मराठी संशोधन कार्यासाठी कृत्रिम बुद्धिमत्तेचा (AI) वापर

Journal of Research and Development A Multidisciplinary International Level Referred and Double Blind Peer Reviewed, Open Access ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(VII)| September2025 36   󰫦 󰭃 󰱮󰰌 (AI)    1,    2 1 󰱳󰰏     󰭈   󰰣 󰫦  󰱳 ,   󰲆 .  2  󰫦󰫦     󰭈   󰰣 󰫦  󰱳 ,   󰲆 .  Manuscript ID: JRD -2025-170910 ISSN: 2230-9578 Volume 17 Issue 9(VII) Pp. 36-41 September 2025 Submitted: 12 Aug. 2025 Revised: 25 Aug. 2025 Accepted: 14 Sept.2025 Published: 30 Sept. 2025  ( Abstract) 󰱿   󰭃󰫥󰫳 󰬴󰬀 󰭃 󰱮󰰌 (AI)   󰭈󰲑󰬴  󰲐  󰫦 󰭈 󰭃󰫥  .   󰫤󰭃 󰱳  ,  󰫽,    󰲒 󰲈   󰬉  󰭈󰲑󰬴   󰲐  .  󰲐󰬎 AI        , 󰰏, , 󰫤,  󰲄, 󰱿󰬎, OCR–HTR ,     󰬌   󰬊 󰱇  . AI 󰫳 󰬇 󰱚 󰭈   󰭈󰲑󰬴  ,  󰬎  󰬎 󰱨 󰬄  ,   󰬀 󰫥 󰲄  ,   󰬌  󰫽  󰫦󰫤 󰱿󰬎  . 󰫳 󰰌 ,   󰭑 , 󰬶–   , 󰲇   󰬄   󰰀󰱳 ( interdisciplinary )      󰫽 󰲐 AI  󰲑. ,   󰬄   , 󰭃󰫥 󰬉  ,  󰫦, 󰫤   󰲓  AI  󰫽 󰫽   󰬋 󰭈󰫦 .   󰬶, AI  󰰗󰭅, WordNet–OCR , 󰭈󰫤 󰫦󰬴  – 󰫦   󰫦   . ,   󰫦 AI     󰲐     󰱮󰰎󰫳 󰬇    󰬎   󰫤 󰱇󰰗 󰰉 . 󰫯 󰬄 ( Keywords) : 󰭃 󰱮󰰌 (AI),  ,  󰬶,  , 󰰏,  ( NMT ), 󰫤 ( Summarization) ,  󰲄,  . 󰭈󰬎      󰭃󰫥  󰬆  .  󰱮󰰌  ,    󰫦󰫤󰫤     󰭃󰫥 󰬆 󰭃 󰱮󰰌.   󰬆    󰲗󰫦   󰬉 󰫤  . Quick Response Code: Website: https://jrdrvb.org/ DOI: Creative Commons (CC BY-NC-SA 4.0) This is an open access journal, and articles are distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License, which allows others to remix, tweak, and build upon the work noncommercially, as long as appropriate credit is given and the new creations ae licensed under the idential terms. Address for correspondence:   ,  󰱳󰰏     󰭈   󰰣 󰫦  󰱳 ,   󰲆 .  How to cite this article: , .  .  ., & , .  .  . (2025).   󰫦 󰭃 󰱮󰰌 (AI) . Journal of Research & Development, 17(9(VII)), 36–41. Original Article Journal of Research and Development A Multidisciplinary International Level Referred and Double Blind Peer Reviewed, Open Access ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(VII)| September2025 37 󰱳, 󰭃󰲒, 󰱳, 󰫤, 󰫦, 󰫽,   󰭈󰫽 󰫤󰭃 AI    .   󰱮, 󰭈  󰫽󰭈   󰫳    󰲑 . 󰬶 󰬶,  󰬎, , ,    󰭃 󰰉 󰫽     󰲑  .  󰲂󰫦 AI     . 󰲑󰱯󰲆   󰫯 󰲑󰱯󰲆 󰭈  : 1.   󰱳 󰰎 󰲄 . 2.  󰫦 󰭃 󰱮󰰌   . 3. AI 󰫳     󰫦 . 4.  󰫳 󰬍  󰲑󰫦 . 󰭃 󰱮󰰌 󰬉   󰭃 󰱮󰰌  󰬉 󰭈  󰬀  󰰏  . 󰫽   󰱘󰲔  "󰱘󰲔 󰬎"  󰭈󰱨 󰭃  󰰉  ,  󰭈󰲅 󰬎 . 󰰈   󰲄 󰫦󰲑   󰬉  Machine Learning, Deep Learning, Neural Networks, Natural Language Processing  󰬀 󰬎 .  Google, Microsoft, OpenAI, IBM 󰫯 󰲇 󰬁 AI 󰲐   󰬎  󰰉  . 󰬶       . 󰭃    󰱳 󰭈󰫳 󰬂󰬂 .   󰱳     󰬎    󰭃󰫥   󰬎    : 1.  󰭓  – 󰬶     󰬄 ;   . 2.   – 󰫦, , ,  󰫽 󰭈     . 3.  󰬄  –   󰬎,  󰭃  󰰩  . 4. 󰲇  󰫦󰲑 – 󰬶  󰬉      . 5. 󰭃 󰬉  – 󰬶   󰲒    .   AI   1.    󰰏  :   󰬆 󰬊 󰲌 ( , 󰁰󰬋, -󰬎, 󰰌󰭃,  ,  󰰣,  .)   󰫼 󰭈󰲑󰬴 . 󰰏 󰬆  󰬉  , 󰭈, , 󰭈   󰬶 .    (󰬎--󰬎): 1.  󰭈󰲅 󰲀  (. “󰲒  󰰌  ”). 2. 󰲌  ( 󰫦,  󰭊, 󰲑󰭅 󰬎). 3. -󰬎󰬴󰲓 / API 󰰉  󰲐 󰫮. (  !). 4. 󰭈-󰭈󰲓: Unicode normalization, 󰬌 HTML/ , 󰬂 . 5. - (language detection) —  -󰫮     . 6. , 󰬎󰫦 󰲐󰬋, — 󰫰  . 7.   (, 󰲑, 󰭓, 󰬎, 󰭑). 8. 󰰏: supervised (  ) 󰲕 unsupervised (󰫮󰬎󰲔, topic modelling) . 9. 󰲓󰫮: semantic search  embeddings//󰲕 inverted index (Elasticsearch/Ignite )  . Journal of Research and Development A Multidisciplinary International Level Referred and Double Blind Peer Reviewed, Open Access ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(VII)| September2025 38  / -:  󰬁 󰰌󰭃 󰫦  󰱳   󰫦 󰰣 .     󰫳 󰰣 󰰏 . /󰭃:  -󰬎󰬴󰲓: Scrapy / BeautifulSoup- 󰬉 (   ).  󰲓󰫮: Elasticsearch / Solr 󰲕 Vector embeddings (sentence-transformers) 󰰉 semantic search.   󰭊 󰬉 (Prodigy/LabelStudio  󰬉). 󰫦 / :  󰲌󰫳  󰬎󰬴󰲓 󰰉 .    󰬉 󰬍󰫦 󰲒   —  󰫮󰲓  󰰍. 2. 󰬎  (Automatic Translation)  :    󰬬  (󰲕 󰬬   ) 󰰉󰲐  — Neural Machine Translation (NMT) 󰫳 󰬇.   : 1. Parallel corpus   (-󰬶 ). 2.  󰬎󰫳, normalization ( ) . 3.  : seq2seq transformers (NMT) 󰲕 API-󰲐 󰬉󰬁. 4. -󰲆 󰬄󰬶 / glossary   (󰲕 post-editing  ). 5.  fine-tune  (    ). 6.  -󰫤 (post-editing) 󰫦.  / -:     󰬶  󰰉 󰲇 󰭈  .  -󰫤   󰬀 󰱚󰫽 󰬄 . /󰭃:  Neural MT (transformer), back-translation   , glossary-based post-processing.  API  (. 󰱿  󰰈 󰰉) 󰲕 -󰫦 NMT 󰭉󰫮󰫦. 󰫦 / :   󰫯 low-resource  parallel data   — 󰫽     .  󰬎 󰫦, 󰬆/ . 󰫰 󰭈 󰲑   — 󰫽   󰬌. :  󰰍󰫳 󰬎  post-editing 󰫦 .  󰬎 glossaries  style-guides  . 3. 󰫤 (Summarization)  : /     — extractive ( 󰫮 ) 󰲕 abstractive (󰱿 󰫮 󰬁 ).   : 1. 󰬎 normalization. 2. 󰬉 󰭈     (󰬍󰫽/-/ 󰲐󰫳). 3. Extractive 󰱮: TextRank  graph-based 󰱮. 4. Abstractive 󰱮: Transformer-based seq2seq 󰬉 (pretrained language models). 5. 󰰀 multi-document summarization:  󰭓 󰭃 . 6. 󰬉: ROUGE 󰬎,   󰭈󰬋 󰰤󰲐. Journal of Research and Development A Multidisciplinary International Level Referred and Double Blind Peer Reviewed, Open Access ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(VII)| September2025 39  / -:    Literature Review .  󰫤   󰫤. 󰫦 / :  Abstractive 󰬉 - “󰬉” ( /  󰫾)   — 󰫽 󰭓 󰰣 󰬌.  󰫦 󰫦   — 󰭓 󰫦 󰱳. :    󰭓 /󰫦 .  ---  . 4.   (Linguistic Research)  :  , 󰱿, 󰬄󰬍󰱘, -, 󰬄  lexical resources  .   : 1.  (corpus)   — , , 󰫽. 2. NLP-: tokenization, POS tagging, lemmatization, dependency parsing. 3. Morphological analyzers  finite-state transducers  ( 󰱮 󰰉 󰫼). 4. Dialect clustering: unsupervised clustering 󰲕 embedding-based clustering . 5. WordNet/semantic network  󰫼 - 󰰗.  / -:   󰬄  󰬅.  󰱿󰫽  󰲆 . /󰭃:  POS taggers, morphological analyzers, treebanks, WordNet-type resources.  󰬴󰲛󰰩    (. 󰱳󰰏/󰬎 󰰣󰰘). 󰫦 / :   󰲓  treebank 󰬌.  -   . :  󰬎  󰲆 󰰉   .   (open) resource policy  󰱿 󰬬   󰱳. 5. 󰭈  󰬎 󰲄 (OCR / HTR)  :   󰬎  󰬎  󰫮󰬎󰬀 󰰉 — Optical Character Recognition (OCR) 󰲕 Handwritten Text Recognition (HTR).   : 1. 󰬎󰲓: 󰱍-󰲐󰬉 󰬎. 2. 󰭈-󰭈󰲓: 󰬎󰫮, , noises . 3. OCR/HTR   (   󰱇󰫤  ). 4. 󰬎-󰫮: language model-󰲐 󰬎-  -󰫤. 5.   󰬶  (󰭓, 󰲑).  / -:   󰬎 󰬶  󰰉 󰱨 󰬄 .  -󰬶 󰬉 󰰣  󰲄. 󰫦 / :   󰬁󰫸,  ,  󰬎󰲓 (orthography)  OCR error 󰬎  . Journal of Research and Development A Multidisciplinary International Level Referred and Double Blind Peer Reviewed, Open Access ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(VII)| September2025 40  󰬎 󰬀󰫦  HTR  󰬋. :  󰬎󰲓  󰭈-󰭈󰲓 󰫤 󰱳; 󰬎-󰫮    .   󰬶 orthography-normalization 󰰉󰬉  . 6. 󰫯󰲒    (Survey & Social Data Analysis)  : 󰰘󰫤, 󰭈󰲅,  󰫦   AI-󰲐 󰲄 󰰉 󰬍󰫦 .   : 1.   : structured (MCQ)  unstructured (󰬉 󰭈󰲅 󰰌). 2. Data cleaning, missing-value treatment, sampling weights. 3. Open-ended 󰰌 topic modelling (LDA), clustering, supervised classification . 4. causal analysis vs correlation — causal inference  󰫦. 5. network analysis:  , 󰭈 󰱿󰱇󰲒 . 6. geospatial analysis: 󰭈   .  / -:    󰬶 󰱳󰰩 󰬅  󰰎 󰫽 󰲄.  󰫦  󰭈󰲑󰬴 󰬾󰰗-. 󰫦 / :  -󰭈󰫽 biased 󰬍󰫦.    󰬉   ethics . :      sampling plan .  󰱿󰱇󰲒  anonymize ; 󰬌 IRB/ethics approvals 󰫱. 7.  󰬌 (Visualization)  : , 󰱇, , 󰫮 󰫦  󰬎󰰉  󰬍󰫦   —  󰫼   󰫦-󰫦 .   : 1. 󰫦󰭈  󰲆 (story)   . 2. 󰫰 󰫦 : - line,  histogram, - bar,    choropleth map. 3. interactive filters, tooltips  — 󰫽󰫦   . 4. explainability: model-󰲐  SHAP/LIME  󰫮󰬂 . 5. Dashboards  accessibility ( 󰬁, color-blind 󰭉󰲒 palettes) 󰫤 .  / -:       .  󰫳 󰫯 󰬍󰰩   󰫦. /󰭃:  Interactive: Plotly/Dash, Streamlit, Power BI, Tableau.  Static: matplotlib, ggplot, d3.js ().  Geospatial: Leaflet, kepler.gl. 󰫦 / :  󰲒 󰲕 󰭋 󰬌 (e.g., axis-manipulation) 󰫦 󰭃  . Journal of Research and Development A Multidisciplinary International Level Referred and Double Blind Peer Reviewed, Open Access ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(VII)| September2025 41  󰱚  interactive visualization performance issue. :  󰭈󰫽 visualization  󰬎󰲆 legend  data source 󰱳.   (storytelling) 󰱮 visuals  —     󰫦󰫦 . 󰭃 󰱮󰰌  1.   󰭑 . 2.  󰭈󰲑󰬴  . 3.     󰫥 . 4. 󰬶–   . 5. 󰰀󰱳 (Interdisciplinary)  . 6.  󰬎   . 󰫦 1.    󰬄 . 2.   󰫤 󰫦 󰫼 󰫮. 3.  󰫦 󰫦 . 4. AI 󰫳 󰲒󰫳  󰲒 󰬍󰫦  . 5.  󰫤    󰭈󰲅.  󰲑   AI 󰭈 󰫼    󰬌  :    󰬶 󰲗󰫦.    󰱳 AI  󰰗󰭅 .  Marathi WordNet, OCR, Speech-to-Text   .   AI 󰭈󰫤  󰫦.       󰱨 󰭈󰬉 .      (Global Network) 󰫦 . 󰬍󰫦   󰫦 󰭃 󰱮󰰌     . 󰫰 󰭈󰫤, 󰰌   󰬄 󰬉    󰬎  󰬎 󰫦 󰰉 . AI   󰱮󰰌 󰬎󰫦    󰱇󰲒 .     ,  󰭃󰫥  󰰉  󰬍  󰲆 󰫦 󰰉 . 󰫦 󰬎 () 1. "󰭃 󰱮󰰌: 󰬉  " – .  , 󰲇 󰫽 󰲐, . 2. " 󰭃󰫥   " – .  ,  󰭈, . 3. "   " – 󰭈.  󰰏, 󰫽 󰭈, . 4. " 󰱮" – . . . ,  󰭈, 󰬉. 5. " 󰭈󰲑󰬴:   " – 󰭈. . . , 󰱳 󰭈, .  󰲌 (󰱇 󰫸)  󰲇   󰲂 – https://marathivishwakosh.maharashtra.gov.in  IIT Bombay  CFILT (Centre for Indian Language Technology) – https://www.cfilt.iitb.ac.in  Google Scholar –   󰫼.  ArXiv.org – AI  NLP    .