Seiichi Koakutsu
Papers
2
Total Citations
7
H-Index
2
About
Seiichi Koakutsu is a researcher whose work focuses on the intersection of neural networks, evolutionary computation, and fault tolerance. His primary research areas include the evolutionary construction of neural network architectures, pulse neural networks, and learning algorithms designed to overcome fundamental limitations in artificial intelligence. Koakutsu made notable contributions by proposing a modified gene coding scheme for the evolutionary construction of Block-Based Neural Networks (BBNNs) that explicitly considers the possibility of component failure, a critical advancement for developing robust, fault-tolerant systems. His 2004 paper on this topic has garnered 4 citations, reflecting its specialized impact in the field of resilient neural design. Additionally, Koakutsu addressed the persistent problem of catastrophic forgetting in neural networks by developing a Hebbian learning rule for pulse neural networks (PNNs) with leaky integrate-and-fire neurons. This 2003 work, with 3 citations, introduced a method that allows new pattern learning without destroying previously acquired knowledge, while also promoting efficient synapse usage. Through these contributions, Koakutsu has advanced the practical reliability and learning stability of neural systems, offering valuable insights for researchers working on adaptive, fault-aware artificial intelligence.
Research Focus
Key Achievements
Top Papers
- 1
- 2