Recently neural networks and multiple instance learning are both attractive topics in Artificial Intelligence related research fields. Deep neural networks have achieved great success in supervised learning problems, and multiple instance learning as a typical weakly-supervised learning method is effective for many applications in computer vision, biometrics, nature language processing, etc. In this paper, we revisit the problem of solving multiple instance learning problems using neural networks. Neural networks are appealing for solving multiple instance learning problem. The multiple instance neural networks perform multiple instance learning in an end-to-end way, which take a bag with various number of instances as input and directly output bag label. All of the parameters in a multiple instance network are able to be optimized via back-propagation. We propose a new multiple instance neural network to learn bag representations, which is different from the existing multiple instance neural networks that focus on estimating instance label. In addition, recent tricks developed in deep learning have been studied in multiple instance networks, we find deep supervision is effective for boosting bag classification accuracy. In the experiments, the proposed multiple instance networks achieve state-of-the-art or competitive performance on several MIL benchmarks. Moreover, it is extremely fast for both testing and training, e.g., it takes only 0.0003 second to predict a bag and a few seconds to train on a MIL datasets on a moderate CPU.

Using top-ranked documents in response to a query has been shown to be an effective approach to improve the quality of query translation in dictionary-based cross-language information retrieval. In this paper, we propose a new method for dictionary-based query translation based on dimension projection of embedded vectors from the pseudo-relevant documents in the source language to their equivalents in the target language. To this end, first we learn low-dimensional vectors of the words in the pseudo-relevant collections separately and then aim to find a query-dependent transformation matrix between the vectors of translation pairs appeared in the collections. At the next step, representation of each query term is projected to the target language and then, after using a softmax function, a query-dependent translation model is built. Finally, the model is used for query translation. Our experiments on four CLEF collections in French, Spanish, German, and Italian demonstrate that the proposed method outperforms a word embedding baseline based on bilingual shuffling and a further number of competitive baselines. The proposed method reaches up to 87% performance of machine translation (MT) in short queries and considerable improvements in verbose queries.

This paper investigates a novel problem of generating images from visual attributes. We model the image as a composite of foreground and background and develop a layered generative model with disentangled latent variables that can be learned end-to-end using a variational auto-encoder. We experiment with natural images of faces and birds and demonstrate that the proposed models are capable of generating realistic and diverse samples with disentangled latent representations. We use a general energy minimization algorithm for posterior inference of latent variables given novel images. Therefore, the learned generative models show excellent quantitative and visual results in the tasks of attribute-conditioned image reconstruction and completion.

An exact Bayesian network learning algorithm is obtained by recasting the problem into a standard Bayesian network learning problem without missing data. To the best of our knowledge, this is the first exact algorithm for this problem. As expected, the exact algorithm does not scale to large domains. We build on the exact method to create an approximate algorithm using a hill-climbing technique. This algorithm scales to large domains so long as a suitable standard structure learning method for complete data is available. We perform a wide range of experiments to demonstrate the benefits of learning Bayesian networks with such new approach.

Microbiology might not be everyone's idea of fun, but a new DIY project from Stanford University aims to bring'playful interaction' to the study of single-celled organisms. A bioengineer has developed a 3-D printed smartphone microscope that allows users to observe microbes called Euglena, to make serious observations or even play games. The researchers have revealed interactive platforms to go along with the device, including a virtual soccer game and a Pac-Man-like maze, which rely on the movement of the tiny light-seeking organisms. A bioengineer has developed a 3-D printed smartphone microscope that allows users to observe microbes called Euglena, to make serious observations or even play games. A small platform on the device holds the microscope slide, where Euglena swim around freely. These microbes respond to light, and four LEDs activated by a joystick can be used to influence the directions by which they travel by varying the intensity.

NEW YORK A federal appeals court on Friday reinstated a 120 million jury award for Apple Inc (AAPL.O) against Samsung (005930.KS), marking the latest twist in the fierce patent war between the world's top smartphone manufacturers. The court said that there was substantial evidence for the jury verdict related to Samsung's infringement of Apple patents on its slide-to-unlock and autocorrect features, as well as quick links, which automatically turn information like addresses and phone numbers into links. Friday's decision was made by the full slate of judges on the U.S. Court of Appeals for the Federal Circuit in Washington, D.C. In an 8-3 ruling, the judges said that a previous panel of the same court should not have overturned the verdict last February. The three-judge panel did not follow U.S. Supreme Court limits on the scope of its review, because it examined evidence outside the record of the case, the decision said.

This file photo taken on March 25, 2015 shows an UBER application shown as cars drive by in Washington, DC. The US ride-sharing service Uber announced August 18, 2016 it had acquired the commercial transport-focused tech startup Otto as the company presses ahead with its pursuit of self-driving technology. The announcement came as the company also announced a 300 million effort with the Sweden-based automaker Volvo Cars to develop driverless cars."If that sounds like a big deal -- well, it is," Uber CEO Travis Kalanick said in a statement. A Philadelphia judge had ordered ride-hailing companies to cease operations but both Uber and Lyft continued to offer rides there on Friday. Judge Linda Carpenter issued the injunction Thursday.

Neural networks are a rapidly developing paradigm for information processing based loosely on how neurons in the brain processes information. A neural network consists of multiple layers of node, where each node performs a unit of computation, and passes the result onto the next node. Multiple nodes can pass inputs to a single node, and vice-versa. The neural network also contains a set of weights, which can be refined over time as the network learns from sample data. The weights are used to describe and refine the connection strengths between nodes.

How you see the world may be very different than how someone else sees it. And recognizing that has long been thought to be a uniquely human ability. But when it comes to understanding others' perspectives, humans might not be alone. "Reading others' minds is not our special skill," says Fumihiro Kano, a comparative psychologist at Kyoto University in Japan. Nonhuman apes can do it, too, according to Dr. Kano's research, published Thursday in the journal Science, a finding that could further blur the line between the cognitive capacities of humans and nonhuman apes.

Neural networks work so well because they roughly reflect the structure of the underlying universe around us, new research suggests.