IBM and the grand challenges of AI and quantum computing

IBM and the grand challenges of AI and quantum computing

AI is back in the news

OpenAI’s ChatGPT and image-generating AI systems like MidJourney and Stable Diffusion got a lot more people interested in and talked about advanced AI. Which is a good thing. It won’t be pretty if the transformative changes that will occur over the next two or three decades surprise most of us.

IBM has been a pioneer in advanced AI for longer than most. One of IBM’s top executives, Alessandro Curioni, joined the London Futurists Podcast to discuss IBM’s current plans in AI, quantum computing and related fields. Alessandro has worked at IBM for 25 years. He is an IBM Fellow, Director of IBM Research and Vice President for Europe and Africa.

IBM’s big challenges

IBM has been inventing the future of computing for 70 years. He staged a series of impressive Grand Challenges, like Deep Blue beating Gary Kasparov in chess in 1996, and Watson beating Ken Jennings in the Jeopardy TV quiz in 2011. Most recently, in 2018, the company developed a machine that could hold its own. . in debates with a world champion debater.

The earliest of these machines were rules-based, and the latest use deep learning, which creates trained models on large amounts of data. Another paradigm shift is happening now, with the arrival of large language models (LLMs), or base models, which use a technique called self-supervision to do training. The system will take a large amount of phrases – hundreds of billions of them – from the web, randomly hide one word from each phrase, and try to guess what the word is. Over time, the system builds a model of which words go into which sentences. This automation of the training process is a significant step forward and has been made possible by the enormous amounts of data and computing power available today.

It turns out that this methodology is not limited to text. It can be used on any type of structured data, including images, videos, or computer code. Or data flows generated by industrial processes. Or the language of science: molecules translated into symbols.

Closer focus

IBM builds great language models, but for particular applications rather than general use, like ChatGPT. For example, it builds systems to specialize in organic chemistry and business. The weakness of general purpose systems is that they are superficial. They can answer most questions at a high level, but if you go any further they get lost. More specialized machines can go further and are less fragile. Being specialized often means you can get better quality data and can eliminate bias more easily.

One of the reasons why ChatGPT performs better than GPT-3 is reinforcement learning with human feedback (RLHF). OpenAI, which created these systems, hired a large number of people to comment on the system’s output and label biased or offensive passages accordingly. This prompts that AI is not synonymous with artificial intelligence, but with affordable Indians, but humans are used during training, not in operation.

IBM hopes to prove that it can develop a large model in a particular area, which can then be trained on proprietary data from client organizations in that area. This would be a major cost and durability improvement over the old approach, which involved developing a new model for each application.

More efficient chip designs

Another area where IBM is looking to improve the efficiency and sustainability of AI and computing is in chip design. Large language models approximate the scale of computation that takes place inside the human brain, but they use the same energy as a small town, while the brain uses the same energy as a light bulb.

Curioni says IBM is taking three steps to reduce the power demands of advanced AI systems. The first step involves neuromorphic chips, such as IBM’s True North and Loihi, which are more closely modeled on human neurons than traditional chips. Their calculations are less precise, and more analog.

The second stage involves memristors, where processing and memory storage take place on the same chip, reducing the power spent retrieving and restoring data between calculations.

The third step is to activate neural networks, which transmit information only when their particular function is required, whereas in traditional chips each neuron transmits information all the time.

Together, these three steps can confer two to four orders of magnitude of energy efficiency improvements.

Breakthrough in quantum computing

IBM may not currently be seen as the world leader in AI, but one area where it is generally acknowledged to be at the forefront is quantum computing, alongside Google and Microsoft. He just announced a breakthrough in quantum cryptography that will allow data transmitted today to remain secure, even when quantum computers are built that can break today’s encryptions. Quantum computers running Shor’s algorithm can factor numbers efficiently, and when they evolve, they will be able to factor very large numbers that classical machines cannot do in reasonable time.

What IBM and a number of university partners have done is develop a new type of encryption called quantum safe crypto. It is based on high-dimensional lattice cryptography and it is believed that it cannot be broken by quantum computers. Over the past decade, an extensive research program has been conducted to evaluate many potential types of secure quantum cryptography, and last July four algorithms emerged as the most powerful. Three of these four were developed in Curioni’s laboratory in Zurich, and the winner has just been selected.

The next step is to migrate data from the old forms of encryption to this new form. This task is becoming urgent. There was a scare in December 2022 when a team of Chinese researchers announced that they had already found a way to breach today’s encryption technologies. Their paper has been dubbed the “quantum apocalypse” paper. It was soon realized that they weren’t there all the way, but it might not be long before someone gets there – maybe as early as two or three years. The US government has ordered all of its agencies to be quantum safe by 2025, and other governments and companies are doing the same. IBM’s breakthrough may have come just in time.

Similar Posts