Overview v1.0.0🚀
This is an overview post for the Waldo data collection site v1.0.0. This will mention all the issues that have been assigned and/or currently assigned. We will also discuss future features we could to improve the project.
What is the goal for this side project? ⭐
The project involves the creation of a data collection site where users can submit their own video clips and vote on others, with the goal of building a diverse and representative dataset for training a machine learning model. Some of the benefits of this project include the potential to improve the fairness and integrity of online gaming, as well as the opportunity to gather valuable data for machine learning research. However, there are also several roadblocks that we will need to tackle, including issues related to data quality and privacy, as well as the challenges of detecting complex and varied cheating behaviors. We will need to carefully consider these and other factors as we move forward with the project.
What features does it have directly to an end user or community? 💫
- The project will allow the submission of video clips; users would be able to upload their own video clips, from a video hosting provider in this case YouTube, of video game gameplay for inclusion in the dataset.
- It will also allow the voting on video clips where users would be able to view and vote on other users' video clips to indicate whether they believe the clip contains the correct gameplay footage type. This could help to improve the quality and relevance of the dataset by providing a way for the community to provide feedback and input. This will benefit us as managing and going through every submission manually is not a valid solution.
How users can get involved, such as by submitting video clips or voting on others' submissions? 🎉
We are inviting the community to participate in the project by submitting video clips of gameplay footage that you believe may contain instances of cheating or hacking. You can do this through our data collection website. In addition, you can help to improve the quality and relevance of the dataset by voting on other users' video clips and indicating whether you believe the video contains the correct gameplay. Your input will be valuable in helping us to build a comprehensive and accurate dataset for training our machine learning model.
What are the timelines for the project, including any milestones or deadlines that have been set? 💎
We have set the following timeline for the project:
- January: Trial run of the data collection site, including user testing and debugging
- Mid-January: Version 1.0 release of the data collection site, including basic features such as submission and voting
We will be closely tracking our progress against these milestones and will update the community as we make progress. If you have any questions or ideas for how you can get involved, please don't hesitate to reach out to us."
Any rules or guidelines that users should be aware of, such as those related to data quality or privacy? 📏
As we develop and test the model, we will be taking a number of ethical and legal considerations into account. Some of the issues that we will be paying particular attention to include:
-
Data privacy: We will be taking steps to protect the privacy of the individuals depicted in the video clips, including by obtaining appropriate consent and anonymizing data where necessary.
-
Fairness and bias: We will be careful to ensure that the model is trained on a diverse and representative dataset and that it is free from bias. We will also be testing the model to ensure that it performs equally well across different groups of players.
-
Transparency: We will be transparent about the methods and techniques used to develop the model, and will be open to feedback and suggestions from the community.
We will be actively seeking input and guidance from experts in the fields of artificial intelligence and online gaming to help us navigate these and other issues as we move forward with the project.
Developers & Contributers 💖
The project involves the development of a data collection site where users can submit their own video clips and vote on others, with the goal of building a diverse and representative dataset for training a machine learning model.
Core Features 🧪
What features are needed for this side project:
-
Create and sign up an account.
-
Upload a video link from a youtube video and submit it to our backend database.
-
Review gameplay with either yes or no on other peoples submissions.
-
Admin Panel
Accounts
We will be using Oauth providers such as Google, Facebook & more.
- Login
- Logout
- Sign up
- View account details
- Ability to delete account
Uploading gameplay
We are using youtube for video hosting.
- Upload links
- Ability to view and delete submitted gameplay
Reviewing gameplay
A place for users to review uploaded gameplay
- Review gameplay that has never been seen
- A way for managing certain videos on a certainty of "Yes" and "No"
Admin Panel
For admins to remove, add or create moderators and other admins and gmaeplay.
- Ability to search all of users
- Search all of gameplay and the ability to remove it
- Way to stop certain services such as site maintenance, upload, reviews and sign ups.