Karishma Thakrar

Chloe Pomeroy, Aleksandar Pramov, Karishma Thakrar et al.

This paper explores the applications of quantum annealing (QA) and classical simulated annealing (SA) to a suite of combinatorial optimization problems in machine learning, namely feature selection, instance selection, and clustering. We formulate each task as a Quadratic Unconstrained Binary Optimization (QUBO) problem and implement both quantum and classical solvers to compare their effectiveness. For feature selection, we propose several QUBO configurations that balance feature importance and redundancy, showing that quantum annealing (QA) produces solutions that are computationally more efficient. In instance selection, we propose a few novel heuristics for instance-level importance measures that extend existing methods. For clustering, we embed a classical-to-quantum pipeline, using classical clustering followed by QUBO-based medoid refinement, and demonstrate consistent improvements in cluster compactness and retrieval metrics. Our results suggest that QA can be a competitive and efficient tool for discrete machine learning optimization, even within the constraints of current quantum hardware.

Karishma Thakrar

Graph Retrieval-Augmented Generation (GRAG or Graph RAG) architectures aim to enhance language understanding and generation by leveraging external knowledge. However, effectively capturing and integrating the rich semantic information present in textual and structured data remains a challenge. To address this, a novel GRAG framework, Dynamic Graph Retrieval-Agumented Generation (DynaGRAG), is proposed to focus on enhancing subgraph representation and diversity within the knowledge graph. By improving graph density, capturing entity and relation information more effectively, and dynamically prioritizing relevant and diverse subgraphs and information within them, the proposed approach enables a more comprehensive understanding of the underlying semantic structure. This is achieved through a combination of de-duplication processes, two-step mean pooling of embeddings, query-aware retrieval considering unique nodes, and a Dynamic Similarity-Aware BFS (DSA-BFS) traversal algorithm. Integrating Graph Convolutional Networks (GCNs) and Large Language Models (LLMs) through hard prompting further enhances the learning of rich node and edge representations while preserving the hierarchical subgraph structure. Experimental results demonstrate the effectiveness of DynaGRAG, showcasing the significance of enhanced subgraph representation and diversity for improved language understanding and generation.