Deep Learning & Neural Networks

Deep Learning & Neural Networks: Technology Overview

Deep learning & neural networks is a subfield of machine learning. Deep learning leverages complex structures called artificial neural networks, loosely modelled after the human brain. Information flows through numerous layers of interconnected neurons, where it is processed and evaluated. Each layer refines the information, connecting and weighting it through nodes. Essentially, AI learns by continuously reassessing its knowledge, forming new connections, and prioritizing information based on new data it encounters. The term deep learning refers to the vast number of layers these networks can utilize. Deep learning-powered AI has achieved remarkable advancements, especially in areas like image and speech recognition. Key trends include the growing scale and sophistication of models (e.g., transformer architectures), the rise of multimodal learning that combines text, images, and other data types, and increasing efficiency through techniques like model pruning, quantization, and edge deployment. There’s also expanding use across fields such as autonomous systems, medical diagnostics, and generative AI.

Global Development of world class patents in deep learning & neural networks

The patent analysis shows that the number of world-class patents in deep learning & neural networks has skyrocketed from only 1,900 in 2015 to more than 77,500 in 2024.

Top 25 Countries

The distribution of world class patents shows that China and the US are far ahead of all other countries in deep learning & neural networks research.

Top 25 Companies

Large Chinese and US tech giants, such as Tencent, Baidu, Amazon and Alphabet, have developed the most world-class patents in deep learning and neural networks. Siemens is the first European company in the global ranking, but only ranks 25th.

Key connected technologies

Advanced deep learning and neural network architectures are the foundation for Generative AI models like Generative Adversarial Networks (GANs) and GPT models. Therefore, there is a very large overlap between deep learning & neural networks and generative AI world class patents. Other important connected technologies include autonomous driving, face recognition, robotics and smart factory.

Key connected technologies for deep learning & neural networks – country activities

The following chart shows which countries are active in key connected technologies for deep learning & neural networks. Only world class patents are shown that are assigned to both deep learning & neural networks and enabling technologies / adjacent fields. Individual countries can be activated or deactivated by clicking on the corresponding circle in the legend.

Reading example: the number next to the listed technologies indicates the number of world-class patents in 2024. China is the leader in world class patents that are assigned both to Deep Learning & Neural Networks and Generative AI. Patent applications involving researchers from more than one country are assigned to the countries of those researchers. Consequently, the sum of top research locations exceeds 100%.