Artificial Intelligence

人工智能

 

人工智能(AI)是一套让计算单元具有在人类看来似乎是模仿人类认知能力的功能的软硬件系统。

人工智能是使用一类自然计算方法近似计算现实世界中数学或传统建模方法不能有效或准确解答的问题。人工智能采用近似人脑的推理方式,利用不准确和不完整的知识,以自适应推理方式得出动作结果,并随时间推移逐渐积累经验

意法半导体一直以来积极参与AI课题研究,并于2019年1月推出了最新产品。现在,STM32微控制器产品组合让嵌入式开发人员能够实现前所未有的开发效率。借助STM32Cube.AI扩展软件包,他们可以安装和运行经过预训练的人工神经网络(ANN),在STM32Arm®Cortex®-M的MCU上启用AI。

边缘人工智能

人工神经网络(ANN) 可以解决日常生活中出现的各种问题。在我们的环境、家里、办公室、汽车、工厂和个人物品中存在大量的传感器,ANN网络可以充分利用这些传感器产生的数据。将传感器的原始数据发送到功能强大的中央远程信息处理中心(云计算)处理是一个常见的网络模型,这种集中式处理对数据带宽和计算能力要求很高。如果考虑来自数亿个终端设备的音视频或图像文件,该模型将会降低网络响应速度。


将集中式智能系统变为分布式
AI enables much more efficient end-to-end solutions 如果把云端完成的某些分析过程下移到传感器和执行器附近,人工智能就可以实现更高效的端到端解决方案。通过利用先进的边缘计算技术,这种分布式智能方法可大幅降低对数据传输带宽和云服务器处理能力的要求。因为在把数据提供给服务商前,先对个人源数据进行预分析和相关性审核,这种模式还有保护客户数据隐私的优势。


  回到顶部

人工智能 @ ST



 


借助意法半导体新的人工智能(AI)解决方案,现在可以在多款STM32微控制器上安装和运行经过预训练的人工神经网络(ANN)。






 

 

 

 

 

 

 

 

 

 

 

 

 

LSM6DSOX (IMU)等先进传感器集成器学习内核、有限状态机(FSM)和先进数字功能,可让所连接的STM32或应用处理器能够从超低功耗状态,切换到到高性能、高精度的AI功能, 适用于电池供电的物联网、游戏、可穿戴和消费电子产品。

人工智能最新消息


 

 

ST Tools for fast AI prototyping Acoustic event recognition powered by STM32Cube.AI Getting started with the FP-AI-SENSING1

 

回到顶部
 

ST发表的人工智能专题论文

Embedding Recurrent Neural Networks in Wearable Systems for Real-Time Fall Detection. E. Torti, A. Fontanella, M. Musci, N. Blago, D. Pau, F. Leporati, M. Piastra, Microprocessors and microsystems, September 2019
Click here 点击此处

Intelligent Recognition of TCP Intrusions for Embedded Micro-controllers, Remi Varenne, Jean Michel Delorme, Emanuele Plebani, Danilo Pau, Valeria Tomaselli, International Conference on Image Analysis and Processing 2019, September 2, 2019
Click here 点击此处

A New Scalable Architecture to Accelerate Deep Convolutional Neural Networks for Low Power IoT Applications Embedded World 2018 – Speeches
Click here 点击此处

Intelligent Embedded and Real-Time ANN-based Motor Control for Multi-Rotor Unmanned Aircraft Systems, George Michael, Nectarios Efstathiou, Kyriacos Mantis, Theocharis Theocharides, Danilo Pau, Proceedings of 25th IFIP/IEEE International Conference on Very Large Scale Integration (VLSI-SoC) Abu Dhabi, UAE October 23 - 25, 2017
Click here 点击此处

Embedded Real-Time Fall Detection with Deep Learning on Wearable Devices; Euromicro DSD/SEAA 2018, August 29 – 31, 2018, Prague | Czech Republic
Click here 点击此处

Automated generation of Single Shot Detector C library from a high level Deep learning framework, 4th International Forum on Research and Technologies for Society and Industry; Palermo, Italy, September 10-13 2018
Click here 点击此处

Intelligent Cyber-Physical Systems for Industry 4.0, First IEEE International Conference on Artificial Intelligence for Industries, Sep 26, 2018 - Sep 28, 2018, Laguna Hills, CA
Click here 点击此处

STM32Cube.AI: AI productivity boosted on STM32 MCU, D Pau, M Durnerin, V D’Alto, M Castro, tinyML Summit. Advances in ultra-low power Machine Learning technologies and applications March 20-21, 2019 Sunnyvale, California

STM32 solutions for Artificial Neural Networks

Artificial intelligence (AI) is a set of hardware and software systems capable of providing computing units with capabilities that, to a human observer, seem to imitate humans’ cognitive abilities.

It uses an assembly of nature-inspired computational methods to approximate complex real-world problems where mathematical or traditional modeling have proven ineffective or inaccurate. Artificial Intelligence uses an approximation of the way the human brain reasons, using inexact and incomplete knowledge to produce actions in an adaptive way, with experience built up over time.

ST has been actively involved in AI research for many years and has applied its knowledge to develop tools that allow embedded developers to take advantage of AI techniques on ST microcontrollers and sensors. 

AI at the Edge

Artificial Neural Networks (ANNs) address a variety of problems which occur in everyday life. They can exploit the data provided by sensors present in our environments, homes, offices, cars, factories, and personal items. A widespread model assumes the raw data from sensors are sent to a powerful central remote intelligence (Cloud), thus requiring significant data bandwidth and computational capabilities. That model would lower responsiveness if you consider the processing of audio, video or image files from 100s millions of end devices.

Switching from a centralized to a distributed intelligence system

AI enables much more efficient end-to-end solutions when the analysis done in the cloud is moved closer to the sensing and actions. This distributed approach significantly reduces both the required bandwidth for data transfer and the processing capabilities of cloud servers, leveraging modern computing capabilities at the edge. It also offers user data sovereignty advantages, as personal source data is pre-analyzed and provided to service providers with a higher level of interpretation.

processing capabilities of cloud servers

 

Artificial Neural Networks on General Purpose Microcontrollers

Artificial Intelligence @ ST

Thanks to ST’s new set of Artificial Intelligence (AI) solutions, you can now map and run pre-trained Artificial Neural Networks (ANN) using the broad STM32 microcontroller portfolio.

Contact us at edge.ai@st.com to find out more on how you can run edge AI applications on STM32 microcontrollers and application processors.

Artificial Neural Networks on Automotive Microcontrollers

Artificial Intelligence @ ST

Thanks to ST’s SPC5Studio.AI component for our fully customizable SPC5Studio Eclipse development environment, you can now convert, analyze and deploy automotive neural network models on our SPC58 Chorus automotive microcontrollers.
 

Machine Learning on Sensors

LSM6DSOX

Advanced sensors, such as the LSM6DSOX (IMU), contain a machine learning core, a Finite State Machine (FSM) and advanced digital functions to provide to the attached STM32 or application central system capability to transition from ultra-low power state to high performant high accuracy AI capabilities for battery operated IoT, gaming, wearable technology and consumer electronics.

 

Latest news about Artificial Intelligence

 
     

 

ST-Published papers on Artificial Intelligence

2021

Change Detection in Multivariate Datastreams Controlling False Alarms; Luca Frittoli, Diego Carrera, Giacomo Boracchi. Proceedings of European Conference on Machine Learning (ECML) 2021
Click here

Exploiting History Data for Nonstationary Multi-armed Bandit; Gerlando Re, Fabio Chiusano, Francesco Trovò, Diego Carrera, Giacomo Boracchi, Marcello Restelli. Proceedings of European Conference on Machine Learning (ECML) 2021
Click here

Profiled Attacks against the Elliptic Curve Scalar Point Multiplication using Neural Networks;  Alessandro Barenghi, Diego Carrera, Silvia Mella, Andrea Pace, Gerardo Pelosi, Ruggero Susella. International Conference on Network and System Security (NSS) 2021.
Not yet available

"A grapevine leaves dataset for early detection and classification of Esca disease in vineyards through machine learning", M. Alessandrini, R. C. Fuentes Rivera, L. Falaschettia, D. Pau, V Tomaselli, C Turchetti; Data in Brief, Jan 2021

Click here

Deep Learning Localization with 2D Range Scanner; Giuseppe Spampinato, Arcangelo Ranieri Bruna, Ivana Guarneri, Davide Giacalone - 2021 International Conference on Automation, Robotics and Applications (ICARA 2021)
Click here

 

The full list of papers is available here