U.S. tech giants, European telecoms firms, Japanese venture capital and the U.K. government has put together a 1 billion-pound ($1.4 billion) investment into the U.K. artificial intelligence industry, as governments weigh how to compete with China. The deal comprises a total of 300 million pounds of private financing, 300 million pounds of new government spending in addition to 400 million pounds the state has already announced. "Artificial intelligence provides limitless opportunities to develop new, efficient and accessible products and services," Business Secretary Greg Clark said in an emailed statement Thursday. In prioritizing AI, Britain joins several other nations that see the emerging technology as central to their future place in the world, such as China, which last year set a goal of becoming the world leader in AI by 2030. Antony Phillipson, the U.K. Trade Commissioner for North America and Consul General in New York, said in an interview ahead of the announcement that the U.K. could not compete with China in terms of total funding or scale of some government-run AI projects.

Ransomware and distributed denial of service (DDoS) attacks have been on the rise, with 232 attacks through January 2018 compared with 106 through January 2017, said the professional services firm's 2018 State of Cyber Resilience study, which investigated focused attacks defined as having the potential to both penetrate network defences and cause damage, or extract high-value assets and processes from within organisations. Some 4,600 enterprise security practitioners representing companies with annual revenues of $1 billion or more in 15 countries were surveyed across the Americas, Europe and Asia-Pacific. Despite the increased pressure of the attacks, organisations are upping their game and preventing 87% of these focused attacks, compared to 70% a year ago. However, with 13% of focused attacks still penetrating defences, organisations are still facing an average of 30 successful security breaches per year. In addition, only two out of five organisations are currently investing in technologies like machine learning, AI and automation, indicating there is even more progress to be made by increasing investment in cyber resilient innovations and solutions.

LONDON (Reuters) - Britain announced on Thursday a 1 billion pound ($1.4 billion) joint investment in the artificial intelligence (AI) industry to capitalize on what the government sees as a competitive advantage in the sector as it prepares for life after Brexit. The deal is the latest in a series of targeted public-private investment pacts in the government's industrial strategy that aims to modernize Britain's economy and address decades of regional and industrial decline. "It's evident that Britain is a place that people want to come to for AI," business minister Greg Clark told Reuters during a visit to a facility in London that nurtures early-stage tech businesses from across the world. "We have a position of strength that we want to capitalize on because if we don't build on it the other countries around the world would steal a march." Governments worldwide are plowing cash AI to keep up with international rivals and seeking to harness its power to transform industries from transport to agriculture.

Those funds will be matched by more than 300 million pounds of public funding, on top of an existing 400 million pound budget. The funds will be spent on teacher training, research and developing regional technology hubs to explore how AI can be used in industries such as law and insurance.

In this paper, we introduce an alternative approach to Temporal Answer Set Programming that relies on a variation of Temporal Equilibrium Logic (TEL) for finite traces. This approach allows us to even out the expressiveness of TEL over infinite traces with the computational capacity of (incremental) Answer Set Programming (ASP). Also, we argue that finite traces are more natural when reasoning about action and change. As a result, our approach is readily implementable via multi-shot ASP systems and benefits from an extension of ASP's full-fledged input language with temporal operators. This includes future as well as past operators whose combination offers a rich temporal modeling language. For computation, we identify the class of temporal logic programs and prove that it constitutes a normal form for our approach. Finally, we outline two implementations, a generic one and an extension of the ASP system clingo.

The recent framework of compressive statistical learning aims at designing tractable learning algorithms that use only a heavily compressed representation-or sketch-of massive datasets. Compressive K-Means (CKM) is such a method: it estimates the centroids of data clusters from pooled, non-linear, random signatures of the learning examples. While this approach significantly reduces computational time on very large datasets, its digital implementation wastes acquisition resources because the learning examples are compressed only after the sensing stage. The present work generalizes the sketching procedure initially defined in Compressive K-Means to a large class of periodic nonlinearities including hardware-friendly implementations that compressively acquire entire datasets. This idea is exemplified in a Quantized Compressive K-Means procedure, a variant of CKM that leverages 1-bit universal quantization (i.e. retaining the least significant bit of a standard uniform quantizer) as the periodic sketch nonlinearity. Trading for this resource-efficient signature (standard in most acquisition schemes) has almost no impact on the clustering performances, as illustrated by numerical experiments.

Sound event detection (SED) methods are tasked with labeling segments of audio recordings by the presence of active sound sources. SED is typically posed as a supervised machine learning problem, requiring strong annotations for the presence or absence of each sound source at every time instant within the recording. However, strong annotations of this type are both labor- and cost-intensive for human annotators to produce, which limits the practical scalability of SED methods. In this work, we treat SED as a multiple instance learning (MIL) problem, where training labels are static over a short excerpt, indicating the presence or absence of sound sources but not their temporal locality. The models, however, must still produce temporally dynamic predictions, which must be aggregated (pooled) when comparing against static labels during training. To facilitate this aggregation, we develop a family of adaptive pooling operators---referred to as auto-pool---which smoothly interpolate between common pooling operators, such as min-, max-, or average-pooling, and automatically adapt to the characteristics of the sound sources in question. We evaluate the proposed pooling operators on three datasets, and demonstrate that in each case, the proposed methods outperform non-adaptive pooling operators for static prediction, and nearly match the performance of models trained with strong, dynamic annotations. The proposed method is evaluated in conjunction with convolutional neural networks, but can be readily applied to any differentiable model for time-series label prediction.

Crowdsourcing has become very popular among the machine learning community as a way to obtain labels that allow a ground truth to be estimated for a given dataset. In most of the approaches that use crowdsourced labels, annotators are asked to provide, for each presented instance, a single class label. Such a request could be inefficient, that is, considering that the labelers may not be experts, that way to proceed could fail to take real advantage of the knowledge of the labelers. In this paper, the use of weak labeling for crowd learning is proposed, where the annotators may provide more than a single label per instance to try not to miss the real label. The main hypothesis is that, by allowing weak labeling, knowledge can be extracted from the labelers more efficiently by than in the standard crowd learning scenario. Empirical evidence which supports that hypothesis is presented.

We investigate deep neural network performance in the textindependent speaker recognition task. We demonstrate that using angular softmax activation at the last classification layer of a classification neural network instead of a simple softmax activation allows to train a more generalized discriminative speaker embedding extractor. Cosine similarity is an effective metric for speaker verification in this embedding space. We also address the problem of choosing an architecture for the extractor. We found that deep networks with residual frame level connections outperform wide but relatively shallow architectures. This paper also proposes several improvements for previous DNN-based extractor systems to increase the speaker recognition accuracy. We show that the discriminatively trained similarity metric learning approach outperforms the standard LDA-PLDA method as an embedding backend. The results obtained on Speakers in the Wild and NIST SRE 2016 evaluation sets demonstrate robustness of the proposed systems when dealing with close to real-life conditions.

This paper describes the methodology followed to build a neural machine translation system in the biomedical domain for the English-Catalan language pair. This task can be considered a low-resourced task from the point of view of the domain and the language pair. To face this task, this paper reports experiments on a cascade pivot strategy through Spanish for the neural machine translation using the English-Spanish SCIELO and Spanish-Catalan El Peri\'odico database. To test the final performance of the system, we have created a new test data set for English-Catalan in the biomedical domain which is freely available on request.