As the risks of deep learning's continued evolution have received greater attention, a growing refrain has focused on how best to prevent AI from being used for harm. From killer robots to prevalent facial recognition, societies are increasingly talking about the need for new legislation and corporate responsibility pledges to halt the spread of harmful AI. Unfortunately, the reality is that deep learning's ease of use and decentralized development across the world means it is simply impossible to constrain how it is used. Instead, societies must focus on how to counteract its most harmful applications. The public, press, pundits and policymakers speak of laws and pledges to halt the harmful use of AI.

Deep learning is one of the most popular models currently being used in real-world, Data Science applications. It's been an effective model in areas that range from image to text to voice/music. With the increase in its use, the ability to quickly and scalably implement deep learning becomes paramount. The rise of deep learning platforms such as Tensorflow, help developers implement what they need to in easier ways. In this article, we will learn how deep learning works and get familiar with its terminology -- such as backpropagation and batch size.

OpenAI, a nonprofit focused on creating human-level artificial intelligence, just released an update to its GPT-2 text generator. I'm not being hyperbolic when I say that, after trying it, I'm legitimately terrified for the future of humanity if we don't figure out a way to detect AI-generated content – and soon. GPT-2 isn't a killer robot and my fears aren't that AI is going to rise up against us. I'm terrified of GPT-2 because it represents the kind of technology that evil humans are going to use to manipulate the population -- and in my opinion that makes it more dangerous than any gun. Here's how it works: you give it a prompt and it near-instantly spits out a bunch of words. What's scary about it is that it works.

Over the past three to four years, there has been an evolving breakthrough in the world of technology. This has put into great knowledge on how powerful machines can become in making decisions based completely on facts and figures that have been around for centuries‐ a feat not totally possible with any amount of human effort. This movement in the world of technology (understanding of data) has led to many studies that individually make dramatic progress in improving the world. One of these areas is what has been known as Deep Learning. But what exactly is it?

Facial mapping example used to construct a deepfake video. As society increasingly wrestles with the impact of "deep fakes" it is important to acknowledge that much of the current hype is overstated. While it is true that deep fakes constitute an important future issue, today's tools are far less developed or user-friendly than the breathless media hype and public fears might suggest. Most importantly, in many ways the rise of deep fakes mirrors the rise of Photoshop nearly 30 years ago, with a rise in falsified images met with increasing scrutiny of image-based evidence. While deep learning approaches promise higher-quality falsifications with greater ease, if Photoshop did not lead us to abandon our trust in images, do deep fakes really represent the threat we believe they do?

Alibaba, the largest e-commerce platform in China, is a powerhouse not only when it comes to e-commerce, but also when it comes to recommender systems research. Their latest paper, Behaviour Sequence Transformer for E-commerce Recommendation in Alibaba, is yet another publication that pushes the state of the art in recommender systems. In this work, they make use of the popular Transformer model to capture sequential signals in user behaviour in online shopping, in order to perform next click prediction. Recommender systems often make use of a 2-stage paradigm of retrieval and ranking, and Alibaba's approach is no different. The retrieval step used at Alibaba consists of selecting, with high recall, a subset of a million relevant candidate items from the entire item set (which is of course much larger than a million possible items), and the ranking step consists of ranking these candidates with high precision.

Researchers at UCLA and NantWorks have developed an artificial intelligence-powered device that detects cancer cells in a few milliseconds--hundreds of times faster than previous methods. With that speed, the invention could make it possible to extract cancer cells from blood immediately after they are detected, which could in turn help prevent the disease from spreading in the body. A paper about the advance was published in the journal Nature Scientific Reports. The approach relies on two core technologies: deep learning and photonic time stretch. Deep learning is a type of machine learning, an artificial intelligence technique in which algorithms are "trained" to perform tasks using large volumes of data.

The simplest form of RNN ("vanilla" RNN) is similar to a regular neural network, only it contains a loop that allows the model to carry forward results from previous neuron layers. The image below "unrolls" how the loop works. The network looks at a series of inputs over time, X0, X1, X2, until Xt. For example, this could be a sequence of words in a sentence. The neural network has one layer of neurons for each input (in our example, one layer for each word).

This post will go point by point to see how these mistakes can manifest in a PyTorch code sample. Andrej says we should overfit a single batch. I've wasted HOURS training on a giant dataset, just to find out it's only 50% accurate because of a minor bug. The results you'll get are a good guess for the optimal performance of your architecture when it perfectly memorizes the input. Maybe that optimal performance is zero, because an exception gets thrown mid-way through.

One of the byproducts of our digitally transformed world is the accumulation of large quantities of data. Online transactions, medical records, social media posts, emails, instant messages, and connected sensors are just a few examples of the kinds of data being captured and stored on a daily basis. Scientists and research organizations have been exploring how to leverage big data for artificially intelligent applications since the 1970s. Nonetheless, until fairly recently, the big data issues for enterprises remained how to store it cost effectively, how to retrieve it efficiently when needed, and how to protect it from unauthorized access. The growth of the cloud opened up a whole new realm of cost-effective data storage and retrieval solutions, but big data was still largely perceived by enterprises as a passive asset that did not contribute significantly to their bottom lines.