TY - GEN
T1 - Nature-Inspired Hyperparameter Optimization for Liquid Neural Networks in Skin Allergy Prediction
AU - Sekhar, M. Chandra
AU - Maram, Balajee
AU - Sahukaru, Jayavardhanarao
AU - Mrunalini, G.
AU - Prasan, U. D.
AU - Sathishkumar, V. E.
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025
Y1 - 2025
N2 - Accurate prediction of skin allergy is essential for early diagnosis and successful treatment. Conventional machine learning models will fail with intricate skin datasets because of inadequate hyperparameter tuning. This paper introduces a new framework that uses Nature-Inspired Optimization Algorithms, i.e., Particle Swarm Optimization (PSO) and Genetic Algorithms (GA), to optimize hyperparameters of Liquid Neural Networks (Liquid NN) to predict skin allergy. Experimental performance using a benchmark dermatological database showed outstanding improvement in performance with 94.7% accuracy, 92.5% precision, and 93.2% recall compared to traditional optimization techniques by 6-9%. Dynamism and usability of nature-inspired search paradigms in the liquidity of the liquid architecture generated phenomenal generalization to novel data. The model described herein has tremendous potential for predictive models to be enhanced in dermatological research and patient care by early diagnosis and correct diagnosis.
AB - Accurate prediction of skin allergy is essential for early diagnosis and successful treatment. Conventional machine learning models will fail with intricate skin datasets because of inadequate hyperparameter tuning. This paper introduces a new framework that uses Nature-Inspired Optimization Algorithms, i.e., Particle Swarm Optimization (PSO) and Genetic Algorithms (GA), to optimize hyperparameters of Liquid Neural Networks (Liquid NN) to predict skin allergy. Experimental performance using a benchmark dermatological database showed outstanding improvement in performance with 94.7% accuracy, 92.5% precision, and 93.2% recall compared to traditional optimization techniques by 6-9%. Dynamism and usability of nature-inspired search paradigms in the liquidity of the liquid architecture generated phenomenal generalization to novel data. The model described herein has tremendous potential for predictive models to be enhanced in dermatological research and patient care by early diagnosis and correct diagnosis.
KW - Dermatological Diagnosis
KW - Hyperparameter Optimization
KW - Liquid Neural Networks
KW - Nature-Inspired Optimization
KW - Skin Allergy Prediction
UR - https://www.scopus.com/pages/publications/105035192022
U2 - 10.1109/ICCCA66364.2025.11325412
DO - 10.1109/ICCCA66364.2025.11325412
M3 - Conference contribution
AN - SCOPUS:105035192022
T3 - 2025 IEEE 7th International Conference on Computing, Communication and Automation, ICCCA 2025
BT - 2025 IEEE 7th International Conference on Computing, Communication and Automation, ICCCA 2025
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2025 IEEE 7th International Conference on Computing, Communication and Automation, ICCCA 2025
Y2 - 28 November 2025 through 30 November 2025
ER -