# Quantum Neural Networks: Concepts, Applications, and Challenges

**Authors:** Yunseok Kwak, Won Joon Yun, Soyi Jung, Joongheon Kim  
**arXiv:** [2108.01468](https://arxiv.org/abs/2108.01468)  
**Date:** 2021-08-02  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2108.01468v1/

## Key points

- Quantum deep learning combines quantum computing techniques with deep neural networks
- Recent discoveries show that quantum circuits can mimic artificial neural networks
- Understanding the concept of "quantum advantage" is a significant challenge in the field
- Quantum Neural Networks (QNN) have universal approximation capabilities similar to classical models
- Researchers need to articulate how QNN-based algorithms offer advantages over classical counterparts to avoid inefficiencies

## Summary

, , , , 

Quantum deep learning is a rapidly growing research field that combines quantum computing techniques with deep neural networks. While traditionally these two realms have operated independently, recent discoveries have shown that quantum circuits can mimic artificial neural networks, leading to widespread adoption of quantum deep learning research. This paper delves into the foundational principles and background of quantum deep learning while also highlighting key achievements in the field. It also explores the challenges faced by researchers from various perspectives. One significant challenge is understanding the concept of "quantum advantage," which emphasizes that quantum algorithms are not inherently superior to classical algorithms in all scenarios. Instead, realizing the benefits of quantum computing requires well-designed algorithms under specific conditions. While Quantum Neural Networks (QNN) possess universal approximation capabilities similar to classical deep learning models, it is crucial for researchers to articulate how QNN-based algorithms offer advantages over their classical counterparts to avoid inefficiencies. In conclusion, this paper outlines the fundamental concepts of Quantum Neural Networks and their applications across diverse fields. It also addresses research challenges within quantum neural network computation and proposes potential solutions. The work was supported by the National Research Foundation of Korea, with Joongheon Kim serving as the corresponding author. Overall, this comprehensive summary provides insights into the evolving landscape of quantum deep learning research and underscores the importance of developing efficient QNN-based algorithms with clear advantages over classical models for optimal performance in various applications.