Background: In the United States, 5.7 million patients are admitted annually to intensive care units (ICU), with costs exceeding $82 billion. Although close monitoring and dynamic assessment of patient acuity are key aspects of ICU care, both are limited by the time constraints imposed on healthcare providers. Methods: Using the University of Florida Health (UFH) Integrated Data Repository as Honest Broker, we created a database with electronic health records data from a retrospective study cohort of 38,749 adult patients admitted to ICU at UF Health between 06/01/2014 and 08/22/2019. This repository includes demographic information, comorbidities, vital signs, laboratory values, medications with date and timestamps, and diagnoses and procedure codes for all index admission encounters as well as encounters within 12 months prior to index admission and 12 months follow-up. We developed algorithms to identify acuity status of the patient every four hours during each ICU stay. Results: We had 383,193 encounters (121,800 unique patients) admitted to the hospital, and 51,073 encounters (38,749 unique patients) with at least one ICU stay that lasted more than four hours. These patients requiring ICU admission had longer median hospital stay (7 days vs. 1 day) and higher in-hospital mortality (9.6% vs. 0.4%) compared with those not admitted to the ICU. Among patients who were admitted to the ICU and expired during hospital admission, more deaths occurred in the ICU than on general hospital wards (7.4% vs. 0.8%, respectively). Conclusions: We developed phenotyping algorithms that determined patient acuity status every four hours while admitted to the ICU. This approach may be useful in developing prognostic and clinical decision-support tools to aid patients, caregivers, and providers in shared decision-making processes regarding resource use and escalation of care.

Summary: Using 3D imaging and artificial intelligence, researchers discovered the shortest distance between two points on the curved surface of the face predicted, with 89% accuracy, which patients had sleep apnea. Facial features analyzed from 3D photographs could predict the likelihood of having obstructive sleep apnea, according to a study published in the April issue of the Journal of Clinical Sleep Medicine. Using 3D photography, the study found that geodesic measurements -- the shortest distance between two points on a curved surface -- predicted with 89 percent accuracy which patients had sleep apnea. Using traditional 2D linear measurements alone, the algorithm's accuracy was 86 percent. "This application of the technique used predetermined landmarks on the face and neck," said principle investigator Peter Eastwood, who holds a doctorate in respiratory and sleep physiology and is the director of the Centre for Sleep Science at the University of Western Australia (UWA).

The algorithmic revolution is here, and nations are losing control of not only their understanding of the potential impact of artificial intelligence but also the governance model that enforced accountability on the advances in science and technology over the years at all levels. While each new technology innovation claims its territory for the economic advances in the human ecosystem with significant ramifications across cyberspace, geospace and/or space (CGS), the rise of artificial intelligence (AI) has not only undermined governance, management, and growth models, but it has also broken all barriers to boundaries defined by human decision makers. In addition, it is both blurring the boundaries between human intelligence and machine intelligence, and the boundaries between man and machine and real and fake. As a result, the power dynamics are shifting away from the select few across nations (and is moving away from humans entirely to algorithms)--re-defining the criteria upon which geopolitics was framed--and thereby threatening the foundations of global peace and security. Since the beginning of the technological age, each new idea, innovation, and invention has helped humans across nations usher in a new era of economic growth, changing the fundamentals of respective nations and their security.

With a fixed model structure, knowledge distillation and filter grafting are two effective ways to boost single model accuracy. However, the working mechanism and the differences between distillation and grafting have not been fully unveiled. In this paper, we evaluate the effect of distillation and grafting in the filter level, and find that the impacts of the two techniques are surprisingly complementary: distillation mostly enhances the knowledge of valid filters while grafting mostly reactivates invalid filters. This observation guides us to design a unified training framework called DGD, where distillation and grafting are naturally combined to increase the knowledge density inside the filters given a fixed model structure. Through extensive experiments, we show that the knowledge densified network in DGD shares both advantages of distillation and grafting, lifting the model accuracy to a higher level.

Experts from the private sector and leading research labs in the US and Europe have joined forces to create a toolkit for turning AI ethics principles into practice. The preprint paper, published last week, advocates paying people for finding risks of bias in artificial intelligence (AI) systems – adapting a model used to check the security of new computer systems, in which hackers are paid'bounties' for identifying weaknesses. The paper also proposes better linking independent third-party auditing operations and government policies to foster a market in regulatory systems, and suggests that governments increase funding for researchers in academia to verify performance claims made by industry. The 80-page paper, Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claims, has been put together by AI specialists from 30 organisations including Google Brain, Intel, OpenAI, Stanford University and the Leverhulme Centre for the Future of Intelligence. "In order for AI developers to earn trust from system users, customers, civil society, governments, and other stakeholders that they are building AI responsibly, there is a need to move beyond [ethics] principles to a focus on mechanisms for demonstrating responsible behaviour," the executive summary reads.

Automatic programming (AP) is an important area of Machine Learning (ML) where computer programs are generated automatically. Swarm Programming (SP), a newly emerging research area in AP, automatically generates the computer programs using Swarm Intelligence (SI) algorithms. This paper presents two grammar-based SP methods named as Grammatical Moth-Flame Optimizer (GMFO) and Grammatical Whale Optimizer (GWO). The Moth-Flame Optimizer and Whale Optimization algorithm are used as search engines or learning algorithms in GMFO and GWO respectively. The proposed methods are tested on Santa Fe Ant Trail, quartic symbolic regression, and 3-input multiplexer problems. The results are compared with Grammatical Bee Colony (GBC) and Grammatical Fireworks algorithm (GFWA). The experimental results demonstrate that the proposed SP methods can be used in automatic computer program generation.

The 2020 Eurovision Song Contest may have been cancelled, but fans of formulaic pop can still get their fill courtesy of the VPRO AI Song Contest. The contestants and their entries were revealed on 10 April and the public have until 10 May to cast their votes. Thirteen teams have entered, with the competition open to anyone residing in a country eligible to take part in the traditional Eurovision extravaganza. The contestants have used a variety of machine learning techniques to help create their songs, with the teams relying on computer input to different degrees. All artists were keen to stress that, rather than pressing a button and letting their trained algorithms create the entire piece, their work is a result of collaboration between AI and humans.

Scientific analyses often rely on slow, but accurate forward models for observable data conditioned on known model parameters. While various emulation schemes exist to approximate these slow calculations, these approaches are only safe if the approximations are well understood and controlled. This workshop submission reviews and updates a previously published method, which has been used in cosmological simulations, to (1) train an emulator while simultaneously estimating posterior probabilities with MCMC and (2) explicitly propagate the emulation error into errors on the posterior probabilities for model parameters. We demonstrate how these techniques can be applied to quickly estimate posterior distributions for parameters of the $\Lambda$CDM cosmology model, while also gauging the robustness of the emulator approximation.

This paper is concerned by the problem of selecting an optimal sampling set of sensors over a network of time series for the purpose of signal recovery at non-observed sensors with a minimal reconstruction error. The problem is motivated by applications where time-dependent graph signals are collected over redundant networks. In this setting, one may wish to only use a subset of sensors to predict data streams over the whole collection of nodes in the underlying graph. A typical application is the possibility to reduce the power consumption in a network of sensors that may have limited battery supplies. We propose and compare various data-driven strategies to turn off a fixed number of sensors or equivalently to select a sampling set of nodes. We also relate our approach to the existing literature on sensor selection from multivariate data with a (possibly) underlying graph structure. Our methodology combines tools from multivariate time series analysis, graph signal processing, statistical learning in high-dimension and deep learning. To illustrate the performances of our approach, we report numerical experiments on the analysis of real data from bike sharing networks in different cities.

We investigate the evolution of the Q values for the implementation of Deep Q Learning (DQL) in the Stable Baselines library. Stable Baselines incorporates the latest Reinforcement Learning techniques and achieves superhuman performance in many game environments. However, for some simple non-game environments, the DQL in Stable Baselines can struggle to find the correct actions. In this paper we aim to understand the types of environment where this suboptimal behavior can happen, and also investigate the corresponding evolution of the Q values for individual states. We compare a smart TrafficLight environment (where performance is poor) with the AI Gym FrozenLake environment (where performance is perfect). We observe that DQL struggles with TrafficLight because actions are reversible and hence the Q values in a given state are closer than in FrozenLake. We then investigate the evolution of the Q values using a recent decomposition technique of Achiam et al.. We observe that for TrafficLight, the function approximation error and the complex relationships between the states lead to a situation where some Q values meander far from optimal.