September 2023 AMA Transcript — Matrix AI Network

Regarding biometric brain waves. For example, I have done the necessary procedures and saved it for the wallet. I want to open my wallet, but that day I am sick, I have a headache, migraine …. etc. Will my biometric data not cause any problems during the process. Was this possibility considered when taking biometric data?

This is a very good question and, at the same time, an important technical challenge that we urgently need to address. Due to the instability of brainwaves, there is significant variation in users’ brainwave data in different states. Currently, our primary targeted solutions are as follows:

1. Increase daily data collection from users to cover brainwave data in different states as comprehensively as possible.
2. Optimize our brainwave data algorithms to enhance their accuracy and resistance to interference.
3. Incorporate user guidance processes in the product, such as asking suggestive questions or playing specific music and videos to guide users into a specific state before they begin using the device.

We are continuously exploring and researching in this area, hoping that our research will consistently improve the reliability and user-friendliness of our product.

The idea of creating digital avatars from brain wave data is groundbreaking. Could you explain the technical process that allows MATRIX to accurately translate complex brain signals into usable digital representations? As the AvI project advances through the phases of MATRIX 3.0, how will the digital avatars evolve to provide a seamless and immersive experience within the Metaverse?

Our distinctive feature is the increased integration of artificial intelligence (AI) technology in the analysis of brainwave data. Artificial intelligence and neuroscience are two fields closely related, where AI can provide powerful tools for neuroscience research and holds significant importance in interpreting brainwave data.

Firstly, AI technology can identify specific signals in EEG (Electroencephalogram) data, such as attention, excitement, or depression, through EEG signal analysis. AI techniques, based on machine learning algorithms, can automatically analyze brainwave data and detect abnormal activities within EEG signals. For instance, algorithms like Support Vector Machines (SVM) can classify brainwave data, enabling automatic detection and recognition of different types of brainwave signals.

Secondly, AI technology can recognize certain features within brainwave data through EEG signal classification. Brainwave signals are complex nonlinear signals, and manually extracting and classifying their features can be challenging. AI technology, using deep learning algorithms, can automatically learn and extract features from EEG signals and classify them. For example, Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) algorithms can automatically classify brainwave data, identifying different types of waveforms like alpha waves, beta waves, and so on, thereby aiding in better understanding the actual information conveyed by brainwave data.

Regarding the second question, our current goal is not to provide users with a seamless metaverse experience. Instead, we aim for AvI to become a vital assistant when living in the Web3 and metaverse worlds. It will provide us with data acquisition and processing capabilities that align with our virtual world presence. AvI will enhance our decision-making abilities and allow us to work and live in the virtual world even when we are offline, ultimately improving our efficiency and quality of life within the virtual world.

The Matrix Bio-Wallet has gained significant media coverage. Could you share some highlights or key takeaways from its featured articles in outlets like Bloomberg, Yahoo, and MarketWatch?Can you provide insights into the partnerships or collaborations that have played a role in the development and promotion of the Matrix Bio-Wallet?

The media primarily summarized some of the product features of our Bio Wallet, including but not limited to the following points:

1. Protection against cyber hacks
2. Corporate risks
3. Private key
4. Asset Privacy
5. Offline Signing
6. Virus Immunity
7. Counterfeit Protection

Regarding your second question, we have invited our partner, Jackson from Neura Matrix, to give a presentation in the past. Similar activities and presentations will continue to be organized in the future.

Brain wave and finger vein technology integration for security is intriguing. Can you discuss the intricacies of the algorithms and hardware required to effectively combine these two biometric authentication methods?

Currently, our most important goal for version 3.0 is Avatar Intelligence, or AvI. This is a new form of intelligence we have created. We aim to collect users’ brainwave data through neuroscience while combining it with AI large models to create an AI counterpart with similar thinking patterns and values to the user. The purpose is to enhance users’ efficiency in the virtual world. Given that the virtual world is inevitably a vast ocean of data, our physical bodies are unable to possess the data acquisition and processing capabilities necessary to match the virtual world. In other words, at this stage, humans do not have the qualifications to truly live in the virtual world. However, with AvI, this problem will be effectively resolved.

Looking ahead, what is the most exciting or ambitious project or development on Matrix AI Network’s roadmap that you’re eager to share with the community?

The most crucial aspect at present is AvI within our 3.0 milestone. This technology and product are our current primary focus, and it will bring unprecedented convenience to all users in their future lives in the virtual world.

For Stage Three, creating a digital asset authentication and trading platform is intriguing. Can you provide insights into the security measures and mechanisms that will be implemented to ensure the integrity and trustworthiness of digital assets on this platform? 2.The concept of Avatar Intelligence (AvI) blending neuroscience, AI, and blockchain for the Metaverse is unique. Could you share more details about the technological hurdles you anticipate in this ambitious journey and how Matrix AI Network plans to overcome them?

Our primary approach involves using blockchain technology to establish ownership rights over data assets. Subsequently, we ensure the secure storage of data assets through distributed storage technology. Finally, we employ a combination of privacy measures and distributed computing to separate user data’s usage rights from ownership, thus enabling user data to be both valuable and invulnerable.

The primary technical challenge here remains in the acquisition and analysis of brainwave data. We currently aim to address this issue with artificial intelligence technology. On one hand, we use AI techniques to enhance the efficiency of data acquisition and the accuracy of analysis. On the other hand, we leverage AI large models to overcome the current limitations in the dimensionality of brainwave data analysis.

What are the key features and improvements in the latest version of Morpheus compared to its previous iterations?

1. The neural network parameters of the entire large model have been increased by approximately 15 times compared to the previous version.
2. The training corpus is about 35 times larger than the previous version.
3. Due to the increase in parameters, the response speed will significantly decrease, even with network optimizations and increased computational power, resulting in response times approximately 4–5 times longer than before.
4. We have added a new corpus, which is effective until the first half of 2022, allowing for responses and replies related to events up to that time.
5. In the new version of pre-training, a self-regressive mechanism has been introduced.
6. A multi-objective pre-training mechanism has been introduced in the new version of pre-training.
7. In terms of model structure, a single linear layer is used to predict the output tokens.
8. The activation function has been changed from ReLU to GeLUS.
9. Rotational position encoding has been introduced to improve stability.
10. Additionally, we have made some adjustments to the user interaction UI.

We will be publishing a completely new article to provide a detailed explanation of this upgrade.

In mythology, Morpheus is associated with dreams and sleep. How does this concept of Morpheus tie into any relevant aspects of Matrix AI’s technology or philosophy?

Certainly! The name “Morpheus” undoubtedly evokes memories of the iconic character from the movie “The Matrix”. In the film, Morpheus is a visionary leader, guiding the protagonist, Neo, through the intricacies of the Matrix, helping him realize his potential and the reality around him. This character stands as a beacon of knowledge, wisdom, and resilience.

Drawing inspiration from this, the Matrix AI Network introduced “Morpheus” as a part of their Matrix 3.0 blueprint. The goal was to create a high-precision bilingual model (Chinese/English) that would be accessible to users via the Matrix Mainnet. The team believed that such a model would be invaluable in the burgeoning Metaverse, especially given the scarcity of high-quality models available to the public at the time.

However, the journey to bring Morpheus to life wasn’t without its challenges. From computational constraints to the need for refined pretraining algorithms and rapid inference methods, the team faced numerous hurdles. They collaborated with Tsinghua University and, after intense deliberation, opted for a GLM model with 500–1000 billion parameters. This choice ensured the model’s precision while being hardware-friendly.

Despite encountering technical difficulties, such as hardware failures and algorithmic challenges, the team persevered. With support from Tsinghua University and high-performance nodes on the Matrix Mainnet, Morpheus was trained on over 100 trillion text identifiers, achieving impressive performance benchmarks.

It’s essential to understand that Morpheus wasn’t developed as a rival to ChatGPT. Instead, it stands as a testament to MATRIX’s capabilities to foster and support the creation of complex systems within its ecosystem. Just as Morpheus in the movie guided Neo, the Matrix AI Network’s Morpheus aims to guide users through the vast digital realm of the Metaverse.