Knowledge-based Exploration Research Articles

Fusing domain knowledge and reinforcement learning for home integrated demand response online optimization

Electricity–gas integrated household energy systems (HESs) expose obvious system uncertainties, requiring their integrated demand response (IDR) programs to be able to adapt automatically and quickly to system changes. Deep Reinforcement Learning (DRL) methods, though having been proven promising to tackle such problems, are typically not efficient in learning from random explorations. This paper proposes a method based on DRL with HESIDR knowledge penetration. By interpreting the domain IDR knowledge as a set of control rules, the DRL agent gains learning samples from knowledge-based exploration in addition to the traditional exploration–exploitation tradeoff. Correspondingly, we develop a cooperation scheme for action selection and replay sampling, which is based on exponential probability functions to balance the penetration of knowledge-based exploration, random exploration and policy exploitation. We conduct case studies in a typical home multi-energy system environment. After determining parameters in the exponential probability functions, the learning and cost reduction performance of the proposed algorithm was tested. The results show that the method proposed in the present study spends 48.78% less training time than the standard DQN, which enables few-minute optimization on a lightweight PC. Also, the proposed method can further reduce energy bills by 26.17% compared to the uncontrolled scenario and by 9.88% to the integrated rule-based controller.

Educing knowledge from text: semantic information extraction of spatial concepts and places

Abstract. A growing body of geospatial research has shifted the focus from fully structured to semistructured and unstructured content written in natural language. Natural language texts provide a wealth of knowledge about geospatial concepts, places, events, and activities that needs to be extracted and formalized to support semantic annotation, knowledge-based exploration, and semantic search. The paper presents a web-based prototype for the extraction of geospatial entities and concepts, and the subsequent semantic visualization and interactive exploration of the extraction results. A lightweight ontology anchored in natural language guides the interpretation of natural language texts and the extraction of relevant domain knowledge. The approach is applied on three heterogeneous sources which provide a wealth of spatial concepts and place names.

An Exploration On-demand Article Recommender System for Cancer Patients Information Provisioning


 
 
 Information provision plays an important role in ed- ucating patients with serious illnesses, like cancer, to cope with their disease conditions and to actively partic- ipate in shared-decision making process. Recent stud- ies suggest that there is a lack of appropriate educa- tional resources for such patients, specifically prostate cancer patients. To address this issue, in this paper, a Knowledge-based Exploration on-demand article Rec- ommender System (called KERS) is proposed that can provide evidence-based information for patients. Rec- ognizing the fact that exploration is expensive when the user of the system is a human, the main idea in KERS is to minimize exploration while achieving the maximum long-term satisfaction. Therefore, using a knowledge- base developed by an expert in the field, KERS learns user interests as quickly as possible and then it ex- ploits this knowledge to recommend the best articles. Furthermore, KERS needs no information from users beforehand and it learns them through interacting with users. The system will help patients make informed de- cisions, and at the same time, will reduce the burden on the healthcare providers. The results of experiments have confirmed the effectiveness of the proposed system compared to baseline methods.
 
 

A Study of Qualitative Knowledge-Based Exploration for Continuous Deep Reinforcement Learning

A Study of Qualitative Knowledge-Based Exploration for Continuous Deep Reinforcement Learning

Immunoformatics and Molecular Simulation Study Unfold the Immunological Repertoire of Sabia Virus Glycoprotein G1

Sabia virus Glycoprotein G1 (GP1) is a peripheral membrane protein involved in viral adsorption and consequential pathogenic infection. Glycoprotein G1 (GP1) was subjected to immunoformatic and molecular simulation study to predict effective B-cell and T-cell epitopes with knowledge-based exploration of probable immunological responses the epitopes may elicit. The linear and conformational epitopes of GP1 for B-cell were predicted by immunoformatic tools housed in IEDB recourse. Both CD4+ and CD8+ T-cell epitopes were predicted exploiting IEDB recourse tools whereas CD8+ T cell epitopes were severed further by immune response evoking ability of epitope-MHC complex. Contribution of GP1 to innate antiviral response was evolved by interferon (IFN)-gamma inducing and imitation capacity by immunoformatic and docking study respectively. The results of the B-cell epitope analysis suggest the occurrence of potential linear and conformational epitope with cross-reactivity. A range of T-cell epitopes were assumed to be involved in MHC class I and MHC class II molecule dependent antigen presentation. Glycoprotein G1 (GP1) may induce the IFNgamma but its IFN-gamma mimicking ability (confirmed by docking study) that adduces a prompt immune defense. Therefore, inherent immunological repertoire of GP1 epitopes can be taken in to advantage in future immunization regiment development against Sabia virus.

Exploring Research Trends with Rexplore

Current systems for exploring scholarly data exhibit a number of shortcomings in their ability to facilitate the identification of research trends and identify 'interesting' connections between researchers. To address these issues we have developed Rexplore, a novel system which combines statistics, human-computer interaction, and semantic technologies, to support knowledge-based exploration and visualization of scholarly data. In this paper we focus on the functionalities provided by Rexplore for visualizing research trends and we use as an example research in Social Networks, which has experienced dramatic growth in the years 2000-2010.

An Extensible Framework to Sort out Nodes in Graph-Based Structures Powered by the Spreading Activation Technique

This paper presents an extensible framework for the Spreading Activation technique. This technique is supported by the ONTOSPREAD framework enabling the development, configuration, customization, and execution of the Spreading Activation method on graph-based structures. It has been used for a long time to the efficient exploration of knowledge bases built on semantic networks in Information and Document Retrieval domains. The emerging Web of Data and the sheer mass of information now available make it possible the deployment of new services and applications based on the reuse of existing vocabularies and datasets. A large amount of this information is published using semantic web languages and formats such as RDF, implicit graph structures developed using W3C standard languages; but new flexible and scalable methods to create added-value services and exploit the data are required. That is why ONTOSPREAD is considered relevant in providing a new way to implement the double process of activation and spreading of concepts in graph-based structures, more specifically to browse and rank resources in the Web of Data realm. The original constraints like weight degradation according to the distance are provided in combination with others coming from the extension of this technique like the converging paths reward. Finally an evaluation methodology and two examples using the well-known ontologies GALEN and SNOMED CT are presented to validate the goodness, the improvement, and the capabilities of this technique.

Crystallographic Refinement by Knowledge-Based Exploration of Complex Energy Landscapes

Although X-ray crystallography remains the most versatile method to determine the three-dimensional atomic structure of proteins and much progress has been made in model building and refinement techniques, it remains a challenge to elucidate accurately the structure of proteins in medium-resolution crystals. This is largely due to the difficulty of exploring an immense conformational space to identify the set of conformers that collectively best fits the experimental diffraction pattern. We show here that combining knowledge-based conformational sampling in RAPPER with molecular dynamics/simulated annealing (MD/SA) vastly improves the quality and power of refinement compared to MD/SA alone. The utility of this approach is highlighted by the automated determination of a lysozyme mutant from a molecular replacement solution that is in congruence with a model prepared independently by crystallographers. Finally, we discuss the implications of this work on structure determination in particular and conformational sampling and energy minimization in general.

Organisation and expertise: An exploration of knowledge bases and the management of accounting and consulting firms

Existing theories have recognised the importance of expert knowledge in the formation and survival of professional service firms (PSFs), such as accounting and consulting firms, but have not fully explained its role. We argue that knowledge is a key determinant of the organisational structure and performance of PSFs. We examine forms of knowledge and knowledge management strategies within PSFs and develop a framework linking the firm's knowledge base and organisational structure. Two contrasting case studies are presented to illustrate these arguments. The paper concludes with a set of propositions to guide further research.