Automatically processing natural language is quite a challenge for a machine. Complex structure and associations that are easy for humans to understand, are often not easy to reflect in machine representations.
This workshop aims at shedding light on the usefulness of neural networks as the goto machine learning algorithm in natural language processing and aims at explaining some of the basic concepts and building blocks. Using multiple exemplary data sets, we will try to convey basic concepts for understanding, working with and implementing neural networks. We will also discuss some recent trends of the field.
There is a broad spectrum of neural network frameworks out there which aim at lowering the threshold of working with neural networks. For one of those frameworks, “Keras“, there will be programming examples. Please note, however, that this workshop can not provide a general introduction to programming.
Objectives of the workshop
- Understanding the basics of neural networks
- Understanding challenges for a computer working with natural language
- Discussion about some recent advances, risks and challenges for neural networks in the NLP context
- Basics in programming a neural network in Keras
Target audience
This workshop is considered a practical application-focussed Introduction. A basic understanding of mathematical principles and programming (in Python) is required. Other than that, the workshop is meant for everybody with curiosity in neural networks and their role in natural language processing.
Curriculum (May be subject to change)
1-2 Introduction and Basics of NNs
3-4 Framework and Basics of Implementation
5-8 Examples and approaches for different tasks
9 Recent Trends