Audio is one of the toughest unstructured data to mine and process. How can machine learning help in processing audio signals.

This technical talk aims at solving audio problems using Machine Learning and Deep Learning to solve problems like audio file classification, audio file tagging/labeling.

As part of the topic below topics will be covered; All the codes shared for the approaches are part of competitions that have led to top 5-10% of ranking in kaggle/crowdai competitions.

1. Structured Feature Extraction for traditional Machine Learning Algorithms - Demo of Feature Extraction and Random Forest/GBM model

2. Multi Dimensional Feature Extraction for Convolution Neural Networks - Demo of Spectrum Analysis using Transfer Learning using ResNet

3. Temporal Dimension Feature Extraction for Recurrent Neural Networks - Demo of Recurrent Neural Networks

 
 

Outline/Structure of the Talk

1. Structured Feature Extraction for traditional Machine Learning Algorithms - Demo of Feature Extraction and Random Forest/GBM model

2. Multi Dimensional Feature Extraction for Convolution Neural Networks - Demo of Spectrum Analysis using Transfer Learning using ResNet

3. Temporal Dimension Feature Extraction for Recurrent Neural Networks - Demo of Recurrent Neural Networks

Learning Outcome

Understanding of Machine Learning Applications to Audio.

Target Audience

Machine Learning Enthusiasts/Researchers/Newbies

Prerequisites for Attendees

Understanding of Machine Learning Concepts.

schedule Submitted 1 year ago

Public Feedback