As machine learning increasingly affects domains protected by anti-discrimination law, there is much interest in the problem of algorithmically measuring and ensuring fairness in machine learning. Across academia and industry, experts are finally embracing this important research direction that has long been marred by sensationalist clickbait overshadowing scientific efforts. This sequence of posts is a sober take on the subtleties and difficulties in engaging productively with the issue of fairness in machine learning. Prudence is necessary, since a poor regulatory proposal could easily do more harm than doing nothing at all. In this first post, I will focus on a sticky idea I call demographic parity that through its many variants has been proposed as a fairness criterion in dozens of papers.

Apple's Siri upgrades will let you dictate LinkedIn messages, verbally search for shoe photos on Pinterest, and hail an Uber, hands-free. When Apple unveils its new iPhone Wednesday, it's expected to push hard on software improvements available for its next mobile operating system, iOS 10, available for current iPhones and a predicted new iPhone model. The iPhone's voice-activated digital assistant Siri will be a key part of this upgrade, say tech analysts and developers. Siri, primarily used for voice commands to dial phone numbers, set reminders and get traffic information, has been opened to third-party developers after five years of existence. Apple, via a spokeswoman, said app developers are creating "completely new ways for users to interact" within apps.

Neural Nets are quite powerful models in machine learning which are used for learning the behavior of high dimensional data. Typically, data is not always present in its purest form, the signal. Most of the data that is made available for machines during training or runtime prediction do have some amount of noise. During training, the Neural Nets can get highly sensitive to the noise in the input and can start overfitting the learning to the noise in the input. In the previous post titled "Algorithms to Improve NeuralNetwork Accuracy", we learnt about the overfitting problem in Neural Nets and how to break it using L1/L2 regularizers, Weight penalties decay and constraints.

Austin startup CognitiveScale has raised 21.8 million to ramp up sales and marketing of its artificial intelligence technology. The company, founded by members of the IBM Watson team, received the funding from Norwest Venture Partners and Intel Capital. It has raised a total of 32 million since inception. CognitiveScale has created a cloud platform that can mine huge databases and use natural language to present data in an easy to understand format. The platform can be plugged into a company's website and extracts insights from the company's existing data.

The Tate Britain art museum has just launched a new artificial intelligence exhibit that applies machine learning technology to images in some pretty unique ways. Called "Recognition," the exhibit is the winner of Tate's annual IK Prize, which was created in association with Microsoft and awards "digital innovation." It compares images from Tate's enormous archive of artwork with up-to-the-minute Reuters news photography, based on various pattern-recognition tools. These include object recognition, facial recognition, composition analysis, and even natural language processing for looking at captions and text. "From the moment it launched, it's continually scanning both databases and comparing images, trying to find works which are comparable -- whether that be visually or thematically -- and then publishing them online in a virtual gallery," Tony Guillan, producer of the IK Prize, told Digital Trends.

Gov. John Kasich says manufacturing is still important to Ohio, but high-tech is the future. Kasich was the keynote speaker at tonight's annual meeting of the Greater Cleveland Partnership, the area's chamber of commerce. Gov. Kasich started off his address by praising Cleveland for recent successes like the Republican National Convention, the Cavalier's championship and the city's growth as a tourist and convention destination. He moved on to talking about the entire state. He says he's working with artificial-intelligence and information- technology experts on ways to help Ohio advance.

Osprey is a tool for hyperparameter optimization of machine learning algorithms in Python. Hyperparameter optimization can often be an onerous process for researchers, due to time-consuming experimental replicates, non-convex objective functions, and constant tension between exploration of global parameter space and local optimization (Jones, Schonlau, and Welch 1998). We've designed Osprey to provide scientists with a practical, easy-to-use way of finding optimal model parameters. As hyperparameter optimization is an embarrassingly parallel problem, Osprey can easily scale to hundreds of concurrent processes by executing a simple command-line program multiple times.

Osprey is a tool for hyperparameter optimization of machine learning algorithms in Python. Hyperparameter optimization can often be an onerous process for researchers, due to time-consuming experimental replicates, non-convex objective functions, and constant tension between exploration of global parameter space and local optimization (Jones, Schonlau, and Welch 1998). We've designed Osprey to provide scientists with a practical, easy-to-use way of finding optimal model parameters. The software works seamlessly with scikit-learn estimators (Pedregosa et al. 2011) and supports many different search strategies for choosing the next set of parameters with which to evaluate a given model, including gaussian processes (GPy 2012), tree-structured Parzen estimators (Yamins, Tax, and Bergstra 2013), as well as random and grid search. As hyperparameter optimization is an embarrassingly parallel problem, Osprey can easily scale to hundreds of concurrent processes by executing a simple command-line program multiple times.

Geoffrey Hinton is a pioneer in the field of artificial neural networks and co-published the first paper on the backpropagation algorithm for training multilayer perceptron networks. He may have started the introduction of the phrasing "deep" to describe the development of large artificial neural networks. He co-authored a paper in 2006 titled "A Fast Learning Algorithm for Deep Belief Nets" in which they describe an approach to training "deep" (as in a many layered network) of restricted Boltzmann machines. Using complementary priors, we derive a fast, greedy algorithm that can learn deep, directed belief networks one layer at a time, provided the top two layers form an undirected associative memory. This paper and the related paper Geoff co-authored titled "Deep Boltzmann Machines" on an undirected deep network were well received by the community (now cited many hundreds of times) because they were successful examples of greedy layer-wise training of networks, allowing many more layers in feedforward networks.

Which authors of this paper are endorsers? Disable MathJax (What is MathJax?)