Foundation/AI/FLEX/Projects/MachineLearningGames: Difference between revisions

From MozillaWiki
< Foundation‎ | AI‎ | FLEX
Jump to navigation Jump to search
(Add Project Information)
(Undo revision 1209578 by Mwalsh (talk) Undo Changes)
Line 3: Line 3:


== Overview ==
== Overview ==
'''Project Lead:''' Becca Ricks
'''Project Lead:''' Chad Sansing


'''Area of Focus:''' Machine Decision Making (Impact Goal)
'''Area of Focus:''' Focusing on how discrimination, a lack of openness and  transparency, and a lack of responsibility all contribute to discimination and negatively impact digital inclusion.


'''Knowledge Level Required:''' Intermediate
'''Knowledge Level Required:''' None / Any


'''This project is actively seeking contributors you can sign up to participate here.'''
'''This project is actively seeking contributors you can sign up to participate here.'''
Line 13: Line 13:
== Project Information ==
== Project Information ==
==== Project Description: ====
==== Project Description: ====
MoFo has an opportunity to educate itself and learn together about this year's impact goal, machine decision-making (AI). The project idea is to develop a set of curricula (suggested readings & questions) based on the challenges raised in the literature review–data privacy concerns, bias in training data, algorithms & behaviors, lack of accountability mechanisms, etc. The idea is that anyone from MoFo can self-organize into small discussion groups to talk about each topic.
We are going to make low-fi and no-fi social, collaborative games that demonstrate how biased inputs lead machine learning products to discriminate against under-represented and wrongly-represented people. Inspirations include Werewolf, Inhuman Conditions (https://robots.management/), indie table-top role-playing games (TTRPGs), storytelling, matching games, and more.


==== Expected Outcomes ====
==== Expected Outcomes: ====
The hope is that MoFo as an organization is able to develop a more in-depth knowledge about the challenges posed by AI that goes reacting to the news cycle. We should all be able to speak and work confidently around the impact goal.  
I want to learn how each part of machine learning functions & connects to the others from trainers & data sets, to general adversarial networks & generators & discriminators, to the products that use the machine learning & the people it affects.


==== Help Required ====
==== Help Required: ====
Based on the challenge areas identified in the literature review, the hope is that contributors will volunteer to tackle a topic and develop out a suggested curriculum on that topic. These are suggestions–groups can diverge from the topics, too.
Everyone should feel welcome to join, learn, & design games with us, & to help us prototype & publish games to share at All-Hands, MozFest, & beyond via CC licensing & .pdf distribution. Gamers, artists, storytellers, layout designers, editors et al.


==== Required Time Commitments ====
==== Required Time Commitments ====
The project will take place over 1-2 months. The Project Lead will spend 1 hour working on the project per week and contributors would need to allow for 1 hour per week to fully get involved.
Gather contributors --> begin cadence of meetings, mini-design sprints --> prototype --> share at All-Hands --> submit to MozFest --> refine for MozFest --> play at MozFest --> publish betas before the end of 2019. The Project Lead will spend 2 hours working on the project and contributors would need to allow for 2 hours per week to fully get involved.


== Project Updates ==
== Project Updates ==
More to come...
More to come...

Revision as of 16:21, 27 March 2019

Machine Learning Games

MoFo FLEX 2019 - Project Proposal

Overview

Project Lead: Chad Sansing

Area of Focus: Focusing on how discrimination, a lack of openness and transparency, and a lack of responsibility all contribute to discimination and negatively impact digital inclusion.

Knowledge Level Required: None / Any

This project is actively seeking contributors you can sign up to participate here.

Project Information

Project Description:

We are going to make low-fi and no-fi social, collaborative games that demonstrate how biased inputs lead machine learning products to discriminate against under-represented and wrongly-represented people. Inspirations include Werewolf, Inhuman Conditions (https://robots.management/), indie table-top role-playing games (TTRPGs), storytelling, matching games, and more.

Expected Outcomes:

I want to learn how each part of machine learning functions & connects to the others from trainers & data sets, to general adversarial networks & generators & discriminators, to the products that use the machine learning & the people it affects.

Help Required:

Everyone should feel welcome to join, learn, & design games with us, & to help us prototype & publish games to share at All-Hands, MozFest, & beyond via CC licensing & .pdf distribution. Gamers, artists, storytellers, layout designers, editors et al.

Required Time Commitments

Gather contributors --> begin cadence of meetings, mini-design sprints --> prototype --> share at All-Hands --> submit to MozFest --> refine for MozFest --> play at MozFest --> publish betas before the end of 2019. The Project Lead will spend 2 hours working on the project and contributors would need to allow for 2 hours per week to fully get involved.

Project Updates

More to come...

Retrieved from "https://wiki.allizom.org/index.php?title=Foundation/AI/FLEX/Projects/MachineLearningGames&oldid=1209579"