Research Associate Punjab University College of Information Technology
Provide insights on model selection and adaptation for HAR data augmentation; leading to the development of effective GAN-based strategies and advancing the field of sensor data processing and ML in wearable technology.
Design and implement specialized GAN architectures, including HM-GANs, Multi-branch GANs, TTS-GANs, MF-GANs, and RTS-GANs tailored to address challenges in sequential and multimodal data processing for HAR applications.
Enable robust model generalization across diverse HAR datasets, including UNIMIBSHAR, WISDM, PAMAP2, and UCI-HAR, as well as capture temporal / long-term dependencies by integrating techniques, such as KANs and TKANs.