PROPOSAL: libre.new-co.ai
Libre is a (proposed) open/free collaborative teams based AI research group. There are very many people who want to become experts in Deep Learning/Machine Learning/Artificial Intelligence and there exists lots of good online resources. What's missing is an organisation where you can learn from your peers and that's what Libre intends to fill.
Libre means:
- Open to everyone - we organise so that you join a team of peers
- Open publication - all work will be written up and published
- Open source - all source code will be released to github under CC BY-NC
- Free to join and collaborate - no fees are charged
In order that we work well together we do ask for commitment:
- At least four hours per day, five days a week for a whole sprint - it's not fair on other team members otherwise
- Respect, patience and compassion - we all come from very different backgrounds
- Honesty and integrity - (please see our code of conduct)
We work in teams:
- Our main aim is to provide practical experience, this comes from working out problems with peers
- We wil try to coordinate teams so all members have the same working hours
- We loosely follow the scrum framework
Practicalities:
- At the start we will have no funding, this means picking problems that don't require masses of GPU (e.g. can be run in colab).
- We need to decide on tools - maybe these are:
- Pytorch
- Github
- Slack of Discord for 'disposable' chat
- Wiki for structured results
FAQ
Q: Why Agile/Scrum?
We are no fan of the religious nature of Agile/Scrum, but we need a common framework so that everyone can work together. We really mean that people will be working in self contained teams with known sprint goals.
Q: What are the prerequisites
Python3: https://dev.to/javinpaul/top-5-places-to-learn-python-programming-for-free-m4c
Maths: https://www.coursera.org/specializations/mathematics-machine-learning
Q: Why CC BY-NC?
We recognise that different people will have different goals. Some people may be very academic and have as their main goal lots of publications. Some people may want to commercially exploit the results, perhaps by forming a start up. We want to support everyone, CC BY-NC was chosen to both ensure that code was available for academic use whilst allowing the key parties to relicence under other terms so not impeeding startup growth.
Q