Article: Dynamic optimisation algorithm for online learning intervention based on cognitive-emotional multimodal fusion Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.70 pp.22 - 46 Abstract: Aiming at the problem that online learning systems are difficult to perceive and optimise learners' internal states in real time, this paper proposes a dynamic optimisation algorithm based on cognitive-emotional load model and multimodal fusion. The algorithm constructed a theoretical framework of cognition and emotion collaborative computing, and used a hierarchical deep reinforcement learning architecture to realise the continuous space optimisation of intervention strategies. The dataset constructed in the simulation environment, compared with a variety of cutting-edge baselines, the algorithm can significantly improve the standardised learning benefit to 85.7, and reduce the incidence of harmful cognitive emotional overload events to 9.3%. This research provides a feasible path with both theory and technology for the construction of 'state adaptation' intelligent education system. Inderscience Publishers - linking academia, business and industry through research

Title: Dynamic optimisation algorithm for online learning intervention based on cognitive-emotional multimodal fusion

Authors: Na Tian

Addresses: Directly Affiliated College, Beijing Open University, Beijing, 100081, China

Abstract: Aiming at the problem that online learning systems are difficult to perceive and optimise learners' internal states in real time, this paper proposes a dynamic optimisation algorithm based on cognitive-emotional load model and multimodal fusion. The algorithm constructed a theoretical framework of cognition and emotion collaborative computing, and used a hierarchical deep reinforcement learning architecture to realise the continuous space optimisation of intervention strategies. The dataset constructed in the simulation environment, compared with a variety of cutting-edge baselines, the algorithm can significantly improve the standardised learning benefit to 85.7, and reduce the incidence of harmful cognitive emotional overload events to 9.3%. This research provides a feasible path with both theory and technology for the construction of 'state adaptation' intelligent education system.

Keywords: cognitive emotional load; multi-modal fusion; deep reinforcement learning; DRL; online learning intervention.

DOI: 10.1504/IJICT.2026.154388

International Journal of Information and Communication Technology, 2026 Vol.27 No.70, pp.22 - 46

Received: 12 Jan 2026
Accepted: 27 Feb 2026
Published online: 25 Jun 2026 *

Title: Dynamic optimisation algorithm for online learning intervention based on cognitive-emotional multimodal fusion

Keep up-to-date