arXiv subjects

Kamal Premaratne

Publications and source records attributed to Kamal Premaratne.

7 recordsLinked to original sources

arXiv2026

Evaluation of Contextual Understanding in Large Language Models

Large Language Models (LLMs) demonstrate impressive performance across diverse NLP tasks, yet their ability to exhibit genuine contextual understanding remains uncertain. Traditional evaluation metrics such as perplexity, BiLingual Evaluation Understudy (BLEU), or surface-level accuracy fail to reveal how well LLMs extract, integrate, and reason over contextual information--a gap particularly critical in question answering, where models must align responses with contextually grounded knowledge rather than memorized associations. We propose a novel knowledge graph-based evaluation framework introducing Semantic Structural Similarity for KGs (S3KG), a hybrid similarity measure integrating structural and semantic similarity into a continuous evaluation score, alongside a diagnostic framework for categorizing reasoning errors. To validate this pipeline, we evaluate S3KG against established metrics on a curated question-answer (QA) benchmark, demonstrating its effectiveness in measuring correctness, faithfulness, and interpretability in LLM-generated responses.

cs.CL↗

arXiv2026

A Graph Neural Network Model for Real-Time Gesture Recognition Based on sEMG Signals

For seemless control of advanced hand prostheses and augmented reality, accurate and immediate hand gestures recognition is essential. Surface electromyography (sEMG) signals obtained from the forearm are commonly employed for this purpose. In this paper, we present a novel approach for sEMG representation that utilizes graph networks which contain information about muscle activation patterns in the forearm. Based on these graph networks, we have developed a machine learning algorithm capable of real-time hand gesture recognition using a graph neural network. The algorithm's performance was evaluated using sEMG signals acquired from myoband, which has 8 electrodes placed around the forearm, involving 8 healthy subjects. The proposed method demonstrated an average classification accuracy of 99\%, surpassing the performance of state-of-the-art techniques. The average time for both graph construction and prediction stood at 48ms utilizing a M1 pro CPU, rendering the approach well-suited for real-time applications.

cs.AI↗

arXiv2026

Semantic Uncertainty Quantification of Hallucinations in LLMs: A Quantum Tensor Network Based Method

Large language models (LLMs) exhibit strong generative capabilities but remain vulnerable to confabulations, fluent yet unreliable outputs that vary arbitrarily even under identical prompts. Leveraging a quantum tensor network based pipeline, we propose a quantum physics inspired uncertainty quantification framework that accounts for aleatoric uncertainty in token sequence probability for semantic equivalence based clustering of LLM generations. This offers a principled and interpretable scheme for hallucination detection. We further introduce an entropy maximization strategy that prioritizes high certainty, semantically coherent outputs and highlights entropy regions where LLM decisions are likely to be unreliable, offering practical guidelines for when human oversight is warranted. We evaluate the robustness of our scheme under different generation lengths and quantization levels, dimensions overlooked in prior studies, demonstrating that our approach remains reliable even in resource constrained deployments. A total of 116 experiments on TriviaQA, NQ, SVAMP, and SQuAD across multiple architectures including Mistral-7B, Mistral-7B-instruct, Falcon-rw-1b, LLaMA-3.2-1b, LLaMA-2-13b-chat, LLaMA-2-7b-chat, LLaMA-2-13b, and LLaMA-2-7b show consistent improvements in AUROC and AURAC over state of the art baselines.

cs.CL↗

arXiv2025

A Quantum Tensor Network-Based Viewpoint for Modeling and Analysis of Time Series Data

Accurate uncertainty quantification is a critical challenge in machine learning. While neural networks are highly versatile and capable of learning complex patterns, they often lack interpretability due to their ``black box'' nature. On the other hand, probabilistic ``white box'' models, though interpretable, often suffer from a significant performance gap when compared to neural networks. To address this, we propose a novel quantum physics-based ``white box'' method that offers both accurate uncertainty quantification and enhanced interpretability. By mapping the kernel mean embedding (KME) of a time series data vector to a reproducing kernel Hilbert space (RKHS), we construct a tensor network-inspired 1D spin chain Hamiltonian, with the KME as one of its eigen-functions or eigen-modes. We then solve the associated Schr{\"o}dinger equation and apply perturbation theory to quantify uncertainty, thereby improving the interpretability of tasks performed with the quantum tensor network-based model. We demonstrate the effectiveness of this methodology, compared to state-of-the-art ``white box" models, in change point detection and time series clustering, providing insights into the uncertainties associated with decision-making throughout the process.

cs.LG↗

arXiv2022

Clustering Edges in Directed Graphs

How do vertices exert influence in graph data? We develop a framework for edge clustering, a new method for exploratory data analysis that reveals how both vertices and edges collaboratively accomplish directed influence in graphs, especially for directed graphs. In contrast to the ubiquitous vertex clustering which groups vertices, edge clustering groups edges. Edges sharing a functional affinity are assigned to the same group and form an influence subgraph cluster. With a complexity comparable to that of vertex clustering, this framework presents three different methods for edge spectral clustering that reveal important influence subgraphs in graph data, with each method providing different insight into directed influence processes. We present several diverse examples demonstrating the potential for widespread application of edge clustering in scientific research.

cs.SI↗

arXiv2016

Consensus in the Presence of Multiple Opinion Leaders: Effect of Bounded Confidence

The problem of analyzing the performance of networked agents exchanging evidence in a dynamic network has recently grown in importance. This problem has relevance in signal and data fusion network applications and in studying opinion and consensus dynamics in social networks. Due to its capability of handling a wider variety of uncertainties and ambiguities associated with evidence, we use the framework of Dempster-Shafer (DS) theory to capture the opinion of an agent. We then examine the consensus among agents in dynamic networks in which an agent can utilize either a cautious or receptive updating strategy. In particular, we examine the case of bounded confidence updating where an agent exchanges its opinion only with neighboring nodes possessing 'similar' evidence. In a fusion network, this captures the case in which nodes only update their state based on evidence consistent with the node's own evidence. In opinion dynamics, this captures the notions of Social Judgment Theory (SJT) in which agents update their opinions only with other agents possessing opinions closer to their own. Focusing on the two special DS theoretic cases where an agent state is modeled as a Dirichlet body of evidence and a probability mass function (p.m.f.), we utilize results from matrix theory, graph theory, and networks to prove the existence of consensus agent states in several time-varying network cases of interest. For example, we show the existence of a consensus in which a subset of network nodes achieves a consensus that is adopted by follower network nodes. Of particular interest is the case of multiple opinion leaders, where we show that the agents do not reach a consensus in general, but rather converge to 'opinion clusters'. Simulation results are provided to illustrate the main results.

cs.SI↗

arXiv2015

An Efficient Polyphase Filter Based Resampling Method for Unifying the PRFs in SAR Data

Variable and higher pulse repetition frequencies (PRFs) are increasingly being used to meet the stricter requirements and complexities of current airborne and spaceborne synthetic aperture radar (SAR) systems associated with higher resolution and wider area products. POLYPHASE, the proposed resampling scheme, downsamples and unifies variable PRFs within a single look complex (SLC) SAR acquisition and across a repeat pass sequence of acquisitions down to an effective lower PRF. A sparsity condition of the received SAR data ensures that the uniformly resampled data approximates the spectral properties of a decimated densely sampled version of the received SAR data. While experiments conducted with both synthetically generated and real airborne SAR data show that POLYPHASE retains comparable performance to the state-of-the-art BLUI scheme in image quality, a polyphase filter-based implementation of POLYPHASE offers significant computational savings for arbitrary (not necessarily periodic) input PRF variations, thus allowing fully on-board, in-place, and real-time implementation.

cs.CE↗