If the design principles of Aeron were applied to Raft, could you create a high-performance consensus implementation?

Come to this talk if you would like to find out what happened when Aeron met Raft.

This talk highlights a path and strategies how to start with a small full stack application that is enabled to grow into a comprehensive, highly available and resilient application handling millions of QPS. While we walk through the different stages, cost performance is always at the top of our minds. Through live demos and metrics, attendees will learn which trade-offs to consider between cost, performance, reliability and scalability.

Alison’s talk will introduce the hardware and software platform at the heart of this Intelligent Industrial Revolution: NVIDIA GPU Computing. She’ll provide insights into how academia, enterprise and startups are applying AI, as well as offer a glimpse into state-of-the-art research from world-wide labs & internally at NVIDIA, for example combining robotics with VR and AI in an end-to-end simulator to train intelligent machines.

Beginners might like to try our free online 40-minute class using GPU’s in Amazon Web Services

In particular with logistic regression using application, behavioural and bureau data, has become an industry standard accepted as an accurate credit risk assessment tool. Recently, however, financial companies have begun considering how Big Data and advanced Data Science techniques (such as neural networks and GBM) could improve their credit decisioning models. The obstacles that must be overcome before Data Science can be used in this way are both significant and varied (including technological, cultural and legal obstacles).

In this talk Roberta will explain how Data Science can work with financial companies to overcome these obstacles and transform their credit decisioning models.

She will then explain the underlying principles of how do we teach computers ‘to see’ using machine learning methods with examples of recognising objects in visual data.

Subsequently Dr Sharmanska will focus on how can you enhance an automatic systems’ capacity to recognise complex real-world scenes like human actions and emotions using a deep-learning approach.

In this talk, Oliver will discuss how some of these novel models can be implemented in the machine learning framework TensorFlow, starting from a collaborative filtering approach and extending that to more complex deep recommender systems.

Building from that you will delve into related problems of sampling in a stream setting, and updating in a batch setting, and highlight some cool tricks such as capturing time-dynamics via data snapshotting. To finish off we will touch upon algorithms to summarize categorical data, most notably count-min sketch. The talk is motivated by my work at Metabase -- an opensoucre analytics tool -- where you heavily utilize histograms in building our "data scientist in a box".

Nevertheless, a fundamental challenge remains largely open: how to digitise and make accessible the vast and scattered cultural heritage which informs both research and public experiences. In this talk we will consider these challenges through the lenses of the Venice Time Machine: a large-scale digitisation and indexation project based at the Archive of Venice, with the ambition to make accessible the evidence on the history of a single city through the application of AI technologies.

The broadening application of AI to cultural heritage is reaching a momentum currently epitomised by the timemachineproject.eu: a European Flagship proposal whose ambition is to create multiple time machines across Europe.

A case will be made for the possibility and ultimate opportunities opened by big science projects in the digital humanities, and the resulting datafication of cultural heritage.

Covering models, dashboards, and other products built in R, you'll look at how Docker containers can make managing dependencies a breeze, allow your code to be hosted anywhere, and have it work in high-scale systems.

Many modules have matured since then, and he'll walk you through the most interesting once you can leverage in your next generation application. He'll also touch upon some of the lesser know Redis Enterprise features for geo distributed big data sets. (Not to be missed if you find live demo's entertaining)

She'll highlight her misunderstandings which lead to a poor-performing cluster in the first place, and then summarise her findings as positive and actionable pieces of advice to follow when setting up your own cluster.

However, this does not come without its challenges. For example, the replication crisis faced by Cognitive Neuroscience, along with other biological sciences, is a very real issue whereby independent researchers fail to replicate findings from publications using the same experimental design.

This issue is attributed to many reasons, such as low statistical power and differences in analytical methods. The advent of big data offers solutions to these challenges. Online platforms designed specifically to allow neuroimaging data and experimental tasks to be shared between researchers is promoting collaboration across the discipline. Such collaboration will make research more accessible for under-funded research groups, could lead to greater statistically-powered findings, and help build a greater understanding of what is happening in the brain.

Alternatively, the use of large-scale behavioural tasks through online platforms such as Amazon Mechanical Turk or on smartphones has surged in popularity in recent years. By developing complex decision-making tasks that people can perform in their homes, researchers can incorporate natural features from everyday life in their analyses, allowing us to understand cognitions and brain functions relevant for real life.

In this talk, Jak will review some of the challenges faced by the Cognitive Neuroscience discipline, and the subsequent solutions that open sourcing and big data can help to reveal. He will then discuss an alternative view, outlining how big data from large-scale behavioural tasks performed in naturalistic environments has the potential to further benefit the discipline.

Chanuki will talk to you about how she exploited recent advances in computer vision and deep learning to attempt to find answers to this question that go beyond the simple explanation ‘what is natural is beautiful’. She will also explore a new deep learning model to predict scenic locations around London.

Chanuki's research shows that beauty – once thought to be in the eye of the beholder and thus an area of investigation impenetrable by computers – can in fact be discerned by computer algorithms.

The Universities UK #StepChange report calls for the alignment of learning analytics to student wellbeing, specifically to identify changes in students' behaviours. From research into the analysis of social media behaviours and message sentiment, it is possible to identify the onset of a depressive period. This is beyond the scope of our role, but is it possible to identify student behaviours that are of concern?

In this session, Samantha will explore why this is deemed necessary, how learning analytics may be used to identify these behaviours and what the ethical and legal issues may be in doing so.

You will cover all steps of the workflow:

• Spinning up your Spark instance on AWS
• Trimming and cleaning data
• Using different storage formats for faster handling
• Browsing subsets of the data to get a feel for which features might be the most useful
• Application of models from the SparkMLliband Scikit-learn libraries
• Viewing results and assessing the quality of our predictions

In this session, Hollie will explore how you can address fairness, accountability and the long-term effects on society when designing with data.

Technology should be a beneficial force in everyone's lives, taking the world in exciting new directions and making everybody better humans. To ensure this, you need to facilitate a conversation between data technology and the human experience.

Keeping social responsibility and ethical behaviour in mind when designing AI systems will allow you to put the right systems in place to contribute to the society you want, fostering higher levels of cognitive and emotional skills.

You will cover all steps of the workflow:

• Spinning up your Spark instance on AWS
• Trimming and cleaning data
• Using different storage formats for faster handling
• Browsing subsets of the data to get a feel for which features might be the most useful
• Application of models from the SparkMLliband Scikit-learn libraries
• Viewing results and assessing the quality of our predictions

The goal was to detect if there was something in the gene mapping of patients which could tell researchers and healthcare professionals how the patient will react to the treatment. In this way, they can prevent sending patients to a very intensive type of treatment (which can lead to numerous negative long-term side effects), if they know the patients are likely to react well to lower-intensity treatment.

In this workshop, you’ll learn how to leverage Kubernetes to deploying complex workloads in the cloud, on bare metal and locally. You’ll learn how Kubernetes provides a fast iteration cycle, flexible scalability, and a lack of boilerplate which makes it ideal for most of the machine learning experiments.

Please install the following software on your computer:

• - Kubectl
• - Docker
• - Bash (please install Cmder if you're on Windows)
• - Git
• - Python 2.7 and pip
• - A text editor such as Atom

You should also create an account on Docker Hub.

To address this, Rose will propose a set of recommendations for how educators can draw on what AI does best to nurture and expand our human capabilities.

In this workshop, you’ll learn how to leverage Kubernetes to deploying complex workloads in the cloud, on bare metal and locally. You’ll learn how Kubernetes provides a fast iteration cycle, flexible scalability, and a lack of boilerplate which makes it ideal for most of the machine learning experiments.

Please install the following software on your computer:

• - Kubectl
• - Docker
• - Bash (please install Cmder if you're on Windows)
• - Git
• - Python 2.7 and pip
• - A text editor such as Atom

You should also create an account on Docker Hub.

Now, as our production systems become more distributed geographically, we can push some ML capabilities to the edge.

Imagine:

• Reusing existing models in distributed and decentralized architectures
• Image and object detection from local video streams
• Event classification from IoT-based sensors
• Interact with WebAssembly-optimized code on all browsers and platforms
• Natural Language Processing of user text and audio

If models aren’t too big and can be cached, the Web architecture will happily support these new use cases.

In this lighting talk, Tom investigates the underlying motives of a company striving to become data-driven and how these may be translated into more tangible goals.