# To Do List ### Chrome Extension * [ ] Finish the Chrome Extension page on this site (info about the features), right now it's just copied from the welcome page tab. * [ ] The popup should render the progress bar if closed and reopened during analysis processing. * [ ] Switch to UnBIAS ternary classification instead of binary classification. * [ ] Add multi-modal image classification feature for TruBias release * [ ] Add "Explain" button to the cards in the Explore Tab. It should use [Mbias](https://huggingface.co/newsmediabias/MBIAS) eventually, but we can start with OpenAI API because it's more cost effective. * [ ] Add a "Debias" option (uses Mbias to retain maximum context, while reducing bias). Could probabily use OpenAI for this until Mbias is hosted too. * [ ] Make a really nice media: demo with Arcade (with pan and zoom), Chrome Store images and banners, etc. ### PyPI Package * [ ] Add a multimodal pipeline for TruBias. * [ ] Add `ner="unbias"` to pipeline args. This would use UnBIAS NER instead of `"gus"` (GUS-Net) so that it would return the biased words instead of the generalizations, unfairness, and stereotype entities in the "ner" dictionary value. * [ ] Add individual model modules (allow people to call specific ones instead of all through the main pipeline. ### Hosted Endpoints * [ ] Write page about the endpoints and how people can use them. * [ ] Need to find someone to cover the costs, using AWS right now, and it's only $50/m per model (BERT models, Mbias would be more). I can add scaling with replicas, and let it scale to 0, so it's efficiently managed. ### Docs (This Site) * [ ] Ask Dr. Raza to add some info on models/datasets she knows. * [ ] Add dataset principles to the datasets tab ([FAIR](https://www.go-fair.org/fair-principles/) principles). * [ ] Write the LLMs section: De-biasing (Mbias), explanation (Mbias), classification LoRA? * [ ] Add the TruBias paper and details when it's ready (Dr. Raza and Maximus working on it). * [ ] Add sections in "Research" for Hate Speech detection and political bias detection. ### Discord * [ ] We should add some channels for sharing active research projects. ### GitHub * [ ] Add pretty readme files for all the subdirs. * [ ] Add a ipynb for using the-faily-project package to the resources --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://ethical-spectacle-research.gitbook.io/rumormill/join/to-do-list.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.