The COVID-19 pandemic has undeniably showcased new ways of delivering care to patients within the NHS, not least through the increased use of digital, technological and AI-based healthcare solutions. The immediacy and necessity of the response paved the way for health tech adoption at a pace unencumbered by prior levels of red tape and underpinned by a desire to work in'new' ways for what was a'new' challenge at the time. Technological innovation has supported health services in numerous ways, including with making significant decisions around capacity and priorities, through the increased remote monitoring of patients, and by allowing more effective communication among the workforce. Having said this, the long-standing barriers to innovation faced by health tech companies have not been, and are unlikely to be, substantially dismantled. While the market access landscape has evolved for companies over recent years, with a number of positive initiatives within the national architecture acting as a'pull' for innovation, the adoption challenge largely remains entrenched.

Artificial Intelligence (AI) is bringing a technological revolution to society. The new emerging digital world carries with it a scary thing: Artificial Intelligence (AI) bias. It is a pressing concern over as AI is becoming extremely powerful and at the same time with a lot of discriminatory thoughts like humans. Human bias is not new. The recent protests across the globe on racial discrimination are a pure example that bias is a major threat to human society.

Moreover, "Cold War," "Modern Warfare" and "Warzone" will continue to feature live service elements, per a blog posted Thursday, with a new battlepass featuring "Cold War" coming in December. More content likely means more required storage space, and while Activision's plan allows players to progress in the battlepass and earn Call of Duty XP regardless of which of the three versions they're playing, the sheer volume of required disk space may push players to choose their favorite version of the franchise, limit space for non-COD games or modify their usual gaming behaviors in other ways.

These are amongst the first detailed legislative proposals to be published internationally, so make for interesting reading for stakeholders worldwide. For AI product producers, these ideas merit careful consideration. Next year, the European Commission said it would issue draft regulations on AI. The Commission could well adopt any of the European Parliament's proposals, or variants on them. Affected stakeholders will have opportunities to engage with any new AI laws during the normal legislative process, but efforts to understand how these proposals could affect your company should start now.

Autonomous AI systems are rapidly making their way into the health care system, presenting regulators with thorny questions about how to protect data, prevent bias, and make sure constantly evolving machines can operate safely in clinical practice. The urgency of those inquiries will be on display Thursday during a key meeting hosted by the Food and Drug Administration, which is convening patients to collect their perspectives on AI development and regulation. The gathering of the Patient Engagement Advisory Committee comes as the agency considers crossing a crucial threshold: the approval of the first adaptive AI product, in which a system's performance changes based on its use in the real world. To date, the FDA has only approved locked systems that produce the same result based on the same input. Unlock this article by subscribing to STAT and enjoy your first 30 days free!

ML models often exhibit unexpectedly poor behavior when they are deployed in real-world domains. We identify underspecification as a key reason for these failures. An ML pipeline is underspecified when it can return many predictors with equivalently strong held-out performance in the training domain. Underspecification is common in modern ML pipelines, such as those based on deep learning. Predictors returned by underspecified pipelines are often treated as equivalent based on their training domain performance, but we show here that such predictors can behave very differently in deployment domains. This ambiguity can lead to instability and poor model behavior in practice, and is a distinct failure mode from previously identified issues arising from structural mismatch between training and deployment domains. We show that this problem appears in a wide variety of practical ML pipelines, using examples from computer vision, medical imaging, natural language processing, clinical risk prediction based on electronic health records, and medical genomics. Our results show the need to explicitly account for underspecification in modeling pipelines that are intended for real-world deployment in any domain.

Continuous-time event data are common in applications such as individual behavior data, financial transactions, and medical health records. Modeling such data can be very challenging, in particular for applications with many different types of events, since it requires a model to predict the event types as well as the time of occurrence. Recurrent neural networks that parameterize time-varying intensity functions are the current state-of-the-art for predictive modeling with such data. These models typically assume that all event sequences come from the same data distribution. However, in many applications event sequences are generated by different sources, or users, and their characteristics can be very different. In this paper, we extend the broad class of neural marked point process models to mixtures of latent embeddings, where each mixture component models the characteristic traits of a given user. Our approach relies on augmenting these models with a latent variable that encodes user characteristics, represented by a mixture model over user behavior that is trained via amortized variational inference. We evaluate our methods on four large real-world datasets and demonstrate systematic improvements from our approach over existing work for a variety of predictive metrics such as log-likelihood, next event ranking, and source-of-sequence identification.

Neural link predictors are immensely useful for identifying missing edges in large scale Knowledge Graphs. However, it is still not clear how to use these models for answering more complex queries that arise in a number of domains, such as queries using logical conjunctions, disjunctions, and existential quantifiers, while accounting for missing edges. In this work, we propose a framework for efficiently answering complex queries on incomplete Knowledge Graphs. We translate each query into an end-to-end differentiable objective, where the truth value of each atom is computed by a pre-trained neural link predictor. We then analyse two solutions to the optimisation problem, including gradient-based and combinatorial search. In our experiments, the proposed approach produces more accurate results than state-of-the-art methods -- black-box neural models trained on millions of generated queries -- without the need of training on a large and diverse set of complex queries. Using orders of magnitude less training data, we obtain relative improvements ranging from 8% up to 40% in Hits@3 across different knowledge graphs containing factual information. Finally, we demonstrate that it is possible to explain the outcome of our model in terms of the intermediate solutions identified for each of the complex query atoms.

You can find more information about formatting under the section "Submission guidelines" https://www.springer.com/journal/146. For inquiries and to submit your abstract and manuscript, please contact: aisocietyncstate@gmail.com