In multiagent systems, social choice functions can help aggregate the distinct preferences that agents have over alternatives, enabling them to settle on a single choice. Despite the basic manipulability of all reasonable voting systems, it would still be desirable to find ways to reach plausible outcomes, which are stable states, i.e., a situation where no agent would wish to change its vote. One possibility is an iterative process in which, after everyone initially votes, participants may change their votes, one voter at a time. This technique, explored in previous work, converges to a Nash equilibrium when Plurality voting is used, along with a tie-breaking rule that chooses a winner according to a linear order of preferences over candidates. In this paper, we both consider limitations of the iterative voting method, as well as expanding upon it. We demonstrate the significance of tie-breaking rules, showing that no iterative scoring rule converges for all tie-breaking. However, using a restricted tie-breaking rule (such as the linear order rule used in previous work) does not by itself ensure convergence. We prove that in addition to plurality, the veto voting rule converges as well using a linear order tie-breaking rule. However, we show that these two voting rules are the only scoring rules that converge, regardless of tie-breaking mechanism.

Sparsity-based subspace clustering algorithms have attracted significant attention thanks to their excellent performance in practical applications. A prominent example is the sparse subspace clustering (SSC) algorithm by Elhamifar and Vidal, which performs spectral clustering based on an adjacency matrix obtained by sparsely representing each data point in terms of all the other data points via the Lasso. When the number of data points is large or the dimension of the ambient space is high, the computational complexity of SSC quickly becomes prohibitive. Dyer et al. observed that SSC-OMP obtained by replacing the Lasso by the greedy orthogonal matching pursuit (OMP) algorithm results in significantly lower computational complexity, while often yielding comparable performance. The central goal of this paper is an analytical performance characterization of SSC-OMP for noisy data. Moreover, we introduce and analyze the SSC-MP algorithm, which employs matching pursuit (MP) in lieu of OMP. Both SSC-OMP and SSC-MP are proven to succeed even when the subspaces intersect and when the data points are contaminated by severe noise. The clustering conditions we obtain for SSC-OMP and SSC-MP are similar to those for SSC and for the thresholding-based subspace clustering (TSC) algorithm due to Heckel and B\"olcskei. Analytical results in combination with numerical results indicate that both SSC-OMP and SSC-MP with a data-dependent stopping criterion automatically detect the dimensions of the subspaces underlying the data. Moreover, experiments on synthetic and real data show that SSC-MP compares very favorably to SSC, SSC-OMP, TSC, and the nearest subspace neighbor (NSN) algorithm, both in terms of clustering performance and running time. In addition, we find that, in contrast to SSC-OMP, the performance of SSC-MP is very robust with respect to the choice of parameters in the stopping criteria.

An emerging design principle in deep learning is that each layer of a deep artificial neural network should be able to easily express the identity transformation. This idea not only motivated various normalization techniques, such as \emph{batch normalization}, but was also key to the immense success of \emph{residual networks}. In this work, we put the principle of \emph{identity parameterization} on a more solid theoretical footing alongside further empirical progress. We first give a strikingly simple proof that arbitrarily deep linear residual networks have no spurious local optima. The same result for linear feed-forward networks in their standard parameterization is substantially more delicate. Second, we show that residual networks with ReLu activations have universal finite-sample expressivity in the sense that the network can represent any function of its sample provided that the model has more parameters than the sample size. Directly inspired by our theory, we experiment with a radically simple residual architecture consisting of only residual convolutional layers and ReLu activations, but no batch normalization, dropout, or max pool. Our model improves significantly on previous all-convolutional networks on the CIFAR10, CIFAR100, and ImageNet classification benchmarks.

There's a harsh reality we need to face--a robotic, AI-driven Shakespeare is nowhere in sight. No robot will write verse that influences English the way Bard's did anytime soon. You won't find an AI spitting rhymes like Rakim or Nas, either. But if your standards aren't too high, there is some AI-constructed poetry you can read today. Take an AI that uses the recurrent neural network language model technique, feed it thousands of romantic novels to learn language from, give it a starting sentence and an ending sentence, instruct it to fill the gap between them, and you'll get something like this: This AI, designed by Google, Stanford University, and the University of Massachusetts, isn't supposed to be the world's first artificial poet--it's just a side effect. But if you keep in mind that the AI generated all of the sentences except the first and last on its own, it's impressive that they all make sense and have a common theme.

I have been pondering the security technology encroachments into public life, particularly regarding CCTV monitoring. There was a time, it seems now very long ago, that the UK was awash in CCTV. Hundreds of millions of dollars, and over four million (and counting) CCTV cameras later, the UK is the most surveilled society on earth. We were assured that would never happen in the US, or other developed countries. Still, if you have nothing to hide… Today, London and Beijing have over 400,000 CCTV each (proving politics is no guarantee either way). In the US there are over 30 million CCTV cameras, mostly in private hands.

Palantir CEO Alex Karp Says Going Public Is'A Possibility' Two false news stories from a website registered in Macedonia. If one looks deeply at the world of "fake news" one finds that much of what makes it possible has its roots in modern journalistic practice. From the inverted pyramid style of summarizing major findings in the headline and lede, while burying the details at the bottom, to writing for one's audience and skipping complex detail, the sacrosanct practices of journalism have been weaponized by the clickbait and fake news communities. Take the standard inverted pyramid format of Western journalism. No matter whether you are writing for a fringe blog or one of the most respected news outlets in the world, the pyramid style of writing means the headline will summarize the article in a few meme-ready share-worthy words, the lead paragraph will offer a tantalizing preview of the major conclusions, and the remaining top-level paragraphs will all convey the big story.

During Microsoft's Build Conference earlier this year, CEO Satya Nadella delivered the three-hour keynote address, in which he highlighted his belief that the future of technology lies in human language. In this new wave of technology, conversation is the new interface, and "bots are the new apps." While not as flashy as virtual reality nor as immediately practical as 3D printing, chatbots are nevertheless gaining major traction this year, with support coming from across the entire tech industry. The big tech enterprises are all entering the chatbot space, and many startups are too. The question is, why is so much attention being given to chatbots recently, and should your business be paying attention too?

What is an evolutionary algorithm?

Want to watch this again later? Report Need to report the video? Report Need to report the video? Need to report the video? This feature is not available right now.

With the advent of photography, a tiny fraction of 19th-century scientists believed they could develop methods of accurately identifying criminals by their facial features. While their hypotheses were eventually discredited, new artificial intelligence technology suggests their claims might've been valid after all. Xiaolin Wu and Xi Zhang from Shanghai Jiao Tong University in China have resurrected this facial recognition tradition and built a neural network that can supposedly pick out criminals by simply looking at their faces. TNW Conference is back for its 12th year. To accomplish this, the researchers used an array of machine-vision algorithms to examine a series of facial juxtapositions between photos of criminals and non-criminals with the goal of finding out whether a neural network can reliably tell them apart.