This work is a further study on the Generalized Constraint Neural Network (GCNN) model [1], [2]. Two challenges are encountered in the study, that is, to embed any type of prior information and to select its imposing schemes. The work focuses on the second challenge and studies a new constraint imposing scheme for equality constraints. A new method called locally imposing function (LIF) is proposed to provide a local correction to the GCNN prediction function, which therefore falls within Locally Imposing Scheme (LIS). In comparison, the conventional Lagrange multiplier method is considered as Globally Imposing Scheme (GIS) because its added constraint term exhibits a global impact to its objective function. Two advantages are gained from LIS over GIS. First, LIS enables constraints to fire locally and explicitly in the domain only where they need on the prediction function. Second, constraints can be implemented within a network setting directly. We attempt to interpret several constraint methods graphically from a viewpoint of the locality principle. Numerical examples confirm the advantages of the proposed method. In solving boundary value problems with Dirichlet and Neumann constraints, the GCNN model with LIF is possible to achieve an exact satisfaction of the constraints.

We consider the problem of learning a non-negative linear classifier with a $1$-norm of at most $k$, and a fixed threshold, under the hinge-loss. This problem generalizes the problem of learning a $k$-monotone disjunction. We prove that we can learn efficiently in this setting, at a rate which is linear in both $k$ and the size of the threshold, and that this is the best possible rate. We provide an efficient online learning algorithm that achieves the optimal rate, and show that in the batch case, empirical risk minimization achieves this rate as well. The rates we show are tighter than the uniform convergence rate, which grows with $k^2$.

To whom correspondence should be addressed; Email: artitw@gmail.com Customer temporal behavioral data was represented as images in order to perform churn prediction by leveraging deep learning architectures prominent in image classification. Supervised learning was performed on labeled data of over 6 million customers using deep convolutional neural networks, which achieved an AUC of 0.743 on the test dataset using no more than 12 temporal features for each customer. Unsupervised learning was conducted using autoencoders to better understand the reasons for customer churn. Images that maximally activate the hidden units of an autoencoder trained with churned customers reveal ample opportunities for action to be taken to prevent churn among strong data, no voice users.

Google's Alphabet has a new walking robot that wouldn't look out of place in Interstellar or science-fiction homes of the future. The reportedly as-yet-unnamed robot was shown off at the New Economic Summit in Tokyo by Alphabet-owned Japanese robotics company Schaft. It has a very different design to Alphabet's other robots made by Boston Dynamics, with a compact two-leg design and central body that can be moved up or down to cope with different tasks. Unlike Alphabet's larger bipedal robots designed either to interact in a human-like fashion with the world - the humanoid Atlas - or to be a robotic packhorse for the US military or dog's plaything, the Schaft robot is designed to be lower cost, lower power and be used by civilians, carrying up to 60kg over uneven terrain and stairs. The robot was demonstrated dealing with unsure footing, compensating for standing on a moving pipe in one instance and walking on shingle in another.

Researchers from the University of Science and Technology of China on Friday revealed a realistic robot they've been working on for the past three years. Called, Jia Jia, the robot is said to be capable of human-like facial expressions, along with talking and interacting with people nearby. While its creators describe it as looking similar to a "real woman," at least it's a step up from that nightmarish home-made Scarlett Johansson robot. Among the details the researchers have incorporated into Jia Jia are the way its eyes will glance around a room in a natural way, as well as mouth movements that align with its speaking. Not only can it respond to humans, but it can recognize when someone is taking a picture and make appropriate comments, such as warning not to stand too close for fear of making her face "look fat."

Picture the scenario - a sentient machine is "living" in the US in the year 2050 and starts browsing through the US constitution. Having read it, it decides that it wants the opportunity to vote. Oh, and it also wants the right to procreate. Pretty basic human rights that it feels it should have now it has human-level intelligence. "Do you give it the right to vote or the right to procreate because you can't do both?" asks Ryan Calo, a law professor at the University of Washington.

The University of Science and Technology of China on Friday officially launched the robot "Jiajia" it invented for interactive experience. HEFEI, April 15 (Xinhua) -- A new interactive robot, named Jia Jia, was unveiled Friday by the University of Science and Technology of China (USTC) in Hefei, capital of east China's Anhui Province. Welcome!" the eye-catching robot said as it greeted the audience at the university's multi-media center. "Don't come too close to me when you are taking a picture. It will make my face look fat," Jia Jia said. Jia Jia was developed by a robot research and development team at the USTC, which also developed the model service robot "Kejia." It took the team three years to research and develop this new-generation interactive robot, which can speak, show micro-expressions, move its lips, and move its body, according to team director Chen Xiaoping. Compared to previous interactive robots, Jia Jia's eyeballs roll naturally and its speech is in sync with its lip movements, in addition to her human-like form, Chen said. Jia Jia can not cry or laugh and these are areas to be developed, Chen added. "We hope to develop the robot so it has deep learning abilities.

PRINCETON, NEW JERSEY – Last month, AlphaGo, a computer program specially designed to play the game go, caused shock waves among aficionados when it defeated Lee Sidol, one of the world's top-ranked professional players, winning a five-game tournament by a score of 4-1. Why, you may ask, is that news? Twenty years have passed since the IBM computer Deep Blue defeated world chess champion Garry Kasparov, and we all know computers have improved since then. But Deep Blue won through sheer computing power, using its ability to calculate the outcomes of more moves to a deeper level than even a world champion can. Go is played on a far larger board (19 by 19 squares, compared to eight by eight for chess) and has more possible moves than there are atoms in the universe, so raw computing power was unlikely to beat a human with a strong intuitive sense of the best moves.

It has been a nightmare of a PR week for Microsoft. It started with the head of Microsoft's Xbox division, Phil Spencer, having to apologise for having scantily clad female dancers dressed as school girls at a party thrown by Microsoft at the Game Developers Conference (GDC). He said that having the dancers at this event "was absolutely not consistent or aligned to our values. That was unequivocally wrong and will not be tolerated". The matter was being dealt with internally and so we don't know who would have been responsible and why they might have thought this was going to be a good idea.

The University of Science and Technology of China has recently unveiled an eerily realistic robot named Jia Jia. While she looks more human-like than that creepy ScarJo robot, you'll probably still find yourself plunging head first into the uncanny valley while looking at her. Jia Jia can talk and interact with real humans, as well as make some facial expressions -- she can even tell you off if she senses you're taking an unflattering picture of her. "Don't come too close to me when you are taking a picture. It will make my face look fat," she told someone trying to capture her photo during the presscon.