We propose a new generative model, Cramer-Wold Autoencoder (CWAE). Following WAE, we directly encourage normality of the latent space. Our paper uses also the recent idea from Sliced WAE (SWAE) model, which uses one-dimensional projections as a method of verifying closeness of two distributions. The crucial new ingredient is the introduction of a new (Cramer-Wold) metric in the space of densities, which replaces the Wasserstein metric used in SWAE. We show that the Cramer-Wold metric between Gaussian mixtures is given by a simple analytic formula, which results in the removal of sampling necessary to estimate the cost function in WAE and SWAE models. As a consequence, while drastically simplifying the optimization procedure, CWAE produces samples of a matching perceptual quality to other SOTA models.

During the last decade, the applications of signal processing have drastically improved with deep learning. However areas of affecting computing such as emotional speech synthesis or emotion recognition from spoken language remains challenging. In this paper, we investigate the use of a neural Automatic Speech Recognition (ASR) as a feature extractor for emotion recognition. We show that these features outperform the eGeMAPS feature set to predict the valence and arousal emotional dimensions, which means that the audio-to-text mapping learning by the ASR system contain information related to the emotional dimensions in spontaneous speech. We also examine the relationship between first layers (closer to speech) and last layers (closer to text) of the ASR and valence/arousal.

When ontologies cover overlapping topics, the overlap can be represented using ontology alignments. These alignments need to be continuously adapted to changing ontologies. Especially for large ontologies this is a costly task often consisting of manual work. Finding changes that do not lead to an adaption of the alignment can potentially make this process significantly easier. This work presents an approach to finding these changes based on RDF embeddings and common classification techniques. To examine the feasibility of this approach, an evaluation on a real-world dataset is presented. In this evaluation, the best classifiers reached a precision of 0.8.

Chico Q. Camargo University of Oxford Deep neural networks generalize remarkably well without explicit regularization even in the strongly over-parametrized regime. This success suggests that some form of implicit regularization must be at work. By applying a modified version of the coding theorem from algorithmic information theory and by performing extensive empirical analysis of random neural networks, we argue that the parameter function map of deep neural networks is exponentially biased towards functions with lower descriptional complexity. We show explicitly for supervised learning of Boolean functions that the intrinsic simplicity bias of deep neural networks means that they generalize significantly better than an unbiased learning algorithm does. The superior generalization due to simplicity bias can be explained using PAC-Bayes theory, which yields useful generalization error bounds for learning Boolean functions with a wide range of complexities. Finally, we provide evidence that deeper neural networks trained on the CIFAR10 data set exhibit stronger simplicity bias than shallow networks do, which may help explain why deeper networks generalize better than shallow ones do.

Whenever you search on Google for something you're thinking about buying, you're using a selection of keywords and signalling your intent to make a purchase. If you search for "brown shoes" and adverts appear in your search results suggesting where to get brown shoes, you've experienced "search retargeting", a technique that brands are combining with artificial intelligence (AI) to figure out what makes consumers tick. By mapping and analysing the keywords that consumers use in their search queries, brands can learn more about what people actually want to buy. "Keywords are a great signal of intent," says Carl White, co-founder of Nano Interactive, which provides search targeting technology to brands. "With the developments of AI techniques, you can feel out what people are really searching for. And you can make strong assumptions about what their intent is based on the kind of combination of keywords."

Korean tech giant Samsung has announced a major investment in artificial intelligence research in the UK. The company is to open an AI research lab in Cambridge, in a move that has been welcomed by the prime minister. The lab will join other Samsung centres dedicated to the topic, based in Moscow and Toronto. The technology is now seen as key to competing in many industries. The UK has been a hotspot for AI research.

Please don't hesitate to get in touch if you have a topic you'd like to talk about or a project you want to present! - [masked]:) Anyline is going to sponsor free drinks at the beginning of the evening. Agenda: 7pm: Grab a welcome drink 7.30pm: Is the Singularity near? Where technology and AI could lead us: Facts, forecasts and disruptive projections. A talk by Michael Sprinzl Abstract: Baseline detection is still a challenging task for heterogeneous collections of historical documents. We present a novel approach to baseline extraction in such settings, turning out the winning entry to the ICDAR 2017 Competition on Baseline detection (cBAD).

The graph represents a network of 3,590 Twitter users whose tweets in the requested range contained "futureofwork ", or who were replied to or mentioned in those tweets. The network was obtained from the NodeXL Graph Server on Tuesday, 22 May 2018 at 14:48 UTC. The requested start date was Tuesday, 22 May 2018 at 00:01 UTC and the maximum number of days (going backward) was 14. The maximum number of tweets collected was 5,000. The tweets in the network were tweeted over the 1-day, 22-hour, 57-minute period from Sunday, 20 May 2018 at 01:02 UTC to Tuesday, 22 May 2018 at 00:00 UTC.

We've all woken up, opened our fridges and realised that we need to do some shopping. How many of us have wondered:"I wish I had I had a robot to deliver my groceries?" Well it looks like that's finally happening now! The Co-op have started to deliver groceries to customers in Milton Keynes by using cute robots that are created by Starship Technologies. These'ground drones' have delivered approximately 200 different products over the past four weeks.

It was the first day of school in Russia, a much-beloved unofficial holiday, and President Vladimir Putin was on stage in a national TV broadcast, chatting with jeans-clad teenagers about the future. "Artificial intelligence is the future," he told them, "not only for Russia, but for all humankind. It comes with colossal opportunities, but also threats that are difficult to predict. Whoever becomes the leader in this sphere will become the ruler of the world." Then, this March, in the final moments of Putin's re-election campaign, came a stern message to lawmakers at his annual address to parliament: "The speed of technological progress is accelerating sharply... Those who manage to ride this technological wave will surge far ahead. Those who fail to do this will be submerged and drown."