文件列表:

api/ (0, 2023-11-05)
api/Dockerfile (193, 2023-11-05)
api/docker-services.yml (60, 2023-11-05)
api/models/ (0, 2023-11-05)
api/models/rf_model_bad_v1.pkl (1065305, 2023-11-05)
api/models/rf_model_v1.pkl (880137, 2023-11-05)
api/models/rf_model_v2.pkl (315321, 2023-11-05)
api/models/rf_model_v3.pkl (635769, 2023-11-05)
api/models/versions.txt (30, 2023-11-05)
api/server.py (2035, 2023-11-05)
api/shap/ (0, 2023-11-05)
api/shap/explainer_bad_v1.bz2 (102475, 2023-11-05)
api/shap/explainer_v1.bz2 (97788, 2023-11-05)
api/shap/explainer_v2.bz2 (35113, 2023-11-05)
api/shap/explainer_v3.bz2 (61582, 2023-11-05)
docker-compose.dev.yml (422, 2023-11-05)
docker-compose.prod.yml (537, 2023-11-05)
lung-vitality/ (0, 2023-11-05)
lung-vitality/.eslintrc.json (40, 2023-11-05)
lung-vitality/components.json (327, 2023-11-05)
lung-vitality/dev.Dockerfile (292, 2023-11-05)
lung-vitality/docker/ (0, 2023-11-05)
lung-vitality/docker/docker-services.dev.yml (285, 2023-11-05)
lung-vitality/docker/docker-services.prod.yml (207, 2023-11-05)
lung-vitality/next.config.js (263, 2023-11-05)
lung-vitality/package-lock.json (188676, 2023-11-05)
lung-vitality/package.json (1126, 2023-11-05)
lung-vitality/postcss.config.js (82, 2023-11-05)
lung-vitality/prettier.config.js (66, 2023-11-05)
lung-vitality/prod.Dockerfile (1016, 2023-11-05)
lung-vitality/public/ (0, 2023-11-05)
lung-vitality/public/info.svg (6082, 2023-11-05)
lung-vitality/public/logo.svg (975, 2023-11-05)
lung-vitality/public/medicine.svg (46499, 2023-11-05)
lung-vitality/public/question.svg (13372, 2023-11-05)
lung-vitality/src/ (0, 2023-11-05)
lung-vitality/src/app/ (0, 2023-11-05)
... ...

# HCAID (Human Centric AI Design) ## 1. Introduction HCAID is part of our minor Artificial Intelligence. This module focus on the design of AI systems by following some guidelines on how to design a user-centric AI-based app. Because AI makes it possible to make predictions, it increases our ability to make decisions. However, the user must be central when designing a prescribed AI system. Otherwise we lose the capacity for free choice. That is why we use Design Thinking to create a solution in which humans and the AI system are complementary to each other and people retain the ability to make free choices. ## 2. Our idea Machine learning can be used to make predictions to give insight in various topics. We have developed an AI model that can predict if you are likely to have lung cancer or not. Of course this topic can impact the users if they don't know how the machine learning model works and how well the predictions are. Our good application "Lung Vitality" will need to account for: - **privacy**: if data gets collected, it will be non-identifiable and the user will get notified if they want to share this information or not. - **bias**: at what level is the application biased ? what could be the impact on the users ? - **explainability**: the application should give the user an explanation on how the predictions are made and how the system works. Our bad application tries to exploit the AI model in a bad way. The application doesn't need to follow all the rules described above: - **privacy**: the application will ask information that is not related to the machine learning model. It will also store your information that is directly related to you. - **bias**: the application doesn't care about bias, it will look at the data that is barely related to the actual prediction, like gender. - **explainability**: the application makes a prediction without explaining how that prediction is made. ## 3. Development Both websites are made with the framework NextJS. To run the website locally, you can use the 'docker-compose.dev.yml' file to launch a development environment for both applications, available on http://localhost:3001 and http://localhost:3002. To do this, you need to have Docker installed on your computer. If you don't have Docker installed, you can also run the applications locally by installing NodeJS. **Launch the development environment with Docker** Inside the root folder: 'docker compose -f docker-compose.dev.yml up -d' ## 4. Deployment Both websites are deployed on a Linux server with Docker installed. Via a webhook, the server is automatically updated when a new commit is pushed to the master branch. The websites are available on: - https://smart-lungs.qloozen.nl/ - https://lung-vitality.qloozen.nl/

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：