ARGUS is a personal AI assistant built using Python and TensorFlow. It is designed to perform various tasks such as recognizing wake words (ARGUS), interacting through natural language conversations, performing web searches, opening and closing applications, retrieving news, and more. ARGUS also incorporates a suite of advanced programming features. Its architecture includes a dynamic reward system that continuously evaluates responses based on relevance, intent matching, and sentiment alignment, while a genetic algorithm employs mutation and crossover techniques to refine and enhance candidate responses. Additionally, ARGUS features a robust sequence-to-sequence conversational model with attention mechanisms, real time audio processing and visualization via a Tkinter GUI, comprehensive logging and data storage for conversation history, and seamless integration with custom scripts and external APIs for tasks like news retrieval, note-taking, and even password breach checking. These components work in harmony to ensure that ARGUS not only responds intelligently but also learns and adapts over time, continually improving its conversational abilities.
ARGUS - Advanced Response and Guidance User System
Its voice activated when you say the wake word "ARGUS"
The program runs on MacOS and only has been tested on MacOS but can be easily adjusted to run on Windows.
Wake Word Detection: ARGUS listens for a specific wake word ("Argus") and activates upon detecting it.
Conversational AI: ARGUS can have natural language conversations with users and respond to queries.
Web Search: ARGUS can perform Google searches based on user queries.
Wikipedia Search: ARGUS can search Wikipedia for information about people or specific topics.
News Updates: ARGUS can fetch and read the latest news headlines from BBC.
Application Management: ARGUS can open and close applications on your computer.
Notes Taking: ARGUS can take notes and save them to a file.
Jokes: ARGUS can tell jokes to the user.
Custom Tools: ARGUS can run custom scripts like Spider Crawler and Password Checker.
Dynamic Reward System: ARGUS evaluates the quality of its responses based on relevance, intent matching, and sentiment analysis, adjusting its behavior accordingly.
Manual Model Adjustment: The user can manually adjust the chatbot's model, train it with new feedback data, and save the model for future use.
Code: ARGUS can code for you. Just tell it what you want it to create and it will to the best of its ability using its trained code model. - https://huggingface.co/bjw333/macosargus-code this is the code model
And more...
To use ARGUS you have to clone the whole repo and then install and change anything required to run ARGUSMAIN.py these changes most likely being pip installs and file path changes.
Then just run python3.10 ARGUSMAIN.py in your terminal.
Object Recognition System this system will give ARGUS the ability to in real time see objects and recognize them for what they are.
Feel free to fork this repository and submit pull requests. Any contributions to improve ARGUS are welcome!
This project is licensed under the MIT License.
This project is still a work in progress!