The material for each day will become available in the morning. Go to the drop-down menu to access the material once it's available. Lectures will be a mixture of theory and coding examples along with coding exercises.

Week 1

Lecture 1 - 29th Jan 2024: 

• General course introduction, Introduction to audio signals

Lecture 2 - 30th Jan 2024: 

• Introduction to Transforms and Spectrograms

Lecture 3 - 31st Jan 2024: 

• Audio augmentation, VGGish and YAMNet

Lecture 4 - 1st Feb 2024: 

• Introduction to machine learning for bioacoustic classifiers

• Applying bioacoustic classifiers to real audio files

• CNNs for bioacoustics 101

Lecture 5 - 2nd Feb 2024: 

• Introduction to SonicVisualiser for bioacoustics annotation 

• Setup Raspberry Pi to create Automatic Recording Units (ARUs)

3rd Feb  2024:

• Field trip to Intaka Island Nature Reserve to deploy ARUs and collect acoustic data

Week 2

Lecture 6 - 5th Feb  2024: 

• Transfer learning for bioacoustics

• CNN-RNNs for bioacoustics binary classification 

Lecture 7 - 6th Feb  2024:

• More advanced applications of CNN-RNNs for whale bioacoustics

• Unsupervised audio clustering and classification

Lecture 8 - 7th Feb  2024:

TBC

Lecture 9 - 8th Feb 2024:

• Student presentations

• Discussion: How to implement a real-time monitoring system

• Discussion: Ethics of machine learning and wildlife monitoring

• Discussion: What other areas, related to animals and ecology, can DL help with?

Class test - 9th Feb 2024: 

• Coding test

Week 3

Lecture 10 - 12th Feb 2024:

• Introduction to Pose Estimation

• Introduction to Animal Pose Estimation

• Introduction to Multiple Animal Pose Estimation

• Annotation for Pose Estimation

Lecture 11 - 13th Feb 2024:

• Basic structure for a pose estimation algorithm

• Advanced techniques for improving pose estimation

• Evaluating a pose estimation algorithm

• Regression based pose estimation

Lecture 12 - 14th Feb 2024:

• Discussion of techniques used for pose estimation improvement

• Introduction to acceleration data 

Lecture 13 - 15th Feb 2024:

• Introduction to machine learning for acceleration data and behavioural identification

• Random forest for behavioral identification from acceleration data

• Deep learning for behavioral identification from acceleration data 

Lecture 14 - 16th Feb 2024:

• Introduction to a fully convolutional neural network; Vnet

• Application of the Vnet on the acceleration data