A Review on Neural Turing Machine (NTM)
- PDF / 2,837,443 Bytes
- 23 Pages / 595.276 x 790.866 pts Page_size
- 105 Downloads / 229 Views
SURVEY ARTICLE
A Review on Neural Turing Machine (NTM) Soroor Malekmohamadi Faradonbe1,2 · Faramarz Safi‑Esfahani1,2 · Morteza Karimian‑kelishadrokhi1,2 Received: 13 June 2019 / Accepted: 22 September 2020 © Springer Nature Singapore Pte Ltd 2020
Abstract One of the significant objectives of artificial intelligence is to design learning algorithms that are executed on general-purpose computational machines inspired by the human brain. Neural Turing Machine (NTM) is a step towards realizing such a computational machine. In the literature, a variety of approaches have been presented for the NTM; however, there is no existing comprehensive survey and taxonomy for NTM methods. This article presents an overview of taxonomies characterizing the critical concepts of the NTM through a comprehensive survey on the related research activities. This in-depth analysis of taxonomies can provide researchers, designers, and application developers with a clear guideline to compare NTM methods. The taxonomy of machine learning, neural networks, and the Turing machine is introduced. The NTM is also inspected in terms of concepts, structure, implemented tasks, and related works. The article further presents research discussions and future challenges in this area. Keywords Neural turing machine (NTM) · Deep learning · Machine learning · Turing machine · Neural networks
Introduction Artificial Intelligence (AI) seeks to construct real intelligent machines. The artificial neural networks constitute a small portion of AI; thus, the human brain should be named as a biological neural network (BNN). The brain is a very complicated nonlinear, and parallel computer [1]. The almost relatively new neural learning technology in AI named ‘Deep Learning’ that consists of multi-layer neural networks learning techniques. Deep Learning provides a computerized system to observe the abstract pattern similar to that of the human brain while providing the means to resolve cognitive problems. As to the development of deep learning, recently, there exist many useful methods for multi-layer neural network training such as Recurrent Neural Network (RNN) [2]. * Faramarz Safi‑Esfahani [email protected] Soroor Malekmohamadi Faradonbe [email protected] Morteza Karimian‑kelishadrokhi [email protected] 1
Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran
Big Data Research Center, Najafabad Branch, Islamic Azad University, Najafabad, Iran
2
What is possible in reality is not so simple in practice. Computation programs are constructed based on three fundamental mechanisms: (1) initial operations (e.g., arithmetic), (2) logical flow control, (e.g., branching), and (3) external memory, to allow reading and writing (Von Neumann 1945). Concerning the success made in complicated data modeling, machine learning usually applies logical flow control by ignoring external memory. Here, RNNs outperform other learning machine methods with a learning capability. Moreover, it is evident that RNNs are turing
Data Loading...