MeetAI London and NeurotechX want to join efforts and bring together a selected panel of experts in diverse aspects of machine learning and neuroscience. An open discussion centred around how this two fields work together, the current achievements, and the future goals and limitations. Neuroscience and artificial intelligence are heavily related and both are living a golden age. Machine learning has been inspired by the nervous systems since its first steps. Terms such as neural networks or reinforcement learning have been borrowed from natural sciences and translated into silicon.

Can you get pregnant using Artificial Intelligence? One American startup is trying to make it so. Mira Care, which SHACK15 News met up at CES 2018, has designed a gadget that harnesses machine learning to track hormone levels, and suggest how likely a user is to conceive a baby at a given time. Users just need to pee on a test stick and insert it in the device; then, Mira's algorithm will blend data from the sample with information about each user's fitness and life habits in order to calculate a fertility score, which will be shown on a mobile app. Mira's AI-powered device and its linked app showing a user's fertility score (via Mira Care) "One difference with any other menstrual cycle tracker is personalization," company CEO Zheng Yang said."It China-born Yang, who earned a PhD in Physics and for ten years has tried to apply mathematical and physical models to biomedical problems, also underlined that the tool and the algorithm have been tested in official clinical trials-- over 200 cases in total-- which have pegged Mira's has a 99 percent accuracy at predicting one person's chances of conceiving a baby. Yang said that the device--which has been approved by US medical authority FDA--is going to be launched on the US market in the second quarter of 2018 (and in Europe shortly after). He added that Mira is going to be followed by a string of similar AI-fuelled home diagnostic gadgets. "We are working on a product line that focuses on several diseases: like infectious diseases, chronic diseases, kidney function,"he told SHACK15 News."We're

The rapidly advancing area of artificial intelligence will require a new field of law and new regulations governing a growing pool of businesses involved, according to Microsoft Corp, a 25-year participant in AI research. Companies making and selling AI software will need to be held responsible for potential harm caused by'unreasonable practices' – if a self-driving car program is set up in an unsafe manner that causes injury or death, for example, Microsoft said. And as AI and automation boost the number of laborers in the gig-economy or on-demand jobs, Microsoft said technology companies need to take responsibility and advocate for protections and benefits for workers, rather than passing the buck by claiming to be'just the technology platform' enabling all this change. Microsoft broaches these ideas in a 149-page book entitled "The Future Computed," which will also be the subject of a panel at the World Economic Forum in Davos, Switzerland, next week. As Redmond, Washington-based Microsoft seeks to be a leader in AI and automating work tasks, it's also trying to get out in front of the challenges expected to arise from promising new technologies, such as job losses and everyday citizens who may be hurt or disadvantaged by malfunctioning or biased algorithms.

Ever since early-nineteenth-century textile workers destroyed the mechanical looms that threatened their livelihoods, debates over automation have conjured gloom-and-doom scenarios about the future of work. With another era of automation upon us, how nervous about the future of our own livelihoods should we be? A recent report by the McKinsey Global Institute estimates that depending on a country's level of development, advances in automation will require 3 to 14 per cent of workers worldwide to change occupations or upgrade their skills by the year 2030. Already, about 10 per cent of all jobs in Europe have disappeared since 1990 during the first wave of routine-based technological change. And with advances in artificial intelligence (AI), which affects a broader range of tasks, that share could double in the coming years.

The surge in robot sales has seen the emergence of four major suppliers, two Japanese, Fanuc and Yaskawa, a Swiss/Swedish concern ABB and Germany's Kuka AG. The rise in robot demand has coincided with a jump in their share prices. Kuka made the news last year not because its robots were building Tesla and Porsche cars, but for its €4.5bn takeover by the Chinese appliance company Medea, which hopes to build small mobile robots for the home and consumer industry. However, the German government was unhappy with the takeover. While it has a right to block any non-EU company from acquiring more than a 25pc stake in any German entity, it is limited to public order being endangered or national security.

People are increasingly reliant on artificial intelligence (AI) -- that is, the machines, systems or applications that are capable of performing tasks that, until recently, could only be performed by a human. Think of your morning routine: maybe a Google Assistant checks your calendar and reminds you of your meetings. Then you survey Twitter, which uses algorithms to curate what you see -- the latest about Trump, trade and technology rise to the top. And at the end of it all, when you settle in for some Netflix, your profile suggests a few thrillers you're likely to binge-watch. Marketing statistics reveal that some 57 percent of consumers expect voice-activated smart assistants to have a major or moderate impact on their daily lives by 2020.

Gaussian process for vector-valued function model has been shown to be a useful method for multi-output prediction. The existing method for this model is to re-formulate the matrix-variate Gaussian distribution as a multivariate normal distribution. Although it is effective in many cases, re-formulation is not always workable and difficult to extend because not all matrix-variate distributions can be transformed to related multivariate distributions, such as the case for matrix-variate Student$-t$ distribution. In this paper, we propose a new derivation of multivariate Gaussian process regression (MV-GPR), where the model settings, derivations and computations are all directly performed in matrix form, rather than vectorizing the matrices as done in the existing methods. Furthermore, we introduce the multivariate Student$-t$ process and then derive a new method, multivariate Student$-t$ process regression (MV-TPR) for multi-output prediction. Both MV-GPR and MV-TPR have closed-form expressions for the marginal likelihoods and predictive distributions. The usefulness of the proposed methods is illustrated through several simulated examples. In particular, we verify empirically that MV-TPR has superiority for the datasets considered, including air quality prediction and bike rent prediction. At last, the proposed methods are shown to produce profitable investment strategies in the stock markets.

Dictionary learning and component analysis models are fundamental in learning compact representations that are relevant to a given task (feature extraction, dimensionality reduction, denoising, etc.). The model complexity is encoded by means of specific structure, such as sparsity, low-rankness, or nonnegativity. Unfortunately, approaches like K-SVD - that learn dictionaries for sparse coding via Singular Value Decomposition (SVD) - are hard to scale to high-volume and high-dimensional visual data, and fragile in the presence of outliers. Conversely, robust component analysis methods such as the Robust Principle Component Analysis (RPCA) are able to recover low-complexity (e.g., low-rank) representations from data corrupted with noise of unknown magnitude and support, but do not provide a dictionary that respects the structure of the data (e.g., images), and also involve expensive computations. In this paper, we propose a novel Kronecker-decomposable component analysis model, coined as Robust Kronecker Component Analysis (RKCA), that combines ideas from sparse dictionary learning and robust component analysis. RKCA has several appealing properties, including robustness to gross corruption; it can be used for low-rank modeling, and leverages separability to solve significantly smaller problems. We design an efficient learning algorithm by drawing links with a restricted form of tensor factorization, and analyze its optimality and low-rankness properties. The effectiveness of the proposed approach is demonstrated on real-world applications, namely background subtraction and image denoising and completion, by performing a thorough comparison with the current state of the art.

This chapter provides a self-contained introduction to the use of Bayesian inference to extract large-scale modular structures from network data, based on the stochastic block model (SBM), as well as its degree-corrected and overlapping generalizations. We focus on nonparametric formulations that allow their inference in a manner that prevents overfitting, and enables model selection. We discuss aspects of the choice of priors, in particular how to avoid underfitting via increased Bayesian hierarchies, and we contrast the task of sampling network partitions from the posterior distribution with finding the single point estimate that maximizes it, while describing efficient algorithms to perform either one. We also show how inferring the SBM can be used to predict missing and spurious links, and shed light on the fundamental limitations of the detectability of modular structures in networks.

A principled approach to characterize the hidden structure of networks is to formulate generative models, and then infer their parameters from data. When the desired structure is composed of modules or "communities", a suitable choice for this task is the stochastic block model (SBM), where nodes are divided into groups, and the placement of edges is conditioned on the group memberships. Here, we present a nonparametric Bayesian method to infer the modular structure of empirical networks, including the number of modules and their hierarchical organization. We focus on a microcanonical variant of the SBM, where the structure is imposed via hard constraints, i.e. the generated networks are not allowed to violate the patterns imposed by the model. We show how this simple model variation allows simultaneously for two important improvements over more traditional inference approaches: 1. Deeper Bayesian hierarchies, with noninformative priors replaced by sequences of priors and hyperpriors, that not only remove limitations that seriously degrade the inference on large networks, but also reveal structures at multiple scales; 2. A very efficient inference algorithm that scales well not only for networks with a large number of nodes and edges, but also with an unlimited number of modules. We show also how this approach can be used to sample modular hierarchies from the posterior distribution, as well as to perform model selection. We discuss and analyze the differences between sampling from the posterior and simply finding the single parameter estimate that maximizes it. Furthermore, we expose a direct equivalence between our microcanonical approach and alternative derivations based on the canonical SBM.