Kaile Su

Jinan University, Griffith University, Institute of Software

Papers

Total Citations

H-Index

About

Kaile Su is a researcher whose work bridges artificial intelligence, speech recognition, and real-time planning systems. His key research areas include recurrent neural network architectures, automatic speech recognition, and efficient pathfinding algorithms. Su made significant contributions to deep learning for audio processing, particularly through his work on Long Short-Term Memory Projection (LSTMP) networks, which optimize both speed and performance for time-series tasks like piano note recognition—a paper that has garnered 31 citations. He also advanced large vocabulary continuous speech recognition by developing a fast learning method for multilayer perceptrons, incorporating a preadjusting strategy and dynamic cosine-based learning rates to improve accuracy, earning 14 citations. Beyond speech, Su’s notable work on the BDD-based Dynamic A* algorithm for real-time replanning (7 citations) demonstrates his versatility in addressing computational efficiency in dynamic environments. His contributions reflect a commitment to making neural networks and planning systems faster and more practical for real-world applications, offering valuable insights for students and researchers exploring the intersection of machine learning and robotics.