from 9 peer-reviewed papers addressing the following research question: What is the peak memory consumption difference between Mul-GAD and standard GNN baselines during anomaly detection on the Reddit and Yelp datasets. BACKGROUND: As mortality rates decline, life expectancy…

Abstract: This report synthesises findings from 3 peer-reviewed papers addressing the following research question: How does the inference throughput of Mul-GAD compare to GraphSAGE and GAT on large-scale graph datasets when measured in graphs per second. MOTIVATION: In recent years, there have been various efforts to overcome…

Abstract: This report synthesises findings from 4 peer-reviewed papers addressing the following research question: How does the robustness of Mul-GAD against adversarial attacks compare to that of SVM and Random Forest when evaluated on perturbed versions of CORA or Citeseer datasets using F1-score and AUC-ROC as. Support…

Abstract: This report synthesises findings from 4 peer-reviewed papers addressing the following research question: How does the performance of Mul-GAD compare to traditional shallow learning methods (e.g., SVM, Random Forest) on benchmark text classification datasets like 20 Newsgroups or Reuters when evaluated. Abstract The…

Abstract: This report synthesises findings from 9 peer-reviewed papers addressing the following research question: How does Directional Preference Alignment affect code generation pass@1 scores on HumanEval when models are subjected to adversarial syntax perturbations compared to standard RLHF. Abstract—Large Language Models…

Abstract: This report synthesises findings from 7 peer-reviewed papers addressing the following research question: What is the inference latency overhead of Directional Preference Alignment versus RLHF when generating code solutions for the MultiPL-E dataset. In this report, we introduce Qwen2.5, a comprehensive series of…

Abstract: This report synthesises findings from 1 peer-reviewed paper addressing the following research question: How does Directional Preference Alignment impact code generation pass@1 scores on low-resource languages in the MultiPL-E benchmark compared to standard RLHF. Large Language Models (LLMs) have garnered remarkable…

Abstract: This report synthesises findings from 12 peer-reviewed papers addressing the following research question: What is the impact of communication round frequency on the convergence speed and final inference throughput of federated vision-language models when using near-optimal compression versus QSGD. This paper provides…

Abstract: This report synthesises findings from 11 peer-reviewed papers addressing the following research question: To what extent does the federated learning setup with diversity-driven client selection improve the generalization of multimodal models (e.g., CLIP, ViLBERT) across unseen domains compared to. Natural Language…

Abstract: This report synthesises findings from 13 peer-reviewed papers addressing the following research question: How does the use of diversity-driven client selection in federated multimodal learning compare to traditional client sampling methods in terms of convergence speed and final retrieval accuracy on. Federated…

Abstract: This report synthesises findings from 1 peer-reviewed paper addressing the following research question: Does the diversity-driven selection in FedDiverse reduce the convergence rounds required for CodeLlama to achieve target accuracy on the MBPP dataset relative to federated averaging in heterogeneous. Large language…

Abstract: This report synthesises findings from 9 peer-reviewed papers addressing the following research question: How does cross-domain alignment in federated multimodal models perform when evaluated on out-of-distribution benchmarks, using metrics like CLIP score or BLEU for multimodal reasoning tasks. Medical…

Abstract: This report synthesises findings from 7 peer-reviewed papers addressing the following research question: How does FedDiverse's client selection strategy impact CodeLlama's pass@1 scores on the HumanEval benchmark under extreme non-IID code distribution compared to standard FedAvg. Federated learning (FL) is a machine…

Abstract: This report synthesises findings from 6 peer-reviewed papers addressing the following research question: Does diversity-driven federated averaging improve the zero-shot reasoning capabilities of CodeLlama on the MATH benchmark when trained on heterogeneous client data distributions. Large Language Models (LLMs) have…

Abstract: This report synthesises findings from 10 peer-reviewed papers addressing the following research question: How does diversity-driven client selection in federated learning affect the pass@1 scores of CodeLlama-7B on the HumanEval benchmark under extreme non-IID code distribution scenarios. Trained on massive publicly…

Abstract: This report synthesises findings from 5 peer-reviewed papers addressing the following research question: What is the impact of feature-oriented regulation methods like \$Psi\$-Net on the inference efficiency of federated multimodal models under non-IID data distributions, measured by throughput and accuracy.…

Abstract: This report synthesises findings from 3 peer-reviewed papers addressing the following research question: How does the dynamic meta-layer aggregation approach in Fed-DPRoC compare to other federated learning defense mechanisms (e.g., Krum, Median) in terms of inference accuracy and communication overhead. The rapid…

Abstract: This report synthesises findings from 11 peer-reviewed papers addressing the following research question: How does the structural similarity of feature embeddings in collaborative multimodal federated learning models scale with increasing degrees of non-IID data, as measured by cosine similarity across. In parallel…

Abstract: This report synthesises findings from 9 peer-reviewed papers addressing the following research question: How does the robustness of Fed-DPRoC scale with the number of Byzantine clients compared to baseline federated averaging in cross-domain settings such as federated natural language processing tasks. Federated…

Abstract: This report synthesises findings from 10 peer-reviewed papers addressing the following research question: How does stochastic layer fusion affect the cross-domain generalization accuracy of multimodal vision-language models on wilds benchmarks compared to deterministic fusion. This review critically distinguishes…

Abstract: This report synthesises findings from 2 peer-reviewed papers addressing the following research question: How does the diversity-driven client selection in FedDiverse compare to FLCert's defense mechanisms in terms of inference efficiency and generalization performance on the LEAF Shakespeare dataset,. In parallel…

Abstract: This report synthesises findings from 13 peer-reviewed papers addressing the following research question: What is the impact of varying the compression ratio in Fed-DPRoC's Johnson-Lindenstrauss-based mechanism on model accuracy and throughput in federated learning with large-scale multimodal datasets. Federated…

Abstract: This report synthesises findings from 10 peer-reviewed papers addressing the following research question: What is the effect of FLCert's client selection strategy on the federated learning performance in a cross-domain setting, evaluated on both Shakespeare and Reddit datasets, measured by model F1. Federated…