The proliferation of computing devices has brought about an opportunity to deploy machine learning models on new problem domains using previously inaccessible data. Traditional algorithms for training such models often require data to be stored on a single machine with compute performed by a single node, making them unsuitable for decentralised training on multiple devices. This deficiency has motivated the development of federated learning algorithms, which allow multiple data owners to train collaboratively and use a shared model whilst keeping local data private. However, many of these algorithms focus on obtaining point estimates of model parameters, rather than probabilistic estimates capable of capturing model uncertainty, which is essential in many applications. Variational inference (VI) has become the method of choice for fitting many modern probabilistic models. In this paper we introduce partitioned variational inference (PVI), a general framework for performing VI in the federated setting. We develop new supporting theory for PVI, demonstrating a number of properties that make it an attractive choice for practitioners; use PVI to unify a wealth of fragmented, yet related literature; and provide empirical results that showcase the effectiveness of PVI in a variety of federated settings.

The vast list of artificial intelligence applications continually increases as researchers, technologists, and scientists try to leverage computing power to gain competitive edges over the more slowly adopting set. Today I want to traipse across the American business and tech landscape and present a few of the new and hopefully intriguing upgrades of these mostly familiar devices and services being brought into the 21st century via AI. First a quick overview of the concept of AI and where it's come from over the past years and decades. Earlier writings comingled two phrases to identify the technology: artificial intelligence, which has become the well-known marketable way to talk about the tech, and "computational intelligence," which might be useful amongst a group of AI--err, CI?--subject matter experts, but doesn't carry the cachet of its more widely accepted phrase. For anyone who uses either phrase, it's generally understood to refer to some sort of machine-based intelligence.

The Advancing Data Justice project is a Global Partnership on AI (GPAI) initiative, led by The Alan Turing Institute. Researchers from the Institute have been collaborating with twelve Policy Pilot Partner organisations from Asia, Oceania, Africa, and South America, and each of these have been working to understand what data justice might look like in their distinct contexts. The aims of the project are 1) to gain a better understanding of the current state of research in the field to better inform future research directions, and 2) to create a guide for policymakers, developers, and communities affected by AI, comprising advice on what they should consider in their practice, use and experience of AI systems. As part of the project, the team have recently launched the first instalment of a documentary series which tracks the work of the project partners. They discuss how data-driven technologies can be deployed in a way which is compatible with values of social justice.

RAMSTEIN AIR BASE, Germany – The United States marshaled 40 allies on Tuesday to furnish Ukraine with long-term military aid in what could become a protracted battle against the Russian invasion, and Germany said it would send dozens of armored anti-aircraft vehicles. It was a major policy shift for a country that had wavered over fear of provoking Russia. The announcement by Germany, Europe's biggest economy and one of Russia's most important Western trading partners, was among many signals on Tuesday pointing to further escalation in the war and disappointment for diplomacy. Germany's shift on weapons also was seen as a strong affirmation of a toughened message by the administration of U.S. President Joe Biden, which has said it wants to see Russia not only defeated in Ukraine but seriously weakened from the conflict that Russian President Vladimir Putin began two months ago. The increasing flow of Western weapons into Ukraine -- including howitzers, armed drones, tanks and ammunition -- also amounted to another sign that a war Putin had expected would divide his Western adversaries had instead drawn them much closer together.

Jerry Levine is Chief Evangelist & General Counsel at ContractPodAi. He helps guide global client success and shape overall product vision. It was 25 years ago when IBM's artificial intelligence system, Deep Blue, defeated Garry Kasparov in a six-game rematch of chess. But this competition did not reveal AI to be smarter than its human opponent, who was at the time the reigning world champion; Deep Blue's success demonstrated that we, humans, could program AI to perform functions we cannot do quickly on our own--analyzing vast amounts of data and processing any number of natural languages, just to name a couple of functions. Today, AI continues to attract more attention and interest than most other innovations, including when it comes to nonfungible tokens (NFTS).

Many real-world multi-objective optimisation problems rely on computationally expensive function evaluations. Multi-objective Bayesian optimisation (BO) can be used to alleviate the computation time to find an approximated set of Pareto optimal solutions. In many real-world problems, a decision-maker has some preferences on the objective functions. One approach to incorporate the preferences in multi-objective BO is to use a scalarising function and build a single surrogate model (mono-surrogate approach) on it. This approach has two major limitations. Firstly, the fitness landscape of the scalarising function and the objective functions may not be similar. Secondly, the approach assumes that the scalarising function distribution is Gaussian, and thus a closed-form expression of an acquisition function e.g., expected improvement can be used. We overcome these limitations by building independent surrogate models (multi-surrogate approach) on each objective function and show that the distribution of the scalarising function is not Gaussian. We approximate the distribution using Generalised value distribution. We present an a-priori multi-surrogate approach to incorporate the desirable objective function values (or reference point) as the preferences of a decision-maker in multi-objective BO. The results and comparison with the existing mono-surrogate approach on benchmark and real-world optimisation problems show the potential of the proposed approach.

Hypernymy plays a fundamental role in many AI tasks like taxonomy learning, ontology learning, etc. This has motivated the development of many automatic identification methods for extracting this relation, most of which rely on word distribution. We present a novel model HyperBox to learn box embeddings for hypernym discovery. Given an input term, HyperBox retrieves its suitable hypernym from a target corpus. For this task, we use the dataset published for SemEval 2018 Shared Task on Hypernym Discovery. We compare the performance of our model on two specific domains of knowledge: medical and music. Experimentally, we show that our model outperforms existing methods on the majority of the evaluation metrics. Moreover, our model generalize well over unseen hypernymy pairs using only a small set of training data.

A long time ago in the machine learning literature, the idea of incorporating a mechanism inspired by the human visual system into neural networks was introduced. This idea is named the attention mechanism, and it has gone through a long development period. Today, many works have been devoted to this idea in a variety of tasks. Remarkable performance has recently been demonstrated. The goal of this paper is to provide an overview from the early work on searching for ways to implement attention idea with neural networks until the recent trends. This review emphasizes the important milestones during this progress regarding different tasks. By this way, this study aims to provide a road map for researchers to explore the current development and get inspired for novel approaches beyond the attention.

The idea of a hub for drones and Jetsons-style flying taxis may sound like a concept from the latest science fiction blockbuster. But such a hub – known as a vertiport – became a reality this week, with the opening of the Air-One site in Coventry. Air-One is a centre for drones and electric vertical take-off and landing (eVTOL) aircraft, including flying cars and taxis, and will serve as a blueprint for more than 200 vertiports planned worldwide over the next five years, according to Urban-Air Port, the UK-based developer of Air-One. MailOnline attended the opening of the Air-One vertiport this week for a look at what the future of transport could be. From a central rising launch pad from which flying taxis could shuttle passengers on their journeys, to a futuristic vending machine where you can order items to be delivered by drones on demand, here's a look at the key features.

The ACM constitution provides that our Association hold a general election in the even-numbered years for the positions of President, Vice President, Secretary/Treasurer, and Members-at-Large. Biographical information and statements of the candidates appear on the following pages (candidates' names appear in random order). In addition to the election of ACM's officers--President, Vice President, Secretary/Treasurer--two Members-at-Large will be elected to serve on ACM Council. The 2022 candidates for ACM President, Yannis Ioannidis and Joseph A. Konstan, are working together to solicit and answer questions from the computing community! Please refer to the instructions posted at https://vote.escvote.com/acm. Please note the election email will be addressed from acmhelp@mg.electionservicescorp.com. Please return your ballot in the enclosed envelope, which must be signed by you on the outside in the space provided. The signed ballot envelope may be inserted into a separate envelope for mailing if you prefer this method. All ballots must be received by no later than 16:00 UTC on 23 May 2022. Validation by the Elections Committee will take place at 14:00 UTC on 25 May 2022. Yannis Ioannidis is Professor of Informatics & Telecom at the U. of Athens, Greece (since 1997). Prior to that, he was a professor of Computer Sciences at the U. of Wisconsin-Madison (1986-1997).