Nearly every day, in war zones around the world, American military forces request fire support. By radioing coordinates to a howitzer miles away, infantrymen can deliver the awful ruin of a 155-mm artillery shell on opposing forces. If defense officials in Washington have their way, artificial intelligence is about to make that process a whole lot faster.
The effort to speed up fire support is one of a handful initiatives that Lt. Gen. Jack Shanahan describes as the “lower consequence missions” that the Pentagon is using to demonstrate how it can integrate artificial intelligence into its weapons systems. As the head of the Joint Artificial Intelligence Center, a 140-person clearinghouse within the Department of Defense focused on speeding up AI adoption, Shanahan and his team are building applications in well-established AI domains—tools for predictive maintenance and health record analysis—but also venturing into the more exotic, pursuing AI capabilities that would make the technology a centerpiece of American warfighting. Read More
Monthly Archives: December 2019
In Event of Moon Disaster – Nixon Deepfake Clips
This Nixon Deepfake Is an Alternate Reality Where Apollo 11 Fails
Deepfake technology makes the impossible, possible—well, at least visually possible. In this case, we’re talking about Richard Nixon and a speech of his that never actually occurred—a speech where he announces the death of all three Apollo 11 astronauts on the surface of the moon. Read More
Is AI About to Hit a Wall?
There have been several stories over the last several months around the theme that AI is about to hit a wall. That the rapid improvements we’ve experienced and the benefits we’ve accrued can’t continue at the current pace. It’s worth taking a look at these arguments to see if we should be adjusting our plans and expectations. Read More
CryptoNN: Training Neural Networks over Encrypted Data
Emerging neural networks based machine learning techniques such as deep learning and its variants have shown tremendous potential in many application domains. However,they raise serious privacy concerns due to the risk of leakage of highly privacy-sensitive data when data collected from usersis used to train neural network models to support predictive tasks. To tackle such serious privacy concerns, several privacy-preserving approaches have been proposed in the literature that use either secure multi-party computation (SMC) or homomorphic encryption (HE) as the underlying mechanisms. However, neither of these cryptographic approaches provides an efficient solution towards constructing a privacy-preserving machine learning model, as well as supporting both the training and inference phases.To tackle the above issue, we propose a CryptoNN framework that supports training a neural network model over encrypted data by using the emerging functional encryption scheme instead of SMC or HE. We also construct a functional encryption scheme for basic arithmetic computation to support the requirement ofthe proposed CryptoNN framework. We present performance evaluation and security analysis of the underlying crypto scheme and show through our experiments that CryptoNN achieves accuracy that is similar to those of the baseline neural network models on the MNIST dataset. Read More
Finland is making its online AI crash course free to the world
Last year, Finland launched a free online crash course in artificial intelligence with the aim of educating its citizens about the new technology. Now, as a Christmas present to the world, the European nation is making the six week program available for anyone to take.
Strictly speaking, it’s a present for the European Union. Finland is relinquishing the EU’s rotating presidency at the end of the year, and decided to translate its course into every EU language as a gift to citizens. But there aren’t any geographical restrictions as to who can take the course, so really it’s to the world’s benefit. Read More
Why some platforms succeed … while most fail
One of the most important business models of the 21st century is, without question, the platform model, which powers many of today’s biggest and most disruptive companies. Innovation platform, such as Google Android, Apple iPhone operating systems, and Amazon Web Services enable third-party firms to add complementary products and services to a core product or technology. Transaction platforms, which include the likes of Amazon Marketplace, Airbnb, and Uber, bring together producers and users in efficient exchanges of value, and they leverage network effects—the more participants, the greater the value produced. Read More
7 Artificial Intelligence Trends to Watch in 2020
Artificial Intelligence offers great potential and, for some, risks for humans in the future. While still in its infancy it is being employed in some interesting ways. X
Here we explore some of the main AI trends predicted by experts in the field. If correct, 2020 should see some very exciting developments indeed. Read More
Self-training with Noisy Student improves ImageNet classification
We present a simple self-training method that achieves 87.4% top-1 accuracy on ImageNet, which is 1.0% better than the state-of-the-art model that requires 3.5B weakly labeled Instagram images. On robustness test sets, it improves ImageNet-A top-1 accuracy from 16.6% to 74.2%, reduces ImageNet-C mean corruption error from 45.7 to 31.2, and reduces ImageNet-P mean flip rate from 27.8 to 16.1.
To achieve this result, we first train an EfficientNet model on labeled ImageNet images and use it as a teacher to generate pseudo labels on 300M unlabeled images. We then train a larger EfficientNet as a student model on the combination of labeled and pseudo labeled images. We iterate this process by putting back the student as the teacher. During the generation of the pseudo labels, the teacher is not noised so that the pseudo labels are as good as possible. But during the learning of the student, we inject noise such as data augmentation, dropout, stochastic depth to the student so that the noised student is forced to learn harder from the pseudo labels. Read More
There’s a new way to tame language AI so it doesn’t embarrass you
In the last two years, the AI subfield of natural-language processing has seen enormous progress. For example, a language model developed by the San Francisco–based research lab OpenAI, called GPT-2, has been used to generate fiction, fake news articles, and a practically infinite Choose Your Own Adventure–style text game.
But these kinds of models are essentially massive text-prediction systems that don’t take sense into account, so the sentences they produce are more likely to be superficially fluent than they are to be truly meaningful. It’s hard to tell a model to stick to a particular topic like health care, for example. Yet models like GPT-2 can still be gamed to produce racist and toxic output, making them even less useful. Read More
When Data Creates Competitive Advantage
Many executives and investors assume that it’s possible to use customer-data capabilities to gain an unbeatable competitive edge. The more customers you have, the more data you can gather, and that data, when analyzed with machine-learning tools, allows you to offer a better product that attracts more customers. You can then collect even more data and eventually marginalize your competitors in the same way that businesses with sizable network effects do. Or so the thinking goes. More often than not, this assumption is wrong. In most instances people grossly overestimate the advantage that data confers. en Data Creates Competitive Advantage.
The virtuous cycles generated by data-enabled learning may look similar to those of regular network effects, wherein an offering—like a social media platform—becomes more valuable as more people use it and ultimately garners a critical mass of users that shuts out competitors. But in practice regular network effects last longer and tend to be more powerful. To establish the strongest competitive position, you need them and data-enabled learning. However, few companies are able to develop both. Nevertheless under the right conditions customer-generated data can help you build competitive defenses, even if network effects aren’t present. In this article we’ll walk you through what those conditions are and explain how to evaluate whether they apply to your business. Read More
Rick's Cafe AI 