Huma Abia Kanta

Huma Abia Kanta (born 24 May 2009) is a computer science scholar as well as an Indian student researcher from Guwahati, Assam. Huma is making contributions to machine learning (ML), artificial intelligence (AI), and computer literacy education in Northeast India. She presented a research paper at an international AI conference in Azerbaijan in 2025.

Early life and education

Guwahati, Assam is where Huma grew up since birth. She went to Royal Global School. She received the Chairman’s Medal for academic excellence in 2025 and a 100% scholarship at Royal Global School in the same year.

Research and publications

Huma primarily focused on research studies in AI applications for sustainability, socio-technological modeling, and computational linguistics.

ML-Based Prediction of Phycocyanin Purity

Huma presented her study (ML-Based Prediction of Phycocyanin Purity) in an International Conference event organized by Nakhchivan Technical University in collaboration with other 5 universities that was held 30–31 October 2025.

In the phycocyanin purity study, the model achieved a mean absolute error of 0.058, compared with a laboratory standard deviation of 0.31, indicating that its predictions were substantially more precise than typical measurement variability. In her paper, she discussed the use of machine-learning models to forecast levels of pigment purity, which is crucial in the sustainable bio-resource industry.

Seq2Seq Reconstruction of Sanskrit Phonology

In the Sanskrit phonology project, her sequence-to-sequence model with Luong attention achieved an accuracy of 0.59, compared to a baseline of 0.21, using a curated Buddhist-lexicon dataset derived from Tang-era Siddham–Hanzi transliteration materials.

In this study, Seq2Seq Reconstruction of Sanskrit Phonology (which has been accepted for RegiCON 2025, organised by Gauhati University and Assam Skill University) she used a Recurrent Neural Network (RNN) encoder-decoder with Luong attention to reconstruct phonetic relationships based on historical Sanskrit corpora.