About My Projects
I am making an AI unparallel to any other AI in existence. It is called M.A.G.I.S.T. and will be able to think like a human! I plan on embedding this into a biomimetic robot dog! This research will push humanity into a new generation of robotic intelligence. Here are some of the topics I research.
- Application of biomimetic robotics in self-supervised AGIs.
- Development of self-supervised, multi-agent, and structured data-driven AIs.
- Developing efficient algorithms that function predominantly on edge computing.
My main project is Project Zeta and MAGIST which work in conjunction to make a revolutionary Artificial General Intelligence. I also have other projects running like DeepQ which is a more accurate IQ test that uses AI and an intelligent, dynamic question system.Learn More
MAGIST - Artificial General Intelligence
MAGIST is my masterpiece. Ever since the creation of The Perceptron (the first AI), people have dreamed of an AI that thinks like us. MAGIST can teach itself various types of information independently! It can logically interpret human language and respond adequately using previously learned information. It also uses a novel data pipeline that stores data in a logical format as opposed to many powerful AIs today like Imagen and GPT3, which use an uninterpretable latent space.
Project Zeta - The Quadruped
This biomimetic platform is loaded with 4 3DOF legs, 17kg/cm motors, 6-axis gyro, dual radio system, 2k camera, and much more. This runs ROS(Robot Operating System) and has full WiFi and Bluetooth capability. It will be the perfect host for my MAGIST algorithm.
Here are the platforms I currently operate on:
All of my code, repositories, issues, etc. are hosted on GitHub making them available to the public. I strongly encourage the community to contribute to this project. I am pushing commits every week and are constantly evolving my code to make it perform better.
All open-source projects require a strong community and crowd-funding to continue research. I chose the OpenCollective platform for its transparency to ensure that the community fully trusts the final destination of their money. I appreciate donations of all sizes.
This is where I can showcase my projects to our community and teach the knowledge I have attained through this platform. I have AI, ROS, and Python tutorials as well as videos of my robots. I will have YouTube Shorts that demonstrate the capabilities of our robots and how you can develop intelligent machines using our codebase.
I regularly write articles regarding robotics, technology, ethics, and programming to inspire the STEAM field and provide my insight and knowledge to the common public. This can help them in their projects as well. For example, there is an article series about ROS that covers everything you need to make a ROS system.
Frequently Asked Questions
Are my projects OpenSource?
Yes! All of my projects will be under the GNU GPL v3 license. Although, restrictions may apply to non-individuals/corporations.
Is MAGIST truly generally intelligent?
My technique walks the fine line between a specific AI and a general AI. MAGIST, in its final form, will function in a method similar to our brains. It will be able to independently collect data from a variety of sources and compute adequate responses based on the concurrent conversations and environmental state. Although many people believe that an AGI(Artificial General Intelligence) is strictly an end-to-end AI, that is only one way to do it. AIs like Google Imagen and OpenAI GPT3 seem generally intelligent, they cannot currently expand their database. My AI can adapt to drastically different data and relate images, audio, text, etc. to a single data instance.
What sets MAGIST apart from other AIs?
MAGIST has a unique data pipeline that can relate all types of different inputs and collected information to a single data instance. This instance can be searched on demand to provide the NLP algorithm with a plethora of information to compute an appropriate response. It also independently collects data from a variety of sources like the environment and the Internet. Finally, it is using a multi-agent approach as opposed to an end-to-end approach that is used by many other research companies.
Where is my research headed in the next 5 years?
In the next five years, I want DeepShift Labs to be a research institution improving the MAGIST algorithm and creating a stable implementation of it into biomimetic robots. MAGIST will be conversationally intelligent and add valuable, non-biased information to the ongoing conversation. It will also use a unique action controller to control Project Zeta.
Please don't hesitate to contact me on my media platforms. I will do my best to answer your question, concern, etc.