Zhihui Luo

A human-computer collaborative approach to identifying common data elements in clinical trial eligibility criteria

OBJECTIVE To identify Common Data Elements (CDEs) in eligibility criteria of multiple clinical trials studying the same disease using a human-computer collaborative approach. DESIGN A set of free-text eligibility criteria from clinical trials on two representative diseases, breast cancer and cardiovascular diseases, was sampled to identify disease-specific eligibility criteria CDEs. In this proposed approach, a semantic annotator is used to recognize Unified Medical Language Systems (UMLSs) terms within the eligibility criteria text. The Apriori algorithm is applied to mine frequent disease-specific UMLS terms, which are then filtered by a list of preferred UMLS semantic types, grouped by similarity based on the Dice coefficient, and, finally, manually reviewed. MEASUREMENTS Standard precision, recall, and F-score of the CDEs recommended by the proposed approach were measured with respect to manually identified CDEs. RESULTS Average precision and recall of the recommended CDEs for the two diseases were 0.823 and 0.797, respectively, leading to an average F-score of 0.810. In addition, the machine-powered CDEs covered 80% of the cardiovascular CDEs published by The American Heart Association and assigned by human experts. CONCLUSION It is feasible and effort saving to use a human-computer collaborative approach to augment domain experts for identifying disease-specific CDEs from free-text clinical trial eligibility criteria.

A residue-free green synergistic antifungal nanotechnology for pesticide thiram by ZnO nanoparticles

Here we reported a residue-free green nanotechnology which synergistically enhance the pesticides efficiency and successively eliminate its residue. We built up a composite antifungal system by a simple pre-treating and assembling procedure for investigating synergy. Investigations showed 0.25 g/L ZnO nanoparticles (NPs) with 0.01 g/L thiram could inhibit the fungal growth in a synergistic mode. More importantly, the 0.25 g/L ZnO NPs completely degraded 0.01 g/L thiram under simulated sunlight irradiation within 6 hours. It was demonstrated that the formation of ZnO-thiram antifungal system, electrostatic adsorption of ZnO NPs to fungi cells and the cellular internalization of ZnO-thiram composites played important roles in synergy. Oxidative stress test indicated ZnO-induced oxidative damage was enhanced by thiram that finally result in synergistic antifungal effect. By reducing the pesticides usage, this nanotechnology could control the plant disease economically, more significantly, the following photocatalytic degradation of pesticide greatly benefit the human social by avoiding negative influence of pesticide residue on public health and environment.

Multi-topic assignment for exploratory navigation of consumer health information in NetWellness using formal concept analysis.

BackgroundFinding quality consumer health information online can effectively bring important public health benefits to the general population. It can empower people with timely and current knowledge for managing their health and promoting wellbeing. Despite a popular belief that search engines such as Google can solve all information access problems, recent studies show that using search engines and simple search terms is not sufficient. Our objective is to provide an approach to organizing consumer health information for navigational exploration, complementing keyword-based direct search. Multi-topic assignment to health information, such as online questions, is a fundamental step for navigational exploration.MethodsWe introduce a new multi-topic assignment method combining semantic annotation using UMLS concepts (CUIs) and Formal Concept Analysis (FCA). Each question was tagged with CUIs identified by MetaMap. The CUIs were filtered with term-frequency and a new term-strength index to construct a CUI-question context. The CUI-question context and a topic-subject context were used for multi-topic assignment, resulting in a topic-question context. The topic-question context was then directly used for constructing a prototype navigational exploration interface.ResultsExperimental evaluation was performed on the task of automatic multi-topic assignment of 99 predefined topics for about 60,000 consumer health questions from NetWellness. Using example-based metrics, suitable for multi-topic assignment problems, our method achieved a precision of 0.849, recall of 0.774, and F1 measure of 0.782, using a reference standard of 278 questions with manually assigned topics. Compared to NetWellness’ original topic assignment, a 36.5% increase in recall is achieved with virtually no sacrifice in precision.ConclusionEnhancing the recall of multi-topic assignment without sacrificing precision is a prerequisite for achieving the benefits of navigational exploration. Our new multi-topic assignment method, combining term-strength, FCA, and information retrieval techniques, significantly improved recall and performed well according to example-based metrics.

Multi-Agent Reinforcement Learning – An Exploration Using Q-Learning

It is possible to exploit automated learning from sensed data for practical applications - in essence facilitating reasoning about particular problem domains based on a combination of environmental awareness and insights elicited from past decisions. We explore some enhanced Reinforcement Learning (RL) methods used for achieving such machine learning using software agents in order to address two questions. Can RL implementations/methods be accelerated by using a Multi-Agent approach? Can an agent learn composite skills in single-pass?

Trial prospector: Matching patients with cancer research studies using an automated and scalable approach

Cancer is responsible for approximately 7.6 million deaths per year worldwide. A 2012 survey in the United Kingdom found dramatic improvement in survival rates for childhood cancer because of increased participation in clinical trials. Unfortunately, overall patient participation in cancer clinical studies is low. A key logistical barrier to patient and physician participation is the time required for identification of appropriate clinical trials for individual patients. We introduce the Trial Prospector tool that supports end-to-end management of cancer clinical trial recruitment workflow with (a) structured entry of trial eligibility criteria, (b) automated extraction of patient data from multiple sources, (c) a scalable matching algorithm, and (d) interactive user interface (UI) for physicians with both matching results and a detailed explanation of causes for ineligibility of available trials. We report the results from deployment of Trial Prospector at the National Cancer Institute (NCI)-designated Case Comprehensive Cancer Center (Case CCC) with 1,367 clinical trial eligibility evaluations performed with 100% accuracy.

Autonomous Robot Control Using Evidential Reasoning

Evidence theory has been widely applied to uncertainty reasoning. In this paper a finite state machine with evidential reasoning is proposed to control autonomous robots. The Khepera robot is used to demonstrate the control system. The experimental results show that the control system is able to automatically control the robot behaviours via evidential control rules, which can be designed by humans or automatically extracted using machine learning approaches. Therefore, the control system provides an efficient approach to transfer human knowledge to an autonomous robot.

Incremental Knowledge Base for Uncertain Reasoning

Evidence theory has been widely applied to uncertain reasoning. However, the evidence space and hypothesis space are each defined as a fixed set. If the theory is applied to solve specific problems, the corresponding evidence and hypothesis spaces have to be defined at the outset. In the real world, there are lots of systems with incremental spaces of evidence and hypotheses. It is difficult to apply the original evidence theory to such dynamic systems. Therefore, in this paper an incremental knowledge base is proposed to deal with dynamic knowledge bases. In the incremental knowledge base, a set of definitions in evidence theory are generalized. Based on the definitions, a set of operations for the incremental knowledge base is provided. A well known example is used to demonstrate the knowledge base in a reasoning system.

Skill combination for reinforcement learning

Recently researchers have introduced methods to develop reusable knowledge in reinforcement learning (RL). In this paper, we define simple principles to combine skills in reinforcement learning. We present a skill combination method that uses trained skills to solve different tasks in a RL domain. Through this combination method, composite skills can be used to express tasks at a high level and they can also be re-used with different tasks in the context of the same problem domains. The method generates an abstract task representation based upon normal reinforcement learning which decreases the information coupling of states thus improving an agents learning. The experimental results demonstrate that the skills combination method can effectively reduce the learning space, and so accelerate the learning speed of the RL agent. We also show in the examples that different tasks can be solved by combining simple reusable skills.

Learning to select relevant perspective in a dynamic environment

When an agent observes its environment, there are two important characteristics of the perceived information. One is the relevance of information and the other is redundancy. The irrelevant and redundant features which commonly exists within an environment, commonly leads to agent state explosion and associated high computational cost within the learning process. This paper presents an efficient method concerning both the relevance of information and the correlation in order to improve the learning of reinforcement learning agent. We introduce a new concurrent online learning method to calculate the match count C(s) and relevance degree I(s) to quantify the redundancy and correlation of features with respect to a desired learning task. Our analysis shows that the correlation relationship of the features can be extracted and projected to concurrent biased learning threads. By comparing the commonalities of these learning threads, we can evaluate the relevance degree of a feature that contributes to a particular learning task. We explain the method using random walk examples and then demonstrate the method on the chase object domain. Our validation results show that, using the concurrent learning method, we can efficiently detect redundancy and irrelevant features from the environment on sequential tasks, and significantly improve the efficiency of learning. After relevant features are extracted, the agent can remarkably accelerate its succeeding learning speed.