http://repository.iiitd.edu.in/xmlui/handle/123456789/1725 Year-2024 Wed, 23 Sep 2026 17:53:34 GMT 2026-09-23T17:53:34Z http://repository.iiitd.edu.in/xmlui/handle/123456789/2167 Title: Parameterized algorithms to find cohesive subgraphs with small diameter Authors: Sakhuja, Raghav; Majumdar, Diptapriyo (Advisor) Abstract: Given an undirected graph G = (V, E), an s-club is a vertex subset S ⊆ V (G) such that G[S] has diameter at most s. Formally, an s-Club problem asks if the input graph has an s-club with at least k vertices. There have been plethora of works on s-Club problem with fixed value of s = 2 as well as with vertex and edge triangle constraints. Vertex-l-Triangle-s-Club and Edge-l-Triangle-s-Club are defined as follows. Vertex-l-Triangle-s-Club asks to find an s-club S with at least k vertices such that every vertex u ∈ S is part of at least l triangles of G[S]. Similarly, the Edge-l-Triangle-s-Club asks to find an s-club S with at least k vertices such that every edge uv ∈ G[S] is part of at least l triangles of G[S]. In this work, we consider the s-Club problem from the perspective of parameterized complexity as follows. In the first part, we focus on s-Club when parameterized by the size of a given cluster edge deletion set of the graph. We prove that s-Club is FPT with a singly exponential running-time. We provide FPT algorithm for both Vertex-l-Triangle-s-Club and Edge-l-Triangle-s-Club when parameterized by the size of a vertex cover of the input graph. In the second part, we focus on Edge-l-Triangle-2-Club and Vertex-l-Triangle-2- Club parameterized by the size of the cluster vertex deletion set. In the third part, we provide a FPT algorithm for Vertex-l-Triangle-2-Club parameterized by the dual parameters treewidth and number of triangle. We provide a FPT algorithm and a compression kernel for s-Club parameterized by cluster edge deletion and provide a linear kernel for Vertex-l-Triangle-s-Club parameterized by feedback edge set. Wed, 27 Nov 2024 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/2167 2024-11-27T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/2158 Title: Book companion: data visualization and web development Authors: Dahiya, Aditya; Bagler, Ganesh (Advisor) Abstract: This project involved developing an interactive online book companion for Prof. Bagler’s book and datasets. It began with the analysis of a recipe dataset containing over 1,18,000 samples and 65+ columns, where key trends and insights were identified. Using tools like Plotly and Matplotlib, over 500 graphs of 20+ types, including interactive world heat maps and word clouds, were created. Chapters of the book were reviewed to provide feedback and suggest reader-friendly static graphs to ensure accessibility without overwhelming the audience. The final output was deployed using HTML and Plotly Express, integrating dynamic visualizations for datasets such as FlavorDB, RecipeDB, FNDDS, and openFoods, providing a seamless, web- based experience. Wed, 27 Nov 2024 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/2158 2024-11-27T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/2157 Title: Constructing knowledge graphs from images Authors: Garg, Chaitanya; Jain, Tanishq; Mutharaju, Raghava (Advisor) Tue, 10 Dec 2024 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/2157 2024-12-10T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/2156 Title: AI/ML models for predicting cardio-vascular diseaeses Authors: Sharma, Naman; Gupta, Anubha (Advisor) Abstract: This project presents a comprehensive framework to address class imbalance in the PTB-XL electrocardiogram (ECG) dataset by integrating diffusion models with state-of-the-art deep learning architectures, including 1D-CNN-LSTM and Gap-5 ST-CNN-LSTM. The diffusion model employs Gaussian noise addition with multiple noise schedules, such as cosine, to generate synthetic ECG samples that augment minority classes. This approach ensures the temporal and physio- logical integrity of the signals while balancing the dataset effectively. The diffusion model is powered by a 1D U-Net architecture with spatio-temporal capabilities, incorporating residual blocks, attention mechanisms, and temporal convolutions. The forward diffusion process progressively adds Gaussian noise, while the reverse process reconstructs clean signals. This facilitates the generation of realistic ECG samples that mimic the original signal distribution and enhance model robustness. For classification, we leverage a hybrid pipeline combining Swin Transformers and 1D-CNN- LSTM. The 1D-CNN extracts fine-grained temporal features, while the LSTM captures sequential dependencies, making it well-suited for time-series ECG data. The Gap-5 ST-CNN-LSTM, a state-of-the-art architecture, further enhances the framework with its optimized spatio-temporal convolutional layers and LSTM units, offering superior feature representation and temporal understanding. Evaluation metrics such as F1 score, AUC, and class-wise recall demonstrate significant improvements in both minority class representation and overall classification accuracy. By combining diffusion-based augmentation with advanced deep learning architectures, this framework delivers a robust solution for multi-label, multi-class ECG diagnostics on imbalanced biomedical datasets. Wed, 27 Nov 2024 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/2156 2024-11-27T00:00:00Z