Machine learning has promised to revolutionise healthcare for several years [1, 2]. Moreover, while there is an extensive literature describing high-performing machine learning models trained on immaculate benchmark datasets [3-5], such promising approaches rarely make it into clinical practice [6]. Often, this is because of an unexpected drop in performance when deploying the model on unseen test data due to domain shift [7, 8], i.e. there is a change in the data distribution between the dataset a model is trained on (source data) and that which it is deployed against (target data). Most common machine learning algorithms rely on an assumption that the source and target data are independent and identically distributed (i.i.d.) [9]. However, with domain shift, this assumption no longer holds, and model performance can be significantly affected. For medical datasets, domain shift is widespread, resulting from differences in equipment and clinical practice between sites [10-13], and models are vulnerable to associating clinically irrelevant features specific to the domain with their predictions, known as shortcut learning [14], which may lead to poor performance on target data. For most medical applications, target data is rarely available prior to real-time deployment; thus, a domain adaptation approach, where pre-trained models are fine-tuned on data from the target distribution is not feasible.

We derive criteria for training adaptive classifier networks to perform unsu(cid:173) pervised data analysis. The first criterion turns a simple Gaussian classifier into a simple Gaussian mixture analyser. The second criterion, which is much more generally applicable, is based on mutual information. This'firm but fair' criterion can be applied to any network that produces probability-type outputs, but it does not necessarily lead to useful behavior. One of the main distinctions made in discussing neural network architectures, and pattern analysis algorithms generally, is between supervised and unsupervised data analysis.

Founder Adam Smith said his business failed to take off because current state-of-the-art machine learning models'don't understand the structure of code'. Kite, a start-up that has been developing artificial intelligence technology to help developers write code for nearly a decade, is sunsetting its business. Based in San Francisco, Kite was founded in 2014 as an early pioneer in the emerging field of AI that assists software developers in writing code – an'autocomplete' for programming of sorts. But now, after eight years of pursuing its vision to be a leader in AI-assisted programming, founder Adam Smith announced on the company website that the business is now wrapping up. "From 2014 to 2021, Kite was a start-up using AI to help developers write code. We have stopped working on Kite and are no longer supporting the Kite software," Smith wrote.

Graphcore has released significant new features for the PopVisionTM family of analysis tools as part of a major Poplar software stack update, Poplar SDK 2.0. We created the PopVision Graph Analyser and System Analyser to help developers maximise the performance of their applications on IPU systems. To mark this update, we are looking at how PopVision tools can be used most effectively to inform and optimise machine learning programs. With its massively parallel architecture, the IPU has been specifically built from the ground up for machine intelligence workloads and is therefore able to deliver state of the art performance across even the most complex AI models. For this reason, many of our users are usually not just looking to run standard machine intelligence models, but to exploit the highest possible performance from IPU systems, beyond what they have been able to achieve with other systems.

A breathalyser that spots Covid-19, a robotic massage table, and an ultraportable electric wheelchair are among the many bizarre inventions unveiled at CES 2022. The event is being held in person in Las Vegas, Nevada, as well as online for people who can't travel - with tens of thousands of ideas, concepts and products on show. The massage table, by Massage Robotics, has two arms and responds to verbal commands in real time, but it has a whopping $310,000 (£228,000) price tag, and was just one of a number of relaxation devices on show at CES this year. Covid-19 is present throughout the event, including in the absence of some major companies such as Amazon and Google, but it is also present in the purpose of a number of products on display, including a breath analyser that detects the virus. The annual event runs until Saturday, and MailOnline has created this roundup of some of the weird and wonderful inventions revealed by firms large and small.

Most of the time organisations have to deal with tough problems like errors, defects, and other complexities while developing intricate software. This is where self-coded applications come into play. Applications that generate the code themselves not only help the programmers to accomplish a task in less time but also increase the programming ability of the developer. Bayou is a system for generating API idioms which are the snippets of code that use APIs in Java. The main task of this system is to use the user's code and the query in order to generate the appropriate program which will most likely solve the task.

Many of the problems of recognition and interpretation encountered in archaeology have close parallels with classic artificial intelligence problems, notably those of scene analysis.

This paper proposes a design for a system to generate constraint solvers that are specialised for specific problem models. It describes the design in detail and gives preliminary experimental results showing the feasibility and effectiveness of the approach.