Monitoring AI enhancements in Urdu
Google Scholar papers focused on Urdu AI, LLMs, and datasets
… the individualLLMoutputs and found that GPT-4 could not predict much data on sentiment… of Gemini in theUrdulanguage. We found that the alphabets ofUrduare derived from the Ara…
We present LEGAL-UQA, the firstUrdulegal question-answering dataset derived fromPakistan's constitution. This parallel English-Urdudataset includes 619 question-answer pairs, …
… Developments within the scope ofartificialintelligenceand language processing have usheredin … It is a part of the GPT-3.5LLMsubtype and is programed to have information from an …
… semantics led to the development ofLLM. These models help computers … Currently,AIdevelopment prioritizes about two dozen high-… This study delves into impartingAIwith the ability to …
… In Baltistan Balti language is significantly influenced byUrdu… gap in the performance ofLLM-based machine translation and … So by using advanced models, technology, andAIwe can …
… The rapid increase in the application of Artificial Intelligence (AI) across a diverse spectrum… LLMs for supportingUrdu, a low resource language with limited data coverage inLLM’s pre-…
…AIassistant, ensuring that the evaluation is both comprehensive and cost-efficient. ForAI…We also performed aLLMbased evaluation of the outputs of the models for the generation …
… 4 Methodology: Abstractive Text Summarization forUrduLanguage UsingLLM(ATSUL)Large language models areartificialintelligence(AI) models that have been trained on massive …
… exclusive with other regional languages such asUrdu, Marathi, or Hindi. Recognizing …work in many other languages, including English,Urdu, Persian, and Chinese. Algorithms on …
… 2.1NLPand non-Latin alphabet languages When we mentionNLP, we invoke the various… alphabets such as Arabic, Chinese,Urdu, Thai, etc. Various works have been carried out in …
…Urdu, a language spoken mainly in Pakistan and India, is considered a low-resource languagein terms ofnaturallanguageprocessing(NLP) research.Urduhas been the focus of this …
… Samples are code-mixed English-Urdutext, whereUrduis written in romanised form. Oursubmission is limited to one of the subtasks-Multi Class classification and we leverage …