arXiv subjects

Raka Moni

Publications and source records attributed to Raka Moni.

4 recordsLinked to original sources

arXiv2026

Reasoning Over Recall: Evaluating the Efficacy of Generalist Architectures vs. Specialized Fine-Tunes in RAG-Based Mental Health Dialogue Systems

The deployment of Large Language Models (LLMs) in mental health counseling faces the dual challenges of hallucinations and lack of empathy. While the former may be mitigated by RAG (retrieval-augmented generation) by anchoring answers in trusted clinical sources, there remains an open question as to whether the most effective model under this paradigm would be one that is fine-tuned on mental health data, or a more general and powerful model that succeeds purely on the basis of reasoning. In this paper, we perform a direct comparison by running four open-source models through the same RAG pipeline using ChromaDB: two generalist reasoners (Qwen2.5-3B and Phi-3-Mini) and two domain-specific fine-tunes (MentalHealthBot-7B and TherapyBot-7B). We use an LLM-as-a-Judge framework to automate evaluation over 50 turns. We find a clear trend: the generalist models outperform the domain-specific ones in empathy (3.72 vs. 3.26, $p < 0.001$) in spite of being much smaller (3B vs. 7B), and all models perform well in terms of safety, but the generalist models show better contextual understanding and are less prone to overfitting as we observe in the domain-specific models. Overall, our results indicate that for RAG-based therapy systems, strong reasoning is more important than training on mental health-specific vocabulary; i.e. a well-reasoned general model would provide more empathetic and balanced support than a larger narrowly fine-tuned model, so long as the answer is already grounded in clinical evidence.

cs.CL↗

arXiv2025

Inside the Degree, Outside the Discipline? Testing an Asymmetric Appraisal Model of the Curricular Legitimacy Gap in Computing Education

Required broader coursework can secure participation without being recognised as legitimate computing knowledge. This study conceptualises this disconnect as a curricular legitimacy gap and tests an asymmetric appraisal model grounded in situated expectancy value theory. The model distinguishes curricular devaluation, judging broader coursework unnecessary or professionally irrelevant, from integrative intention, or willingness to reuse its learning. Survey data from 212 Computer Science and Engineering undergraduates in Bangladesh recruited through snowball sampling were analysed using robust structural equation modelling. Primary inference combined robust direct-path estimates with 5,000 respondent-level bootstrap resamples; alternative measurement, response-quality, and ordinal-estimator specifications were also examined. Perceived burden was positively associated with devaluation, which was negatively associated with intention. The standardised indirect association of burden with intention through devaluation was -0.350, 95% CI [-0.589, -0.169]. Perceived benefits were associated with stronger intention through lower devaluation, indirect association 0.141, 95% CI [0.058, 0.253], and an additional positive direct pathway. The model explained 47.6% of the variance in devaluation and 40.6% in intention. The direct burden-to-intention pathway was unsupported under the primary estimator but significant in the opposite-to-hypothesised direction under ordinal estimation; this residual path is therefore treated as estimator-dependent. The results support a distinction among requirements, valuation, intention, and behaviour and suggest that cost reduction and utility development address different curricular problems. Given the cross-sectional, nonprobability design and developing measures, all pathways are interpreted as associations rather than causal mediation.

cs.CY↗

arXiv2025

NormEval: A Unified Multi-Metric Framework for Evaluating Semantic Fidelity in Text Normalization

Text normalization methods such as stemming and lemmatization are fundamental components of NLP pipelines. As new normalization tools are developed for diverse languages, evaluation methodologies remain fragmented, relying on Compression Ratio, downstream accuracy, or sequence-to-sequence prediction scores in isolation, failing to distinguish between beneficial vocabulary reduction and harmful semantic distortion. Moreover, text normalization underpins intelligent systems in high-stakes domains, including clinical decision support and legal document analysis, and principled evaluation methodology is essential. This paper proposes NormEval, a unified, multilingual evaluation framework comprising five complementary metrics: Compression Ratio (CR), Model Performance Delta (MPD), Information Retention Score (IRS), Algorithm Effectiveness Score (AES), and Average Normalized Levenshtein Distance (ANLD). These metrics assess normalization quality across three dimensions: macro-level efficiency, downstream utility, and micro-level morphological fidelity. The framework operationalizes a Safety Gate hypothesis: ANLD functions as an intrinsic structural hygiene check, utilizing character-level divergence ($\Delta$) to reveal aggressive mutations that macro-level embeddings and downstream tasks mask. Comprehensive ablation experiments on both Bangla and English datasets show that all the components are indispensable, and that the removal of any individual metric leads to a decrease in at least one evaluation aspect, which ultimately results in misleading algorithm rankings.

cs.CL↗

arXiv2025

Less Is More: An Explainable AI Framework for Lightweight Malaria Classification

Background and Objective: Deep learning models have high computational needs and lack interpretability but are often the first choice for medical image classification tasks. This study addresses whether complex neural networks are essential for the simple binary classification task of malaria. We introduce the Extracted Morphological Feature Engineered (EMFE) pipeline, a transparent, reproducible, and low compute machine learning approach tailored explicitly for simple cell morphology, designed to achieve deep learning performance levels on a simple CPU only setup with the practical aim of real world deployment. Methods: The study used the NIH Malaria Cell Images dataset, with two features extracted from each cell image: the number of non background pixels and the number of holes within the cell. Logistic Regression and Random Forest were compared against ResNet18, DenseNet121, MobileNetV2, and EfficientNet across accuracy, model size, and CPU inference time. An ensemble model was created by combining Logistic Regression and Random Forests to achieve higher accuracy while retaining efficiency. Results: The single variable Logistic Regression model achieved a test accuracy of 94.80 percent with a file size of 1.2 kB and negligible inference latency (2.3 ms). The two stage ensemble improved accuracy to 97.15 percent. In contrast, the deep learning methods require 13.6 MB to 44.7 MB of storage and show significantly higher inference times (68 ms). Conclusion: This study shows that a compact feature engineering approach can produce clinically meaningful classification performance while offering gains in transparency, reproducibility, speed, and deployment feasibility. The proposed pipeline demonstrates that simple interpretable features paired with lightweight models can serve as a practical diagnostic solution for environments with limited computational resources.

cs.CV↗