The female server technician stands next to the cabinet in the database of the RACK server series. Runs diagnostically on her computer
aging
What does it mean for the data transport engineering at AI age? More than some top professionals would say that this is even more important now than in the past. In addition, artificial intelligence works with the principle of obtaining information and processing them – in ways that are very similar to what is happening in our human brains.
Thus, the information itself is valuable and the procedures are valuable.
Also, modern developments in large linguistic models have given us different ways to consider data transfer. I will enter some of them with thoughts of three lightning conversations given to the imagination in April, in the Lightning Talks section of the wide program.
Unlock Data Siled
The first such presentation was by William Lindskog Munzing who was talking about an application called Flower, suggesting that traditionally, data has been trapped in silo.
The goal, he said, is to move the AI to the point where the data is. This is much easier in times when working on quantification and foundation models and lower bit systems has led to the Edge AI-the AI tracking capacity where the data is already, instead of transferring it to central data centers.
In this spirit, the flower community, which has drawn up about 5800 developers, with 2000 active projects, works on what Munzing calls the “Collective-1” platform.
“What we believe is that the data must remain where they come,” he said. “It’s … never transmitted. It remains on your device, in your car, whatever it is, or in hospitals. ”
The ISO project, he added, is also adaptable.
“We have done a lot of things in the development of execution time,” Munzing explained. “So it is very easy for you now to perform your experiment in CPU, GPU, and then scale to the real development of the real world if you want to add safe mechanisms, authentication and much more.”
AI for drivers
The next discussion came from Marco Celto, who worked in a project called Meshify.
He explained that the data suggest that small to medium -sized enterprises (SMEs) are on average about 50% as effective as companies and that collectively, media lose up to $ 500 billion in revenue from poor lead management only.
As a solution, Meshify, he explained, will wipe the inbox of a professional, follow and provide them with automated CRM knowledge.
I thought it was also interesting that this project is using the idea of Nanda Decentralized Network that pioneered MIT by some of my colleagues. In fact, so is the flower, which shows the initiative to create an operating tissue protocol for AI gets steam.
Regulation and control
The third presentation came from Peyush Aggarwal in Deloitte, where he talked about the dimensions of change in the AI era.
This was different because Aggarwal did not promote a specific start or product. Instead, he was talking about circles – a cycle of auxiliary increase in automation and priorities for the development of AI, such as:
- Regulatory efforts
- Treatment of data and fears to protect privacy
- Checking complex procedures
- Transparency
“Human advantages need to be strengthened,” Aggarwal said. “How do you make a banker and train them to make cognitive as well as decision making, when people are designed to think in a straight line, right?
Referring to the need to control the culture of change in the banking industry, Aggarwal went through many aspects of the analysis of AI activities from the point of view of the meeting room, citing a gradual change and evolution from classic to digital banking and then to smart banking.
At the end of his presentation, he went on an interesting philosophical path, talking about the management of AI people and agents.
“The most important part is when you bring with the people and agents of AI. Is it. Do you check AI or check the decision of AI and the people who meet?” he asked. “This is the most important aspect. What is the case of use?” When a regulator examines a case of use, he raises the question: “Can I repeat this use case? Can I repeat the question asked?” And if you can’t, you can’t really approve of their use. ”
In other words, managers who manage human agents and AI have a different role than those who only manage humans themselves.
You manage the intersection of people with technology. How does it work?
I thought these were some interesting eyes-opening at a time when we were trying to adapt to a rapidly changing goal of technology. Stay tuned.
