Find open roles in Artificial Intelligence (AI), Machine Learning (ML), Natural Language Processing (NLP), Computer Vision (CV), Data Engineering, Data Analytics, Big Data, and Data Science in general, filtered by job title or popular skill, toolset and products used.

Booz Allen Hamilton announced the creation of HELworks, an innovative developer of directed energy and high energy laser (HEL) weapon systems designed to meet the needs of warfighters in the modern battlespace. "Booz Allen's significant investment in HEL technology maturation and operational prototypes ensures that HELworks solutions are operationally relevant and producible at scale--ready for use by warfighters, today. We are proud to be on the forefront of innovation and mission-focused leaders in this area." Booz Allen has made significant investment in independent research and development projects focused on developing directed energy solutions over the past 5 years, based in the firm's understanding of Department of Defense (DOD) needs and mission requirements. HELworks leverages Booz Allen's 25-plus year heritage of directed energy expertise to optimize size, weight, and power (SWaP); deliver enhanced military utility; and provide rapid deployment of first-of-its-kind HEL solutions.

SharkNinja Operating LLC, a subsidiary of JS Global, an innovation leader in the consumer floorcare industry, announced that the International Trade Commission issued its Initial Determination in an ongoing patent case brought by iRobot Corp. (iRobot) against SharkNinja, largely in SharkNinja's favor. The Initial Determination, issued by the Administrative Law Judge, found for SharkNinja, and against iRobot, on two of the four remaining patents asserted by iRobot, and one claim of the remaining two patents. None of SharkNinja's top-selling AI Ultra products and none of SharkNinja's auto-empty robot products were found to infringe any valid patent claim. The Initial Determination found for iRobot on certain claims of two patents, which were asserted against a small subset of SharkNinja's product line. The Initial Determination is non-final, and is subject to review by the International Trade Commission, which should be completed by the beginning of February 2023.

In all cases, data and AI are enabling the public sector to achieve its missions with more pace, efficiency, and security. The implementation of large-scale automation fosters engagement, as liberated citizens can interact with public servants and processes around the clock. Moreover, the level of security and service is markedly improved, with automation powering real-time threat, incident, and anomaly detection. When taken together, data and AI generates insight that can be leveraged to feed a better decision-making process – from understanding a situation to suggesting next-best actions. It is important to remember that data and AI are tools, just like a carpenter's electric saw. They make the work easier, add precision, and improve efficiency.

New data from Tech Nation and Dealroom found that VC investment in UK AI companies hit a record in Q2 2022 and is on track to at least equal last year. Around 2,009 AI companies are based in the UK, the AI capital of Europe. Last year saw a record investment of £6.6 billion flow into UK AI companies. In 2022 so far, £3.2 billion has been invested. Tech Nation notes that investment into AI firms is historically "loaded towards the end of the calendar year" and is on track to "at least equal" last year.

Cyber security companies are turning to artificial intelligence and machine learning tools to ward off growing number of attacks on networks, Finland-based internet security firm F-Secure said. As the world is fast moving towards Internet of Things and connected devices, deployment of artificial intelligence (AI) has become inevitable for cyber security firms to analyse huge amount of data to save networks from infiltration attempts, F-Secure's Security Advisor Sean Sullivan said. Networks are persistently exposed to threats like malware, phishing, password breaches and denial of service attacks. On a daily basis, F-Secure Labs on an average receives sample data of 500,000 files from its customers that include 10,000 malware variants and 60,000 malicious URLs for analysis and protection, Sullivan said. For humans, it is a big task to go through such huge amount of data and machine learning tools and AI are lending a helping hand at this stage, he said.

The study of mortality patterns is a popular research topic in many areas. We are particularly interested in mortality patterns among main causes of death associated with age-gender combinations. We use symbolic data analysis (SDA) and include three dimensions: age, gender, and patterns across main causes of death. In this study, we present an alternative method to identify clusters of EU countries with similar mortality patterns in the young population, while considering comprehensive information on the distribution of deaths among the main causes of death by different age-gender groups. We explore possible relationships between mortality patterns in the identified clusters and some other sociodemographic indicators. We use EU data of crude mortality rates from 2016, as the most recent complete data available.

Avoiding overfitting is a central challenge in machine learning, yet many large neural networks readily achieve zero training loss. This puzzling contradiction necessitates new approaches to the study of overfitting. Here we quantify overfitting via residual information, defined as the bits in fitted models that encode noise in training data. Information efficient learning algorithms minimize residual information while maximizing the relevant bits, which are predictive of the unknown generative models. We solve this optimization to obtain the information content of optimal algorithms for a linear regression problem and compare it to that of randomized ridge regression. Our results demonstrate the fundamental trade-off between residual and relevant information and characterize the relative information efficiency of randomized regression with respect to optimal algorithms. Finally, using results from random matrix theory, we reveal the information complexity of learning a linear map in high dimensions and unveil information-theoretic analogs of double and multiple descent phenomena.

This document provides extensive guidelines and examples for Rhetorical Structure Theory (RST) annotation in Mandarin Chinese. The guideline is divided into three sections. We first introduce preprocessing steps to prepare data for RST annotation. Secondly, we discuss syntactic criteria to segment texts into Elementary Discourse Units (EDUs). Lastly, we provide examples to define and distinguish discourse relations in different genres. We hope that this reference manual can facilitate RST annotations in Chinese and accelerate the development of the RST framework across languages.

IceCube, a cubic-kilometer array of optical sensors built to detect atmospheric and astrophysical neutrinos between 1 GeV and 1 PeV, is deployed 1.45 km to 2.45 km below the surface of the ice sheet at the South Pole. The classification and reconstruction of events from the in-ice detectors play a central role in the analysis of data from IceCube. Reconstructing and classifying events is a challenge due to the irregular detector geometry, inhomogeneous scattering and absorption of light in the ice and, below 100 GeV, the relatively low number of signal photons produced per event. To address this challenge, it is possible to represent IceCube events as point cloud graphs and use a Graph Neural Network (GNN) as the classification and reconstruction method. The GNN is capable of distinguishing neutrino events from cosmic-ray backgrounds, classifying different neutrino event types, and reconstructing the deposited energy, direction and interaction vertex. Based on simulation, we provide a comparison in the 1-100 GeV energy range to the current state-of-the-art maximum likelihood techniques used in current IceCube analyses, including the effects of known systematic uncertainties. For neutrino event classification, the GNN increases the signal efficiency by 18% at a fixed false positive rate (FPR), compared to current IceCube methods. Alternatively, the GNN offers a reduction of the FPR by over a factor 8 (to below half a percent) at a fixed signal efficiency. For the reconstruction of energy, direction, and interaction vertex, the resolution improves by an average of 13%-20% compared to current maximum likelihood techniques in the energy range of 1-30 GeV. The GNN, when run on a GPU, is capable of processing IceCube events at a rate nearly double of the median IceCube trigger rate of 2.7 kHz, which opens the possibility of using low energy neutrinos in online searches for transient events.