Those worried that the rise of artificial intelligence means that robots will take their job might feel comforted by the fact that many AI tools are actually being designed not to replace humans, but to help them do their jobs better. Though the field is still in its infancy, many young startups came to Europe's largest tech conference, Web Summit, last week to showcase how their AI tools are working to make people more efficient and productive, in both their personal and professional lives. Here are a few that stood out. Does it feel like you've never got enough time for the things you really want to do in life? Paris-based AI startup Smarter Time is helping its 80,000 users find the time that always seems to be missing by tracking their habits and providing feedback.

The European real estate sector continues to flourish in regions such as the UK and Germany, despite strong and unpredictable economic and political headwinds. Successful transactions depend on high quality and detailed due diligence, but competition for the most lucrative deals can sometimes lead organisations to compromise on this stage of the process. The biggest challenge is that the size of real estate transactions is increasing exponentially because of regulatory and compliance requirements and also because of the broader volumes and types of documents involved. This means that more manual processes are required simply to find the right data. The trend for higher volumes and larger transactions in real estate, including higher levels of risk and multiple languages, has important implications for the way in which investment professionals manage the greater complexity of due diligence.

Natural Language Processing (NLP) comprises a set of techniques that can be used to achieve many different objectives. Take a look at the following table to figure out which technique can solve your particular problem. We are going to talk about parsing in the general sense of analyzing a document and extracting its meaning. So, we are going to talk about actual parsing of natural languages, but we will spend most of the time on other techniques. When it comes to understanding programming languages parsing is the way to go, however you can pick specific alternatives for natural languages. In other words, we are mostly going to talk about what you would use instead of parsing, to accomplish your goals. For instance, if you wanted to find all for statements a programming language file, you would parse it and then count the number of for. Instead, you are probably going to use something like stemming to find all mentions of cats in a natural language document. This is necessary because the theory behind the parsing of natural languages might be the same one that is behind the parsing of programming languages, however the practice is very dissimilar. In fact, you are not going to build a parser for a natural language. That is unless you work in artificial intelligence or as researcher. You are even rarely going to use one. Rather you are going to find an algorithm that work a simplified model of the document that can only solve your specific problem. In short, you are going to find tricks to avoid to actually having to parse a natural language. That is why this area of computer science is usually called natural language processing rather than natural language parsing. Now check your email to confirm your subscription. There was an error submitting your subscription. I'd like to learn more about NLP and language engineering We are going to see specific solutions to each problem. Mind you that these specific solutions can be quite complex themselves. The more advanced they are, the less they rely on simple algorithms. Usually they need a vast database of data about the language. A logical consequence of this is that it is rarely easy to adopt a tool for one language to be used for another one. Or rather, the tool might work with few adaptations, but to build the database would require a lot of investment. So, for example, you would probably find a ready to use tool to create a summary of an English text, but maybe not one for an Italian one.

When you upload photos to Facebook, have you noticed that the website already seems to know who's in them? It's remarkable, and you can give the credit to big data. Face recognition software, like fraud detection and ad matching algorithms, draws on deep libraries of content in order to deliver the correct results. And these data collections are hard at work across the web and in many of your favorite apps. It comes as no surprise that developers have been hard at work on face recognition software since it's an integral part of security programs.

Whether it's getting your partner to do more housework or making your colleagues back your latest idea, we all end up spending a considerable amount of time trying to persuade or even manipulate others. So can science offer any clever tricks to get people to do what we want, without resorting to bullying them? In an article for The Conversation, Dr Harriet Dempsey-Jones, a postdoctoral researcher in Cognitive Neurosciences at University of Oxford explains the methods worth trying. Whether it's getting your partner to do more housework or making your colleagues back your latest idea, we all end up spending a considerable amount of time trying to persuade or even manipulate others (stock image) 'Misattribution of arousal' is a popular theory in social psychology that suggests people sometimes mislabel feelings from their body. It involves manipulating individuals into thinking they are experiencing particular emotions, such as believing they are attracted when they're actually scared, Somewhat counter-intuitively, if you want to get something from someone – you should give them something yourself.

The technology is a form of deep-learning artificial intelligence software developed to fit onto mobile computer chips. This allows artificial intelligence to be used in a range of devices, from smartphones to industrial robots. This portability would enable devices to operate independent of the Internet while using artificial intelligence that performs equivalent to tethered neural networks. With this, a hosting chip embedded in a smartphone could run a speech-activated virtual assistant and undertake other intelligent features, such as controlling data usage. Other applications include operating drones and surveillance cameras in remote areas.

Russia's Ministry of Defence has posted what it called "irrefutable proof" of the US aiding so-called Islamic State - but one of the images was actually taken from a video game. The ministry claimed the image showed an IS convoy leaving a Syrian town last week aided by US forces. Instead, it came from the smartphone game AC-130 Gunship Simulator: Special Ops Squadron. The ministry said an employee had mistakenly attached the photo. The Conflict Intelligence Team fact-checking group said the other four provided were also errors, taken from a June 2016 video which showed the Iraqi Air Force attacking IS in Iraq.

Lexical Simplification is the process of replacing complex words in a given sentence with simpler alternatives of equivalent meaning. This task has wide applicability both as an assistive technology for readers with cognitive impairments or disabilities, such as Dyslexia and Aphasia, and as a pre-processing tool for other Natural Language Processing tasks, such as machine translation and summarisation. The problem is commonly framed as a pipeline of four steps: the identification of complex words, the generation of substitution candidates, the selection of those candidates that fit the context, and the ranking of the selected substitutes according to their simplicity. In this survey we review the literature for each step in this typical Lexical Simplification pipeline and provide a benchmarking of existing approaches for these steps on publicly available datasets. We also provide pointers for datasets and resources available for the task.

Testing (conditional) independence of multivariate random variables is a task central to statistical inference and modelling in general - though unfortunately one for which to date there does not exist a practicable workflow. State-of-art workflows suffer from the need for heuristic or subjective manual choices, high computational complexity, or strong parametric assumptions. We address these problems by establishing a theoretical link between multivariate/conditional independence testing, and model comparison in the multivariate predictive modelling aka supervised learning task. This link allows advances in the extensively studied supervised learning workflow to be directly transferred to independence testing workflows - including automated tuning of machine learning type which addresses the need for a heuristic choice, the ability to quantitatively trade-off computational demand with accuracy, and the modern black-box philosophy for checking and interfacing. As a practical implementation of this link between the two workflows, we present a python package 'pcit', which implements our novel multivariate and conditional independence tests, interfacing the supervised learning API of the scikit-learn package. Theory and package also allow for straightforward independence test based learning of graphical model structure. We empirically show that our proposed predictive independence test outperform or are on par to current practice, and the derived graphical model structure learning algorithms asymptotically recover the 'true' graph. This paper, and the 'pcit' package accompanying it, thus provide powerful, scalable, generalizable, and easy-to-use methods for multivariate and conditional independence testing, as well as for graphical model structure learning.

Many modern unsupervised or semi-supervised machine learning algorithms rely on Bayesian probabilistic models. These models are usually intractable and thus require approximate inference. Variational inference (VI) lets us approximate a high-dimensional Bayesian posterior with a simpler variational distribution by solving an optimization problem. This approach has been successfully used in various models and large-scale applications. In this review, we give an overview of recent trends in variational inference. We first introduce standard mean field variational inference, then review recent advances focusing on the following aspects: (a) scalable VI, which includes stochastic approximations, (b) generic VI, which extends the applicability of VI to a large class of otherwise intractable models, such as non-conjugate models, (c) accurate VI, which includes variational models beyond the mean field approximation or with atypical divergences, and (d) amortized VI, which implements the inference over local latent variables with inference networks. Finally, we provide a summary of promising future research directions.