Robust benchmark datasets crucial for improving machine translation systems: VC KU

Posted by admin_kas on 2025-03-02 22:42:16 |

Share: WhatsApp| Facebook| Twitter| Visits: 175

KS News Desk 

Srinagar, Mar 02: The Vice Chancellor (VC) Kashmir University (KU) said that developing robust benchmark datasets was crucial for improving machine translation systems and fostering linguistic inclusivity in the digital era.

She was speaking during a four-day workshop organised by the varsity in collaboration with the International Institute of Information Technology (IIIT-Hyderabad), on Developing IL-IL Benchmark Data at the varsity from February 25 to 28, 2025.

The event was organised to advance research in Indian-language-to-Indian-language (IL-IL) machine translation.

As per the KU handout issued here, the initiative aligned with the university’s commitment to interdisciplinary research in Artificial Intelligence and Natural Language Processing.

Vice Chancellor KU, Prof. Nilofer Khan, in her remarks highlighted the significance of the initiative in strengthening AI-driven language technologies.

The workshop was part of the HIMANGY (Hindustani Machini Anuvaad Technology) consortium under the broader BHASHINI Project funded by the Ministry of Electronics and Information Technology (MeitY), Government of India, aimed to create benchmark datasets for enhancing machine translation accuracy in multiple Indian languages, including Hindi, Telugu, Kashmiri, Kannada, Gujarati, Sindhi, Dogri, Odia, Punjabi, Urdu, and English.

Prof. Dipti Mishra from IIIT-Hyderabad, emphasized the importance of building strong linguistic benchmarks to support digital inclusivity for Indian languages.

“Ensuring representation of diverse languages in machine translation will bridge communication gaps and make digital content more accessible,” she said.

Prof. Aadil Amin Kak, Department of Linguistics KU, stressed on the workshop’s role in ensuring greater visibility for Kashmiri in the digital space and said that such initiatives would help break linguistic barriers and enhance communication across different language communities.

The event saw the participation of 70 researchers from leading institutions across India, including IIT-Patna, Punjabi University-Patiala, IIIT-Bhubaneswar, CDAC-Noida, University of Hyderabad, DAIICT-Ahmedabad, and GCW-Jammu.

Experts engaged in collaborative efforts to refine machine translation models, ensuring better accuracy and effectiveness in Indian language processing.