Deep Learning
Deep learning is a machine learning technique that uses layered artificial neural networks inspired by the human brain.
Deep learning is a machine learning technique that learns by building an artificial neural network modeled after the neuron connections in the human brain into multiple layers. As you go through the layers, your understanding deepens from simple features to complex concepts, such as recognizing lines and edges in a photo first and combining them to recognize eyes, nose, and faces.
Unlike the existing method where people design features one by one, its strength is that it can find features on its own if there is enough data. It received attention for its overwhelming performance in an image recognition competition in 2012, and has since become the technological foundation for the entire current AI boom, including voice recognition, translation, and generative AI.
Instead, it requires massive amounts of data and high-performance hardware such as GPUs, and there is a black box problem that makes it difficult for humans to interpret the internal decision process. It's also a common misconception that just because something is modeled after the brain, it doesn't mean it works like the brain.
✅ Why it matters
- It is the core technology that created the current AI boom and is the foundation of generative AI
- Reduces the design burden on humans by discovering features in data on its own
- Has versatility across fields such as image, voice, and language.
⚠️ Limits and debates
- It requires large amounts of data and expensive computing resources
- There is a black box problem that makes it difficult to interpret the basis for judgment
- It is a structure modeled after the brain and does not think like a real brain.