Week 5 update for my 12 Week Deep Learning Syllabus

I took a couple of weeks off for a mini vacation to Washington DC to attend my convocation with my family and then ended up getting sick which laid me up for a few days. Haven't been very productive the last couple weeks, it's time to get things back on track.

Week 5 Learning Material

Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization - Week 3

This week covered the following topics:

• Master the process of hyperparameter tuning
• Describe softmax classification for multiple classes
• Apply batch normalization to make your neural network more robust
• Build a neural network in TensorFlow and train it on a TensorFlow dataset
• Describe the purpose and operation of GradientTape
• Use tf.Variable to modify the state of a variable
• Apply TensorFlow decorators to speed up code
• Explain the difference between a variable and a constant

This course completes course 2 of 5 in the deep learning specialization.

Applied Techniques

I wrote in week 3 that I was going to deviate from the suggestions for applied techniques a bit an look at AI Scribes. I'm still exploring the tools for doing this and a good test case for a first application. In looking at using the ChatGPT API I decided I'd rather avoid the costs and test on my own system as long as it will scale to what I'm working on. I repurposed my gaming PC (goodbye BG3, Cyberpunk and a handful of other games I haven't completed) to use for my test lab.

OS: Debian 12
CPU: 12th Gen Intel(R) Core(TM) i5-12400 4.4 GHz
GPU: GeForce RTX 3070
Memory: 24G
Disk: 2G SSD

I have used whisper to translate the leaked support videos from Oracle in their recent data breach so I'm experimenting with using it to translate recordings of patient conversations to text, then using a local LLM to translate to SOAP.

ChatGPT suggested I classify a small image set of lesions which is also an interesting problem and in line with the Coursera material. MCP is getting a lot of attention right now and Hugging Face just released a free MCP Course that I'm interested in so I may start exploring that as well. Still figuring out tools and which problem I want to start solving for the applied learning work. Definitely no lack of exciting options.