Motivated by applications in recommender systems, web search, social choice and crowdsourcing, we consider the problem of identifying the set of top $K$ items from noisy pairwise comparisons. In our setting, we are non-actively given $r$ pairwise comparisons between each pair of $n$ items, where each comparison has noise constrained by a very general noise model called the strong stochastic transitivity (SST) model. We analyze the competitive ratio of algorithms for the top-$K$ problem. In particular, we present a linear time algorithm for the top-$K$ problem which has a competitive ratio of $\tilde{O}(\sqrt{n})$; i.e. to solve any instance of top-$K$, our algorithm needs at most $\tilde{O}(\sqrt{n})$ times as many samples needed as the best possible algorithm for that instance (in contrast, all previous known algorithms for the top-$K$ problem have competitive ratios of $\tilde{\Omega}(n)$ or worse). We further show that this is tight: any algorithm for the top-$K$ problem has competitive ratio at least $\tilde{\Omega}(\sqrt{n})$.

Most existing learning to hash methods assume that there are sufficient data, either labeled or unlabeled, on the domain of interest (i.e., the target domain) for training. However, this assumption cannot be satisfied in some real-world applications. To address this data sparsity issue in hashing, inspired by transfer learning, we propose a new framework named Transfer Hashing with Privileged Information (THPI). Specifically, we extend the standard learning to hash method, Iterative Quantization (ITQ), in a transfer learning manner, namely ITQ+. In ITQ+, a new slack function is learned from auxiliary data to approximate the quantization error in ITQ. We developed an alternating optimization approach to solve the resultant optimization problem for ITQ+. We further extend ITQ+ to LapITQ+ by utilizing the geometry structure among the auxiliary data for learning more precise binary codes in the target domain. Extensive experiments on several benchmark datasets verify the effectiveness of our proposed approaches through comparisons with several state-of-the-art baselines.

Cross-domain visual data matching is one of the fundamental problems in many real-world vision tasks, e.g., matching persons across ID photos and surveillance videos. Conventional approaches to this problem usually involves two steps: i) projecting samples from different domains into a common space, and ii) computing (dis-)similarity in this space based on a certain distance. In this paper, we present a novel pairwise similarity measure that advances existing models by i) expanding traditional linear projections into affine transformations and ii) fusing affine Mahalanobis distance and Cosine similarity by a data-driven combination. Moreover, we unify our similarity measure with feature representation learning via deep convolutional neural networks. Specifically, we incorporate the similarity measure matrix into the deep architecture, enabling an end-to-end way of model optimization. We extensively evaluate our generalized similarity model in several challenging cross-domain matching tasks: person re-identification under different views and face verification over different modalities (i.e., faces from still images and videos, older and younger faces, and sketch and photo portraits). The experimental results demonstrate superior performance of our model over other state-of-the-art methods.

A SpaceX capsule returned to Earth on Wednesday with precious science samples from NASA's one-year space station resident. SpaceX reported a good splashdown, with three red-and-white striped parachutes slowing the final descent. The Dragon had been at the station for a month, dropping off supplies as well as an experimental, inflatable room that will pop open in two weeks. It was set free by the station's big robot arm. "Dragon spacecraft has served us well, and it's good to see it departing full of science," Peake radioed from 250 miles up.

Scientists at Los Alamos National Laboratory and the State Key Laboratory for Mechanical Behavior of Materials in China have used a combination of machine learning, supercomputers, and experiments to speed up discovery of new materials with desired properties. The idea is to replace traditional trial-and-error materials research, which is guided only by intuition (and errors). With increasing chemical complexity, the possible combinations have become too large for those trial-and-error approaches to be practical. The scientists focused their initial research on improving nickel-titanium (nitinol) shape-memory alloys (materials that can recover their original shape at a specific temperature after being bent). But the strategy can be used for any materials class (polymers, ceramics, or nanomaterials) or target properties (e.g., dielectric response, piezoelectric coefficients, and band gaps).

No one can predict the future of technology with 100 percent accuracy. But these four pillars are certainly at the forefront of innovation in the years ahead. Speaking at a machine learning and artificial intelligence event hosted by Madrona Venture Group in Seattle on Wednesday, Joseph Sirosh, corporate VP of the Data Group at Microsoft, outlined five trends to watch in a world he described as "ACID": Algorithm, Cloud, IoT, and Data. "We live in a time of great change in computing, where unreasonable effectiveness of algorithms, cloud, IoT, and data are changing how applications are built, period," he said. "Even if you are on the right track, if you don't hop on this bandwagon and actually build things and deploy them and take advantage of their strength, you won't be very effective."

PointGrab has announced a new white paper that explores the impact of deep learning-based smart sensor technology on building automation management. The white paper was developed to help building automation industry stakeholders, from device manufacturers to building managers, better understand the long-term benefits of deep learning-based technology. The paper, "Smarter Sensors: How Deep Learning is Transforming Building Automation," addresses the challenges and opportunities presented by Internet of Things (IoT) device proliferation and data collection and analytics throughout the smart building ecosystem. In this data-rich environment, sensors can be much smarter by sourcing and analyzing richer levels of data and enabling the execution of more sophisticated tasks that go beyond traditional energy consumption management.

Endowing robots and prosthetics with a human-like sense of touch could dramatically advance these technologies. Toward this goal, scientists have come up with various smart skins to layer onto devices. But boosting their sensitivity has involved increasing the numbers of electrodes, depending on the size of the skin. This leads to a rise in costs. Other systems require external batteries and wires to operate, which adds to their bulk. Haixia Zhang and colleagues wanted to find a more practical solution.

Artificial intelligence, or AI, no longer simply exists in science fiction movies and books. Scientists warn AI has and will continue to change almost every aspect of how people conduct business and live. Researchers say artificial intelligence can be a threat, as well as helpful, to humans. From the iPhone personal assistant Siri, to doing searches on the Internet, to the autopilot function, simple artificial intelligence has been around for some time, but is quickly getting more complex and more intelligent. "If we are going to make systems that are going to be more intelligent than us, it's absolutely essential for us to understand how to absolutely guarantee that they only do things that we are happy with," said Stuart Russell, computer science professor at the University of California Berkeley.

Researchers from Stanford University have created a humanoid robot or robot mermaid to explore sunken treasures and relics. Tagged as OceanOne, the robo-mermaid uses artificial intelligence and virtual reality technology to allow human beings to operate it remotely, as per Stanford News. The robot mermaid looks like a human with hands that are installed with sensors to enable OceanOne to discern if an item is fragile or not. It also has two cameras as its eyes and an artificial human brain for navigating the deep sea and analyzing data. According to CNN, OceanOne's first journey to the deep water was to retrieve a vase from the ruins of Louis XIV's ship La Lune.