Digests » 23
this week's favorite
In this video let's cover a simple and practical example to understand neural networks.
Skeletor attempts to provide a lightweight wrapper for research code with two goals: (1) make it easy to track experiment results and data for later analysis and (2) orchestrate many experiments in parallel without worrying too much. The first goal is satisfied using track for logging experiment metrics. You can get the experiment results in a nice Pandas DataFrame with it, it logs in a nice format, and it can back up to S3. The second goal is satisfied using ray to parallelize multi-gpu grid searches over various experiment configurations. This is an improvement over some other setups because it allows us to use a proper distributed execution framework to handle trial scheduling.
These notes and tutorials are meant to complement the material of Stanford’s class CS230 (Deep Learning) taught by Prof. Andrew Ng and Prof. Kian Katanforoosh. For questions / typos / bugs, use Piazza. These posts and this github repository give an optional structure for your final projects. Feel free to reuse this code for your final project, although you are expected to accomplish a lot more. You can also submit a pull request directly to our github.
Ethics are too subjective to guide the use of AI, argue some legal scholars.
AI is a field filled with technical terms. It can be difficult to pin down exactly what they mean, particularly if you don’t work directly with data every day. That’s why we’ve created a glossary of 50 AI terms that frequently come up in discussions about AI. If you can lock down these basics, you should be able to hold your own in any discussion about machine learning. Let’s run through them in alphabetical order.