Despite its great success, machine learning can have its limits when dealing with insufficient training data. A potential solution is the additional integration of prior knowledge into the training process which leads to the notion of informed machine learning.In this paper, we present a structured overview of various approaches in this field. We provide a definition and propose a concept for informed machine learning which illustrates its building blocks and distinguishes it from conventional machine learning. We introduce a taxonomy that serves as a classification framework for informed machine learning approaches. It considers the source of knowledge, its representation, and its integration into the machine learning pipeline. Based on this taxonomy, we survey related research and describe how different knowledge representations such as algebraic equations, logic rules, or simulation results can be used in learning systems. This evaluation of numerous papers on the basis of our taxonomy uncovers key methods in the field of informed machine learning.
Bronchiolitis obliterans syndrome (BOS) after allogeneic hematopoietic cell transplantation (HCT) is associated with high mortality. Purpose: We hypothesized that FAM (inhaled Fluticasone, Azithromycin, and Montelukast) with a brief steroid pulse could avert progression of new-onset BOS. Experimental design: We tested this in a phase II, single-arm, open label, multicenter study (NCT01307462). Results: Thirty-six patients were enrolled within 6 months of BOS diagnosis. The primary endpoint was treatment failure, defined as 10% or greater FEV1% decline at 3 months. At 3 months, 6% (2/36, 95% CI 1%–19%) had treatment failure (vs. 40% in historical controls, p<0.001). FAM was well tolerated. Steroid dose was reduced by 50% or more at 3 months in 48% of patients who could be evaluated (n=27). Patient-reported outcomes at 3 months were statistically significantly improved for SF-36 social functioning score and mental component score, FACT emotional well-being, and Lee symptom scores in lung, skin, mouth, and the overall summary score compared to enrollment (n=24). At 6 months, 36% had treatment failure (95% CI 21%–54%, n=13/36, with 6 documented failures, 7 missing pulmonary function tests). Overall survival was 97% (95% CI 84%–100%) at 6 months. These data suggest that FAM was well tolerated and that treatment with FAM and steroid pulse may halt pulmonary decline in new-onset BOS in the majority of patients and permit reductions in systemic steroid exposure, which collectively may improve quality of life. However, additional treatments are needed for progressive BOS despite FAM.
Several distinct graft-versus-host disease (GVHD)-related syndromes have been defined by the NIH Consensus Conference. We enrolled a prospective cohort of 911 hematopoietic cell transplantation (HCT) recipients at 13 centers between March 2011 and May 2014 to evaluate four GVHD syndromes: late acute GVHD, chronic GVHD, bronchiolitis obliterans syndrome, and cutaneous sclerosis. The median age at HCT was 53.7 years. Most patients received peripheral blood stem cell transplant (81%) using a non myeloablative or reduced intensity conditioning (55%). Pediatric age group and use of bone marrow and umbilical cord blood were underrepresented in our cohort (<=11%). The cumulative incidence of late acute GVHD (late onset and recurrent) was 10% at a median of 5.5 months, chronic GVHD was 47% at a median of 7.4 months, bronchiolitis obliterans was 3% at a median of 12.2 months, and cutaneous sclerosis was 8% at a median onset of 14.0 months after HCT. Late acute GVHD and bronchiolitis obliterans had particularly high non-relapse mortality of 23% and 32% by 2 years after diagnosis. The probability of late acute- and chronic-GVHD-free, relapse-free survival at one and two years after HCT was 38% and 26%. This multi-center, prospective study confirms the high rate of late acute and chronic GVHD syndromes and supports the need for continuous close monitoring and development of more effective GVHD treatment strategies to improve HCT success.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.