3,263 research outputs found
Acute dialysis quality initiative (ADQI)
No full textA new star is born. Its name is ADQI which stands for Acute Dialysis Quality Initiative. In the firmament of guidelines and directions for appropriate management of renal diseases, little has been done so far concerning acute renal failure and its treatment. For this reason we felt that a process seeking consensus and evidence-based statements in the field of acute renal failure was needed
Nomenclature for continuous renal replacement therapies
No full textContinuous renal replacement therapies (CRRTs) have evolved over the last decade, but there is no standard terminology for the defferent methods in use, At an International Conference on CRRT, held in San Diego, CA, November 9-10, 1995, an international panel of experts developed a proposed nomenclature for these therapies, The nomenclature was developed to define common terms and to use a standardized language when papers in the field of CRRT are reviewed and published, This article describes the definition for each technique, It is hoped that these definitions will be used as a framework for subsequent descriptions of new techniques in the literature. (C) 1996 by the National Kidney Foundation, Inc
Proceedings of the International Conference on Continuous Renal Replacement Therapy - November 9-10, 1995 - Introduction
No full textTeacher-apprentices RL (TARL): leveraging complex policy distribution through generative adversarial hypernetwork in reinforcement learning
No full textTypically, a Reinforcement Learning (RL) algorithm focuses in learning a single deployable policy as the end product. Depending on the initialization methods and seed randomization, learning a single policy could possibly leads to convergence to different local optima across different runs, especially when the algorithm is sensitive to hyper-parameter tuning. Motivated by the capability of Generative Adversarial Networks (GANs) in learning complex data manifold, the adversarial training procedure could be utilized to learn a population of good-performing policies instead. We extend the teacher-student methodology observed in the Knowledge Distillation field in typical deep neural network prediction tasks to RL paradigm. Instead of learning a single compressed student network, an adversarially-trained generative model (hypernetwork) is learned to output network weights of a population of good-performing policy networks, representing a school of apprentices. Our proposed framework, named Teacher-Apprentices RL (TARL), is modular and could be used in conjunction with many existing RL algorithms. We illustrate the performance gain and improved robustness by combining TARL with various types of RL algorithms, including direct policy search Cross-Entropy Method, Q-learning, Actor-Critic, and policy gradient-based methods.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Interactive Intelligenc
BADDr: Bayes-Adaptive Deep Dropout RL for POMDPs
No full textWhile reinforcement learning (RL) has made great advances in scalability, exploration and partial observability are still active research topics. In contrast, Bayesian RL (BRL) provides a principled answer to both state estimation and the exploration-exploitation trade-off, but struggles to scale. To tackle this challenge, BRL frameworks with various prior assumptions have been proposed, with varied success. This work presents a representation-agnostic formulation of BRL under partially observability, unifying the previous models under one theoretical umbrella. To demonstrate its practical significance we also propose a novel derivation, Bayes-Adaptive Deep Dropout rl (BADDr), based on dropout networks. Under this parameterization, in contrast to previous work, the belief over the state and dynamics is a more scalable inference problem. We choose actions through Monte-Carlo tree search and empirically show that our method is competitive with state-of-the-art BRL methods on small domains while being able to solve much larger ones.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Interactive Intelligenc
The first international consensus conference on continuous renal replacement therapy
No full textThe first international consensus conference on continuous renal replacement therapy.BackgroundManagement of acute renal failure (ARF) in the critically ill is extremely variable and there are no published standards for the provision of renal replacement therapy in this population. We sought to review the available evidence, make evidence-based practice recommendations, and delineate key questions for future study.MethodsWe undertook an evidence-based review of the literature on continuous renal replacement therapy (CRRT) using MEDLINE searches. We determined a list of key questions and convened a 2-day consensus conference to develop summary statements via a series of alternating breakout and plenary sessions. In these sessions, we identified supporting evidence and generated practice guidelines and/or directions for future research.ResultsOf the 46 questions considered, we found consensus for 20. We found inadequate evidence for 21 questions and for the remaining five we found data but no consensus. Full versions of workgroup findings are available on the Internet at http://www.ADQI.net.ConclusionsDespite limited data, broad areas of consensus exist for use of CRRT and guideline development appears feasible. Equally broad areas of disagreement also exist and additional basic and applied research in acute renal failure is needed
Acute renal failure - definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group
No full textIntroduction There is no consensus definition of acute renal failure (ARF) in critically ill patients. More than 30 different definitions have been used in the literature, creating much confusion and making comparisons difficult. Similarly, strong debate exists on the validity and clinical relevance of animal models of ARF; on choices of fluid management and of end-points for trials of new interventions in this field; and on how information technology can be used to assist this process. Accordingly, we sought to review the available evidence, make recommendations and delineate key questions for future studies. Methods We undertook a systematic review of the literature using Medline and PubMed searches. We determined a list of key questions and convened a 2-day consensus conference to develop summary statements via a series of alternating breakout and plenary sessions. In these sessions, we identified supporting evidence and generated recommendations and/or directions for future research. Results We found sufficient consensus on 47 questions to allow the development of recommendations. Importantly, we were able to develop a consensus definition for ARF. In some cases it was also possible to issue useful consensus recommendations for future investigations. We present a summary of the findings. (Full versions of the six workgroups' findings are available on the internet at http://www.ADQI.net) Conclusion Despite limited data, broad areas of consensus exist for the physiological and clinical principles needed to guide the development of consensus recommendations for defining ARF, selection of animal models, methods of monitoring fluid therapy, choice of physiological and clinical end-points for trials, and the possible role of information technology
Refined Risk Management in Safe Reinforcement Learning with a Distributional Safety Critic
No full textSafety is critical to broadening the real-world use of reinforcement learning (RL). Modeling the safety aspects using a safety-cost signal separate from the reward is becoming standard practice, since it avoids the problem of finding a good balance between safety and performance. However, the total safety-cost distribution of different trajectories is still largely unexplored. In this paper, we propose an actor critic method for safe RL that uses an implicit quantile network to approximate the distribution of accumulated safety-costs. Using an accurate estimate of the distribution of accumulated safetycosts, in particular of the upper tail of the distribution, greatly improves the performance of riskaverse RL agents. The empirical analysis shows that our method achieves good risk control in complex safety-constrained environments.AlgorithmicsIntelligent Electrical Power Grid
- …
