> For the complete documentation index, see [llms.txt](https://aaai.gitbook.io/aaai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://aaai.gitbook.io/aaai/how-aaai-works.md). # How AAAI works * **Live Data Streaming:** Continuously captures data from CEX's to provide real-time insights. * **Anomaly Detection:** Utilizes[ Bayesian Anomaly Detection](https://www.kaggle.com/code/mikepetrouleas/anomaly-detection-with-bayesian-neural-networks) and [Z-Scores](https://www.kaggle.com/discussions/general/414467) to detect and classify anomalies. * **Real-time Alerts:** Posts instant anomaly detection notifications to traders and stakeholders initially through , and later through a $AAAI token holder exclusive community. * **Adaptive Learning:** AAAI uses [Deep Reinforcement Learning ](https://www.kaggle.com/code/alexisbcook/deep-reinforcement-learning)with a reward function that combines anomaly prediction accuracy with feedback and social media engagement to reinforce behaviour. * **User-Centric Design:** Designed with user feedback in mind, ensuring intuitive interaction and usability.