Machine Learning

The Machine Learning (ML) community aims to provide a platform for anyone interested in helping computers learn from data. We mainly target people with backgrounds in data science, machine learning and data visualisation, but also statisticians, computer scientists and anyone looking to learn about the technical side of ML are more than welcome.

Topics

The goal would be to have a platform for technical (focus on math and/or implementation) questions about:

• supervised learning
• unsupervised/self-supervised learning
• deep learning and neural networks
• reinforcement learning
• data collection and pre-processing
• data visualisation
• data-driven computing
• applied statistics
• statistical learning theory
• predictive modelling
• Bayesian modelling
• applying ML models to data
• deployment of ML pipelines
• computational neuroscience
• prompt engineering
• ...

Before asking a question, users are expected to have invested at least some effort to find an answer by themselves (cf. this meta discussion). Questions with readily available answers on e.g. Wikipedia will probably be closed. If this search was unsuccessful or the found answer was incomprehensible or has dubious origins, we will be happy to answer the question. However, the question should clarify why the search results do not answer the question. Answered questions from the StackExchange network can be asked again here, as long as licences are respected (do not copy questions you do not own).

Off-Topic

To keep the community focused on ML, questions about the following topics should be asked elsewhere:

• pure coding questions (should be asked on software.codidact.com).

  E.g. How can I visualise this data with matplotlib? would be on-topic, but How can I change the background colour of plots in matplotlib? should not be a question for this community.

• pure math and statistics questions (should be asked on math.codidact.com).

  E.g. What integrals do I need to compute the expectation of this random variable? would be okay, but How do I compute this complex integral? would not be okay.

• system administration questions (should probably be asked on powerusers.codidact.com or linux.codidact.com).

  E.g. Are there any libraries that I can use to deploy my model using docker? is allowed, but How do I set up docker? would not be suited.

• questions about other parts of Artificial Intelligence (AI) that have nothing to do with ML or ML algorithms. For these questions, a computer science community (cf. https://cs.stackexchange.com/) might be better suited.

  E.g. How does the AlphaZero chess engine learn to play chess? would be okay because AlphaZero involves learning, but How can I build a chess engine using alpha-beta pruning? does not involve any ML and is therefore not suited for this community.

• requests for prompts or improvements to prompts for generating data (this requires a different kind of knowledge and is often considered to be the alchemy branch of ML).

  E.g. How does chain-of-thought prompting help the model to improve predictions? is a valid question, but How can I fix this prompt for generating images with better lightning? is not okay.

• bug reports or other complaints about software (look for a bug tracker instead).

• questions about non-technical aspects of ML (e.g. opinions, ethics, legal, etc.).

• anything that is not (indirectly) related to machine learning.

Community Situation

Overlap on codidact will probably be mostly software, math and powerusers/linux (as indicated above). Maybe the AI Tech proposal could also be incorporated into this community (also, see below).

Currently, the community is shattered over Cross Validated and Data Science. There is also a sub-reddit that is (at the time of writing) protesting against what Reddit is doing.

Additional Features

This community will definitely need support for MathJax and code blocks, but not sure if these are really additional.

Given that machine learning is becoming more and more well-known in non-technical contexts, it might be useful to set up a separate discussion zone where non-technical discussions, prompt engineering, ethics, etc. could get a place in the community. This might make it possible to incorporate the AI Tech proposal. However, I didn't completely think this through yet.

I welcome any sort of feedback and/or suggestions for improvement!

PS: I would greatly appreciate it if someone would have time/be interested to ask some questions on ML. I tend to find it easier to answer questions than to ask them.

posted over 1 year ago

1y ago

mr Tsjolder‭

11 9 3