![Near-sensor and In-sensor Computing](/uploads/posts/2022-10/1667000361_4481_n_ar_s_nsor_and_in_s_nsor_computing.jpg)
Автор: Yang Chai, Fuyou Liao
Издательство: Springer
Год: 2022
Страниц: 237
Язык: английский
Формат: pdf (true)
Размер: 17.1 MB
This book provides a detailed introduction to near-sensor and in-sensor computing paradigms, their working mechanisms, development trends and future directions. The authors also provide a comprehensive review of current progress in this area, analyze existing challenges in the field, and offer possible solutions. Readers will benefit from the discussion of computing approaches that intervene in the vicinity of or inside sensory networks to help process data more efficiently, decreasing power consumption and reducing the transfer of redundant data between sensing and processing units.
In recent years, Deep Learning (DL) technology has made great progress, especially in image and video recognition and natural language processing (NLP). Unlike previous scientific computing tasks involving a small amount of data, the artificial neural network (ANN) algorithm involves heavy vector matrix multiplication, requiring a large amount of data movement in the traditional von Neumann hardware with separation of storage and computing, thus limiting computational throughput and energy efficiency. In order to solve this problem, a direct technical route is to break the boundary between computing and storage and perform VMM operations in the storage array. In order to solve this problem, a direct technical route is to break the boundary between computing and storage and perform VMM operations in the memory array. Memristor is a new principle nano-device with resistive switching behavior, which was realized in experiment by HP in 2008. Memristor has the characteristics of fast operation speed, low power consumption, scalability, good reliability, simple manufacturing process, and so on, which has been widely favored. Its emergence provides a new course to develop non-von Neumann computing architecture with strong computing capability and high energy efficiency.
Neuromorphic computing dreams of human-level artificial general intelligence by emulating the brain, which contrasts the pervasive von Neumann computing architecture. Up to now, artificial synapses have been long and widely adapted to function mostly as signal transmission with a memory effect in neuromorphic computing; significant efforts have been made to mimic mostly the memory function (e.g., memristors). However, emulating synaptic computation, which is vital for information processing, working memory, and decision-making by using short-term plasticity (STP), remains technically challenging to be demonstrated without using numerous CMOS devices. In the human brain, a quadrillion synapses are present in a massively parallel architecture, which highlights the issue of integration of CMOS devices for an efficient neuromorphic chip.
Provides readers with a detailed introduction to the near-sensor and in-sensor computing paradigms;
Includes in-depth and comprehensive summaries of the state-of-the-art development in this field;
Discusses and compares various neuromorphic sensors and neural networks:
Describes integration technology for near-/in-sensor computing;
Reveals the relationship between near-/in-sensor computing and other computing paradigms, such as neuromorphic computing, edge computing, intuitive computing, and in-memory computing.
Contents:
1. Neuromorphic Computing Based on Memristor Dynamics
2. Short-Term Plasticity in 2D Materials for Neuromorphic Computing
3. Bioinspired In-Sensor Computing Devices for Visual Adaptation
4. Neuromorphic Vision Based on van der Waals Heterostructure Materials
5. Neuromorphic Vision Chip
6. Collision Avoidance Systems and Emerging Bio-inspired Sensors for Autonomous Vehicles
7. Emerging Devices for Sensing-Memory-Computing Applications
8. Neural Computing with Photonic Media
9. Multimodal Sensory Computing
Скачать Near-sensor and In-sensor Computing
![](/templates/Mirlib/images/MirLib.png)
[related-news] [/related-news]
Комментарии 0
Комментариев пока нет. Стань первым!