Special Topics on Bioinformatics (2KV)
| Course no.: | 365.036 |
| Lecturer: | Stefan Kremer |
| Times/locations: | Mon, 3:30-5:00pm, room K223B (Start: Mon, Oct 1, 2007) |
| Mode: | KV, 2h, weekly |
| Registration: | KUSSS |
Motivation:
Neural Networks for Bioinformatics
This course presents artificial neural networks, recurrent neural networks and recursive neural networks as a paradigm for solving problems in bioinformatics.We begin with an overview of artificial neural networks, move on to some specific architectures that prove useful for bioinformatics. We cover a number of successful exemplars.
Topic List:
Introduction:- Neural networks in the context of computation and artificial intelligence
- McCulloch and Pitts
- Perceptrons
- Multi-layer perceptrons
- Strengths, weaknesses and limitations of neural networks
- Tricks of the trade
- Overfittiing and Parameter Selection
- Training Regimens and Other Considerations
- Classes of problems in bioinformatics
- Challenges for neural network applications to bioinformatics
- IOHMMs
- Recurrent Networks (including LSTM)
- Recursive Networks
- Protein Folding
- Bond Prediction
- Primer Design
- Promoter Detection
- NUMT Identification
- Others
Requirements:
- Students should be versed in elementary calculus, probability, algorithms and programming.