Shrager J.,Stanford University |
PLoS ONE | Year: 2010
The Internet has enabled profound changes in the way science is performed, especially in scientific communications. Among the most important of these changes is the possibility of new models for pre-publication review, ranging from the current, relatively strict peer-review model, to entirely unreviewed, instant self-publication. Different models may affect scientific progress by altering both the quality and quantity of papers available to the research community. To test how models affect the community, I used a multi-agent simulation of treatment selection and outcome in a patient population to examine how various levels of pre-publication review might affect the rate of scientific progress. I identified a "sweet spot" between the points of very limited and very strict requirements for pre-publication review. The model also produced a u-shaped curve where very limited review requirement was slightly superior to a moderate level of requirement, but not as large as the aforementioned sweet spot. This unexpected phenomenon appears to result from the community taking longer to discover the correct treatment with more strict pre-publication review. In the parameter regimens I explored, both completely unreviewed and very strictly reviewed scientific communication seems likely to hinder scientific progress. Much more investigation is warranted. Multi-agent simulations can help to shed light on complex questions of scientific communication and exhibit interesting, unexpected behaviors. © 2010 Jeff Shrager.
CollabRx | Date: 2014-01-10
Applications for providing web-based and mobile expert systems for clinical decision support; software applications for providing expert systems for clinical decision support; mobile and web-based applications with reference tools, social media and expert systems; software applications with reference tools, social media and expert systems. Information services for providing web-based and mobile expert systems for clinical and medical decision support; information services for providing clinical and medical reference tools, social media and expert systems; subscription services; data analytics services based on user data and use patterns.
CollabRx | Date: 2015-06-03
An evidence-based computerized method for forming treatment plans can utilize an actionability framework, which can include a basis of actionability and a rationale for actionability for biomarkers. Treatment rules derived from data in literature, such as from clinical trials, case studies, research and published literature, can allow the method to form treatment plans with support reasoning and citations, similar to those of medical professionals. Thus the treatment plans proposed by the evidence-based treatment process can survive the inspection and scrutiny of treating physicians due to the available and cited support documents.
CollabRx | Date: 2011-08-26
Methods to treat cancer patients, especially melanoma patients, who have BRAF mutations and have become resistant to BRAF mutant kinase inhibitors employ inhibitors of multiple receptor tyrosine kinases. In addition, methods are described for identifying pharmaceutical compositions and drugs that will be successful in treating these patients.
Shrager J.,CollabRx |
Shrager J.,Stanford University
AI Magazine | Year: 2011
Cancer kills millions of people each year. From an AI perspective, finding effective treatments for cancer is a high-dimensional search problem characterized by many molecularly distinct cancer subtypes, many potential targets and drug combinations, and a dearth of high-quality data to connect molecular subtypes and treatments to responses. The broadening availability of molecular diagnostics and electronic medical records presents both opportunities and challenges to apply AI techniques to personalize and improve cancer treatment. We discuss these in the context of Cancer Commons, a "rapid learning" community where patients, physicians, and researchers collect and analyze the molecular and clinical data from every cancer patient and use these results to individualize therapies. Research opportunities include adaptively planning and executing individual treatment experiments across the whole patient population, inferring the causal mechanisms of tumors, predicting drug response in individuals, and generalizing these findings to new cases. The goal is to treat each patient in accord with the best available knowledge and to continually update that knowledge to benefit subsequent patients. Achieving this goal is a worthy grand challenge for AI. Copyright © 2011, Association for the Advancement of Artificial Intelligence. All rights reserved.