Yann LeCun, owner of the main prize VinFuture 2024, said many people are afraid of AI because they think it will be smarter than humans, but “we still have a long way to go before we reach that level.”
Coming to Vietnam in the early days of December to participate in a number of talks on artificial intelligence (AI) and receive the VinFuture award, Professor Yann LeCun, 64 years old, Vice President, Director of AI Science at Meta, made many announcements. message about acceptance of AI. He believes that the view that the existence of AI creates existential dangers for humanity is not correct. “AI is actually a repository of knowledge for all humanity, so it needs to be built using open platforms. This approach also helps countries ensure autonomy in AI,” he said.
As a pioneer in the field, LeCun’s contributions have been instrumental in shaping modern AI. He is honored as the “godfather” of AI, not only helping to revolutionize the way machines learn and interpret data but also changing the way people interact with technology every day.
The VinFuture 2024 Award Council recognized him for his great contributions to the field of deep learning and convolutional neural networks (CNN). CNN is the foundation of many products and services deployed by global technology giants, such as Facebook, Google, Microsoft, Baidu, AT&T… and can be used by billions of people around the world. everyday. CNNs have now become the standard in artificial intelligence applications that billions of people use every day, playing a central role in the development of technologies such as facial recognition and medical image processing.
Professor Yann LeCun spoke when receiving the award. Video: VinFuture
LeCun was born into an intellectual family in Soisy-sous-Montmorency, a suburb of Paris, France, in 1960. From a young age, LeCun was deeply interested in science and technology. His father was an engineer and his mother was a teacher, both of whom encouraged his curiosity and passion for learning. His early exposure to the world of science and technology helped shape his career and contributions to AI.
LeCun attended Lycée Louis-le-Grand, a prestigious high school in Paris, excelling in math and physics. His good academic performance helped him enroll at École Superieure d’Ingénieurs en Électronique et Électrotechnique (ESIEE), the leading engineering school in France. He studied electrical engineering and computer science, receiving his engineering degree in 1983.
After graduating from ESIEE, LeCun continued his doctoral studies at the Université Pierre et Marie Curie, now the Sorbonne University. His doctoral research focused on machine learning, specifically the development of convolutional neural networks (CNN), a type of deep learning model now widely used in image and speech recognition tasks. He received his doctorate in computer science in 1987.
Yann LeCun joined AT&T Bell Labs in 1988, after doing research with Geoffrey Hinton, another neural network pioneer and winner of the 2024 Nobel Prize in Physics. LeCun worked to apply his research to neural networks. Convolutional logic on the problem of optical character recognition.
In 1989, the United States Postal Service (USPS) provided LeCun’s team with a set of 9,298 handwritten postal codes scanned from parcels at a sorting office in Buffalo, New York. The team used 7,291 of these scans to train the convolutional neural network and the remaining 2,007 scans to test its effectiveness. As a result, the success rate was up to 95%, paving the way for widespread application of this system in the early 1990s. The system was recognized by Guinness World Records as the first neural network that could recognize handwritten character form.
A groundbreaking development in the field of machine learning occurred in 1998, when the LeCun team announced the LeNet-5 convolutional neural network. LeNet-5 consists of 7 layers, including two convolutional layers interleaved with subsampling layers, followed by a fully connected layer and an output layer. The convolutional layer is designed to automatically and adaptively learn spatial hierarchies, an important aspect of human visual perception. Subsampling layers reduce the order of the input, making the network less sensitive to small changes and distortions. The fully connected layer then aggregates these features to make the final prediction.
The success of LeNet-5 in handwriting and character recognition has led to the application of convolutional neural networks for a variety of tasks. In 2012, the AlexNet convolutional neural network outperformed all previous methods in the ImageNet Large-Scale Image Recognition Challenge. This event was the turning point that led to the widespread adoption of convolutional neural networks for image-related tasks.
Professor LeCun on the day of receiving the VinFuture 2024 award in Hanoi. Image: Ngoc Thanh
LeCun, along with Yoshua Bengio and Geoffrey Hinton, was awarded the 2018 Turing Award for pioneering research in the fields of deep learning and neural networks. The Turing Award is often compared to the Nobel Prize in the field of computer science, awarded annually by the Association for Computing Machinery (ACM). The Turing Award is a worthy recognition of LeCun’s achievements, and it is likely that his research will continue to shape AI for years to come.
In addition to the prestigious Turing Award, LeCun is also a Silver Professor of the Courant Institute of Mathematical Sciences at New York University, and Meta’s AI Scientific Director. He is also a member of the US National Academy of Engineering, received the IEEE Neural Network Pioneer Award in 2014 and the PAMI Outstanding Researcher Award in 2015.
While Sam Altman, Elon Musk and many technology leaders draw the prospect of an “AI apocalypse”, LeCun thinks this warning is extremely ridiculous. “Will AI rule the world? No, it’s just a projection of human nature onto machines. One day, computers will be smarter than humans, but it will be a long time before we reach that state.” there”, BBC quoted LeCun in June 2023.
During his visit to Vietnam earlier this month, LeCun shared about the application of AI technology in scientific and technological research. “In general scientific research, AI technology has been applied a lot in science and medicine. In medicine, AI is applied in discovering new drugs or understanding the mechanisms of life. In science, AI can be applied to find new materials. This is very important in responding to climate change as well as promoting progress in physics and chemistry. , biology…”, he said.