This is the web version of Eye on A.I., a weekly newsletter on the intersection of artificial intelligence and industry. Sign up to get it delivered free to your inbox. Amazon, like other tech giants, is desperately hunting for workers who have an expertise in artificial intelligence. The online retailer has many businesses--its core e-commerce division, the Alexa voice-activated digital service, and the AWS cloud computing unit--that depend on machine learning. But there are relatively few computer scientists who know the technology, and those who do are in high demand.

Machine learning is powering most of the recent advancements in AI, including computer vision, natural language processing, predictive analytics, autonomous systems, and a wide range of applications. Machine learning systems are core to enabling each of these seven patterns of AI. In order to move up the data value chain from the information level to the knowledge level, we need to apply machine learning that will enable systems to identify patterns in data and learn from those patterns to apply to new, never before seen data. Machine learning is not all of AI, but it is a big part of it. While building machine learning models is fundamental to today's narrow applications of AI, there are a variety of different ways to go about realizing the same ends.

In the video tutorial below, I take you through this process using the demo example of a handwritten digit recognition model trained on the MNIST dataset.

Machine learning in embedded systems allows the use of that data in automated business processes to make more educated predictions. Running machine learning models on embedded devices is generally known as embedded machine learning. Machine learning leverages a large amount of historic data to enable electronic systems to learn autonomously and use that knowledge for analysis, predictions, and decision making. Devices such as these can fulfill many tasks in the industry. Such devices allow machine learning algorithms on low-power devices like microcontrollers.

It's hard to find reliable sources of information when starting out in deep learning. These will save you time, effort, and headaches. There are tons of free machine learning resources out there, but the sheer volume makes it difficult to sift through them. Not all of them are reliable or well-written. When you're starting out in deep learning, a poorly written tutorial can do more harm than good.

This course will get you started in building your FIRST artificial neural network using deep learning techniques. Following my previous course on logistic regression, we take this basic building block, and build full-on non-linear neural networks right out of the gate using Python and Numpy. All the materials for this course are FREE. We extend the previous binary classification model to multiple classes using the softmax function, and we derive the very important training method called "backpropagation" using first principles. I show you how to code backpropagation in Numpy, first "the slow way", and then "the fast way" using Numpy features.

The graph represents a network of 1,418 Twitter users whose tweets in the requested range contained "#iiot", or who were replied to or mentioned in those tweets. The network was obtained from the NodeXL Graph Server on Tuesday, 01 June 2021 at 20:15 UTC. The requested start date was Tuesday, 01 June 2021 at 00:01 UTC and the maximum number of tweets (going backward in time) was 7,500. The tweets in the network were tweeted over the 2-day, 6-hour, 54-minute period from Saturday, 29 May 2021 at 17:03 UTC to Monday, 31 May 2021 at 23:58 UTC. Additional tweets that were mentioned in this data set were also collected from prior time periods.

Then Peetak Mitra devised a new way to prune networks, removing unimportant nodes and connections. Pruning reduced network size by 90%, making it 10 times faster--while simultaneously increasing accuracy. That's because large networks adapt to any information in the system, which makes them good at generalizing to many scenarios, but they can learn from the noise in the system. If you're applying machine learning to a regular environment--similar types of cylinders--you can afford to shrink the network, thereby filtering out the noise. The team also used a process called quantization, which reduces the excessive precision of the network's values.

In the race to build the underlying technologies that can power the next wave of AI revolution, a Chinese lab just toppled OpenAI, the venerated US-based research lab, in terms of who can train a gigantic deep learning model with the most training parameters--as for whether or not there is a race, at least ranking members of the lab believe so. The Beijing Academy of Artificial Intelligence, styled as BAAI and known in Chinese as 北京智源人工智能研究院, launched the latest version of Wudao 悟道, a pre-trained deep learning model that the lab dubbed as "China's first," and "the world's largest ever," with a whopping 1.75 trillion parameters. Unlike conventional deep learning models that are usually task-specific, Wudao is a multi-modal model trained to tackle both text and image, two dramatically different sets of problems. At BAAI's annual academic conference on Tuesday, the institution demonstrated Wudao performing tasks such as natural language processing, text generation, image recognition, image generation, etc. The model is capable of writing poems and couplets in the traditional Chinese styles, answer questions, write essays, generate alt text for images, and generate corresponding images from natural language description with a decent level of photorealism. It is even able to power "virtual idols", with the help of XiaoIce, a Chinese company spun off of Microsoft--so there can be voice support too, in addition to text and image.

We propose an end-to-end deep learning framework that comprehensively solves the inverse wave scattering problem across all length scales. Our framework consists of the newly introduced wide-band butterfly network coupled with a simple training procedure that dynamically injects noise during training. While our trained network provides competitive results in classical imaging regimes, most notably it also succeeds in the super-resolution regime where other comparable methods fail. This encompasses both (i) reconstruction of scatterers with sub-wavelength geometric features, and (ii) accurate imaging when two or more scatterers are separated by less than the classical diffraction limit. We demonstrate these properties are retained even in the presence of strong noise and extend to scatterers not previously seen in the training set. In addition, our network is straightforward to train requiring no restarts and has an online runtime that is an order of magnitude faster than optimization-based algorithms. We perform experiments with a variety of wave scattering mediums and we demonstrate that our proposed framework outperforms both classical inversion and competing network architectures that specialize in oscillatory wave scattering data.