92 research outputs found
Improving citation network scoring by incorporating author and program committee reputation
No full textImproving Citation Network Scoring by Incorporating Author and Program Committee Reputation
No full textPublication venues play an important role in the
scholarly communication process. The number of publication
venues has been increasing yearly, making it difficult for
researchers to determine the most suitable venue for their
publication. Most existing methods use citation count as the metric
to measure the reputation of publication venues. However, this
does not take into account the quality of citations. Therefore, it is
vital to have a publication venue quality estimation mechanism.
The ultimate goal of this research project is to develop a novel
approach for ranking publication venues by considering
publication history. The main aim of this research work is to
propose a mechanism to identify the key Computer Science
journals and conferences from various fields of research. Our
approach is completely based on the citation network represented
by publications. A modified version of the PageRank algorithm is
used to compute the ranking scores for each publication. In our
publication ranking method, there are many aspects that
contribute to the importance of a publication, including the
number of citations, the rating of the citing publications, the time
metric and the authors’ reputation. Known publication venue
scores have been formulated by using the scores of the
publications. New publication venue ranking is taken care by the
scores of Program Committee members which derive from their
ranking scores as authors. Experimental results show that our
publication ranking method reduces the bias against more recent
publications, while also providing a more accurate way to
determine publication quality
Hindi Reading Comprehension : Do Large Language Models Exhibit Semantic Understanding?
No full textIn this study, we explore the performance of four advanced Generative AI models—GPT-3.5, GPT-4, Llama3, and HindiGPT, for the Hindi reading comprehension task. Using a zero-shot, instruction-based prompting strategy, we assess model responses through a comprehensive triple evaluation framework using the HindiRC dataset. Our framework combines (1) automatic evaluation using ROUGE, BLEU, BLEURT, METEOR, and Cosine Similarity; (2) rating-based assessments focussing on correctness, comprehension depth, and informativeness; and (3) preference-based selection to identify the best responses. Human ratings indicate that GPT-4 outperforms the other LLMs on all parameters, followed by HindiGPT, GPT-3.5, and then Llama3. Preference-based evaluation similarly placed GPT-4 (80%) as the best model, followed by HindiGPT(74%). However, automatic evaluation showed GPT-4 to be the lowest performer on n-gram metrics, yet the best performer on semantic metrics, suggesting it captures deeper meaning and semantic alignment over direct lexical overlap, which aligns with its strong human evaluation scores. This study also highlights that even though the models mostly address literal factual recall questions with high precision, they still face the challenge of specificity and interpretive bias at times
A statistical machine translation approach to Sinhala Tamil language translation
No full textData-driven approaches to Machine Translation have come to the fore of Language Processing Research over the past decade. The relative success in terms of robustness of Example Based and Statistical approaches have given rise to a new optimism and an exploration of other data-driven approaches such as Maximum Entropy language modeling. Much of the work in the literature however, largely report on translation between languages within the European Family of languages. This research is an attempt to cross this language family divide in order to compare the performance of these techniques on Asian languages. In particular, this work reports on Statistical Machine Translation experiments carried out between language pairs of the three major languages of Sri Lanka: Sinhala, Tamil and English. Results indicate that current models perform significantly better for the Sinhala-Tamil pair than the English-Sinhala pair. This in turn appears to confirm the assertion that these techniques work better for languages that are not too distantly related to each other
Bootstrapping the Lexicon Building Process for Machine Translation between ‘New’ Languages
No full textOntological Knowledge Inferring Approach based on Term-Clustering and Intra-Cluster Permutations
No full text
Auto Generation of Gold Standard, Class Labeled Data Set and Ontology Extension Tool [QuadW]
No full text- …
