Patient representation learning refers to learning a dense mathematical representation of a patient that encodes meaningful information from Electronic Health Records (EHRs). This is generally performed using advanced deep learning methods. This study presents a systematic review of this field and provides both qualitative and quantitative analyses from a methodological perspective. We identified studies developing patient representations from EHRs with deep learning methods from MEDLINE, EMBASE, Scopus, the Association for Computing Machinery (ACM) Digital Library, and Institute of Electrical and Electronics Engineers (IEEE) Xplore Digital Library. After screening 362 articles, 48 papers were included for a comprehensive data collection. We noticed a typical workflow starting with feeding raw data, applying deep learning models, and ending with clinical outcome predictions as evaluations of the learned representations. Specifically, learning representations from structured EHR data was dominant (36 out of 48 studies). Recurrent Neural Networks were widely applied as the deep learning architecture (LSTM: 13 studies, GRU: 11 studies). Disease prediction was the most common application and evaluation (30 studies). Benchmark datasets were mostly unavailable (28 studies) due to privacy concerns of EHR data, and code availability was assured in 20 studies. We show the importance and feasibility of learning comprehensive representations of patient EHR data through a systematic review. Advances in patient representation learning techniques will be essential for powering patient-level EHR analyses. Future work will still be devoted to leveraging the richness and potential of available EHR data. Knowledge distillation and advanced learning techniques will be exploited to assist the capability of learning patient representation further.

Reinforcement learning in heterogeneous multi-agent scenarios is important for real-world applications but presents challenges beyond those seen in homogeneous settings and simple benchmarks. In this work, we present an actor-critic algorithm that allows a team of heterogeneous agents to learn decentralized control policies for covering an unknown environment. This task is of interest to national security and emergency response organizations that would like to enhance situational awareness in hazardous areas by deploying teams of unmanned aerial vehicles. To solve this multi-agent coverage path planning problem in unknown environments, we augment a multi-agent actor-critic architecture with a new state encoding structure and triplet learning loss to support heterogeneous agent learning. We developed a simulation environment that includes real-world environmental factors such as turbulence, delayed communication, and agent loss, to train teams of agents as well as probe their robustness and flexibility to such disturbances.

Software often produces biased outputs. In particular, machine learning (ML) based software are known to produce erroneous predictions when processing discriminatory inputs. Such unfair program behavior can be caused by societal bias. In the last few years, Amazon, Microsoft and Google have provided software services that produce unfair outputs, mostly due to societal bias (e.g. gender or race). In such events, developers are saddled with the task of conducting fairness testing. Fairness testing is challenging; developers are tasked with generating discriminatory inputs that reveal and explain biases. We propose a grammar-based fairness testing approach (called ASTRAEA) which leverages context-free grammars to generate discriminatory inputs that reveal fairness violations in software systems. Using probabilistic grammars, ASTRAEA also provides fault diagnosis by isolating the cause of observed software bias. ASTRAEA's diagnoses facilitate the improvement of ML fairness. ASTRAEA was evaluated on 18 software systems that provide three major natural language processing (NLP) services. In our evaluation, ASTRAEA generated fairness violations with a rate of ~18%. ASTRAEA generated over 573K discriminatory test cases and found over 102K fairness violations. Furthermore, ASTRAEA improves software fairness by ~76%, via model-retraining.

The National Security Commission on Artificial Intelligence called for the creation of two new talent initiatives that would help "fundamentally re-imagine" the way the federal government builds its digital workforce in Congressional testimony Tuesday. Members of the NSCAI, led by Eric Schmidt, former chief executive officer of Google, provided an interim review of the commission's work in testimony to a House Armed Services subcommittee Tuesday. In addition to urging lawmakers to legislate the U.S. Digital Service Academy and National Reserve Digital Corps workforce initiatives, NSCAI outlined provisions in five areas included in the 2021 National Defense Authorization Act the commission views as "crucial." But it was Rep. Elissa Slotkin, D-Mich., who brought the hearing down to earth. The former intelligence analyst and Defense Department official asked NSCAI commissioners a pointed question: How can DOD turn the commission's ambitious dreams into realities?

The impacts of the recent COVID-19 pandemic on global trade and our daily lives have led to an acceleration of technological innovations in different sectors. At the same time, our society is transitioning from the era of Big Data towards Hyper Digitization. To help industries prepare for these trends and navigate the post-pandemic world, this year's Taiwan Innotech Expo (TIE 2020) was highlighting the latest smart living technologies that can spark new imaginations. Since its transformation into a global trade show, the TIE continues to draw international attention to Taiwan's strength in R&D and innovation. This year's event showcases how Taiwan stays resilient in face of a global crisis.

Drone startup Skydio today announced the U.S. Federal Aviation Administration (FAA) has granted the North Carolina Department of Transportation (NCDOT) statewide approval to fly Skydio drones beyond visual line of sight to inspect bridges. Skydio, which describes the waiver as the first of its kind, says the NCDOT will be able to conduct maintenance activities without the use of visual observers like trained pilots or staff. A recent study by the American Association of State Highway and Transportation Officials found that taxpayer cost per bridge inspection can be reduced 75% by switching from traditional methods to drones. The Minnesota Department of Transportation found that using drones for bridge inspection offsets some or all of the costs, depending on the bridge configuration and location, with a trial of drone-assisted inspections saving an average of 40% over traditional methods and providing ostensibly superior data and reporting. Going forward, the NCDOT's inspectors can send Skydio 2 drones to inspect critical structures below bridges in North Carolina instead of conducting rappels or using "snooper trucks."

The Fates, it sometimes seems, prefer extreme outcomes. While humans usually reject predictions of futures dramatically changed from the present, information technology has produced a never-ending stream of upheavals in the economy, warfare, our very way of life. Thus, cyberspace in 2030 could be a very different place than it is today, for good or ill. How we deploy artificial intelligence and machine learning to attack and to defend networks will make the difference. Today cyberspace is a hostile environment.

The USA has been named as the country best prepared to realise the benefits of artificial intelligence (AI) technologies in public service delivery, topping the 2020 Government AI Readiness Index. Meanwhile Singapore, which led the 2019 list, has fallen to sixth place. The index – compiled by UK-based consultants Oxford Insights and Canada's International Development Research Centre (IDRC) – examines how well-placed nations are to take advantage of the benefits of AI in their internal operations and the delivery of public services. This year, 172 countries were reviewed. The ranking measures AI readiness across three criteria: government willingness to adopt AI, and the ability to adapt and innovate to do so; availability of AI expertise and tools from the technology sector; and capabilities in building AI tools, providing them with high-quality data, and building them into public services.

The Department of Defense wants industry to give its Joint Artificial Intelligence Center ideas for a project aimed at creating a model to automate and accelerate AI technology acquisition processes. JAIC is looking to use other transaction authority for prototyping an automated business process system, according to a sources sought notice published Friday. Interested parties have until Oct. 9 to provide input. Goals of the project include driving JAIC's collaboration with non-traditional contractors within the commercial and academic sectors and increasing its adoption of Agile and DevSecOps approaches to manage programs of record, the Aug. 28 RFI notice says. The center initially posted a request for information to determine a potential nonprofit manager for its planned AI acquisition business model.

Astronauts on NASA's Apollo missions to the moon captured astounding movies of the lunar surface, but recent enhancements with artificial intelligence (AI) have really made the films out of this world. In remastered movies shared online by by DutchSteamMachine, a YouTube channel run by a film restoration specialist in the Netherlands, details from lunar scenes are astonishingly crisp and vivid; from mission commander Neil Armstrong's first steps on the moon in 1969 to bumpy lunar rover drives during Apollo 15 and 16 in 1971 and 1972, respectively. The film restorer behind DutchSteamMachine, who also goes by "Niels," used AI to stabilize shaky footage and generate new frames in NASA moon landing films; increasing the frame rate (the number of frames that play per second) smoothed the motion and made it look more like movement in high-definition (HD) video. Related: Can machines be creative? The Apollo program launched 11 lunar spaceflight missions between 1968 and 1972; of those, four missions tested equipment and six landed on the moon, allowing 12 men to walk, drive and/or leap over the dusty, cratered lunar surface, according to NASA.