Five scientists have received the $3 million VinFuture 2024 Prize for their research on “Deep Learning.” So, what is “Deep Learning” that has led these scientists to win this prestigious award?
At the VinFuture 2024 award ceremony held on the evening of December 6 in Hanoi, the VinFuture Grand Prize valued at $3 million was awarded to five scientists, including Professor Yoshua Bengio and Professor Geoffrey E. Hinton from Canada; Mr. Jen-Hsun Huang, Professor Yann LeCun, and Professor Fei-Fei Li from the United States.
These scientists earned the prestigious award due to their groundbreaking contributions to advancing Deep Learning.
Scientists receiving the VinFuture Grand Prize 2024 for their groundbreaking contributions to advancing Deep Learning (Photo: Manh Quan).
So, what is “Deep Learning” that has helped these scientists win the VinFuture Grand Prize 2024?
In reality, “Deep Learning” is not a new concept, and it has been frequently mentioned recently, especially as the race to develop artificial intelligence (AI) intensifies.
However, not everyone fully understands the concept and its applications in reality.
What is “Deep Learning”?
Deep Learning is a branch of Machine Learning and Artificial Intelligence (AI). Deep Learning focuses on teaching computers to learn and self-improve their abilities to perform tasks through an artificial neural network that simulates how the human brain works.
The highlight of deep learning compared to traditional machine learning methods is its ability to automatically extract information from data without requiring manual programming intervention.
Deep Learning enables computer systems to make decisions autonomously based on learned data (Illustration: Pinterest).
Simply put, you can think of deep learning as teaching a child to recognize the world around them. For example, you guide a child on how to recognize a cat.
Initially, when the child sees a picture of a cat, their brain gradually learns the features of a cat, such as pointy ears, whiskers, a long tail, and four legs… Each time the child sees a new cat, they will automatically recognize, “This is a cat” based on the features they have learned.
Deep Learning works similarly. It is a method of teaching computers to “learn” from many examples, much like the human brain. The computer will automatically identify important features from the data (like the ears, whiskers, and tail of a cat) through multiple layers of processing (which is why it is called “deep learning”) and use these features to recognize new things.
Illustration of a deep learning system recognizing features of cats to identify them (Illustration: AI).
For instance, if you provide a deep learning model with 1,000 pictures of cats, it will learn the characteristics of this species. Then, when you show it a new picture that the deep learning system has never seen, it can automatically recognize “This is a cat!” based on what it has learned, just like a child would.
The main difference of deep learning compared to previous methods is that instead of humans having to point out each specific feature (such as inputting commands like “find pointy ears,” “find whiskers,”…), the computer is free to explore and filter out which features are important. This allows it to perform complex tasks that are difficult for humans to describe with specific rules.
“Deep Learning” is also like the learning process of a student, where the more examples the student sees, the better they learn, deriving methods to solve problems without needing teachers to instruct every step. This is why AI systems often require vast amounts of data to learn.
The Development History of “Deep Learning”
The origins of deep learning date back to the 1940s when two American scientists, Warren McCulloch and Walter Pitts, created the first mathematical neuron model.
Professor Geoffrey Everest Hinton is regarded as one of the “fathers of AI” (Illustration: LinkedIn).
However, it wasn’t until the 1980s, with the advent of the backpropagation algorithm invented by Canadian computer scientist Geoffrey Everest Hinton, that multi-layer mathematical neural networks became truly feasible and effective. Geoffrey Everest Hinton is also one of the five scientists who recently received the special VinFuture 2024 award.
The real explosion of deep learning began in the early 2010s, thanks to three main factors: a dramatic increase in computing power of computer hardware, the massive amount of available data, and significant improvements in neural network architecture.
Practical Applications of “Deep Learning”
Today, deep learning has been practically applied across many fields of life.
In natural language processing, AI-integrated chatbots such as ChatGPT, Gemini, and Claude AI have made remarkable advancements in translation, text summarization, answering questions, and even artistic creation. These AI chatbots are becoming increasingly intelligent due to their ability to understand context and interact naturally with humans.
In the field of computer vision, deep learning has revolutionized applications for facial recognition, object detection, and more. Intelligent surveillance systems can detect abnormal behavior, while photo editing applications can automatically enhance or create entirely new images.
Deep Learning helps doctors diagnose diseases more accurately based on clinical signs (Illustration: Getty).
In healthcare, deep learning is assisting doctors in diagnosing diseases earlier and more accurately by analyzing X-ray, CT, and MRI images. Deep learning models also support the development of new drugs and predicting protein structures.
In industrial manufacturing, deep learning is applied in automated quality control and optimizing production processes. AI-equipped robots can perform complex tasks that require high flexibility and adaptability thanks to deep learning.
Another prominent application of deep learning is in voice recognition. This technology has transformed how humans interact with machines, from virtual assistants to tools that aid individuals with disabilities.
Additionally, deep learning is used for big data analysis in areas such as finance, fraud detection, and e-commerce, where it recommends products based on user behavior.
However, the rapid development of deep learning also presents numerous challenges. One such challenge is the ethical and privacy issues that arise when personal data is used to train models. Furthermore, deep learning consumes a significant amount of energy and computational resources, raising concerns about sustainability.
More importantly, understanding and controlling complex deep learning models remains a significant challenge, especially since they can make decisions that are difficult for humans to interpret.
In the future, deep learning promises to continue evolving, opening up new opportunities in fields such as medicine, energy, and education. However, to maximize the potential of this technology, there needs to be close collaboration among researchers, governments, and businesses to establish ethical standards and social responsibility.
Deep learning is not just a technological tool but also a driving force for advancing humanity if properly guided and managed.
