Julien Henriet

Case-Based Reasoning adaptation of numerical representations of human organs by interpolation

Case-Based Reasoning (CBR) and interpolation tools can provide solutions to unknown problems by adapting solutions from other problems already solved. We propose a generic approach using an interpolation tool during the CBR-adaptation phase. The application EquiVox, which attempts to design three dimensional representations of human organs according to external measurements, was modelled. It follows the CBR-cycle with its adaptation tool based on Artificial Neural Networks and its performances are evaluated and discussed. The results show that this adaptation tool meets the requirements of radiation protection experts who use such prototypes and also what the limits are of such tools in CBR-adaptation. When adaptations are guided by experience grained through trial and error by experts, interpolation tools become well-suited methods for automatically and quickly providing adaptation strategies and knowledge through training phases.

Towards an optimistic management of concurrency: a probabilistic study of the pilgrim protocol

In CSCW applications, users modify shared objects in real-time. Thus, concurrency management protocols are required in order to maintain consistency. Such protocols can be classified as optimistic or pessimistic. Our Pilgrim protocol is pessimistic since it is based on ownership. Our new version of this protocol is optimistic and designed to minimize the delay before writing. This paper presents this new version based on atomization and multi-versioning and compares it to the former one through a probabilistic study. Finally, this study allows us to highlight the parameters that make it possible to choose between the two protocols studied.

Introduction of a combination vector to optimise the interpolation of numerical phantoms

Phantoms are 3-dimensional (3D) numerical representations of the contours of organs in the human body. The quality of the dosimetric reports established when accidental overexposures to radiation occur is highly dependent on the phantoms reliability with respect to the subject. EquiVox is a Case-Based Reasoning platform which proposes an interpolation of the 3D Lung Contours (3DLC) of subjects during its adaptation phase. This interpolation is conducted by an Artificial Neural Network (ANN) trained to learn how to interpolate the 3DLC of a learning set (LS). ANN is a well-suited tool when known results are numerous. Since the cardinality of our learning set is restrained, the imperfections of each 3DLC have a great impact on interpolations. Thus, we explored the possibility of ignoring some of the 3DLC of LS via implementation of a new learning algorithm which associated Combination Vectors (CV) to LS. The results proved that this method could optimise interpolation accuracy. Furthermore, this study highlights the fact that some of the 3DLC were harmful for some interpolations whereas they increased the accuracy of others.

Adapting numerical representations of lung contours using Case-Based Reasoning and Artificial Neural Networks

In case of a radiological emergency situation involving accidental human exposure, a dosimetry evaluation must be established as soon as possible. In most cases, this evaluation is based on numerical representations and models of subjects. Unfortunately, personalised and realistic human representations are often unavailable for the exposed subjects. However, accuracy of treatment depends on the similarity of the phantom to the subject. The EquiVox platform (Research of Equivalent Voxel phantom) developed in this study uses Case-Based Reasoning principles to retrieve and adapt, from among a set of existing phantoms, the one to represent the subject. This paper introduces the EquiVox platform and Artificial Neural Networks developed to interpolate the subject’s 3D lung contours. The results obtained for the choice and construction of the contours are presented and discussed.

Study of an optimistic protocol for concurrency management in CSCW

Users need to access to shared objects concurrently in CSCW applications. Concurrency management protocols have been designed in order to maintain the consistency. Those protocols are either optimistic or pessimistic, like our own protocol called Pilgrim. This protocol is based on ownership and lock-unlock mechanisms. In order to minimize the delay before writing, we defined an optimistic protocol based on mechanisms such as atomization and multiversion. This paper presents both of those protocols through a finite state automaton, and we compare them computing the probabilities to be able to modify a shared object using the pessimistic version on the one hand, and the optimistic one on the other hand. Finally, we propose an accurate study of parameters that permits to choose one of the both protocols studied.

An iterative precision vector to optimise the CBR adaptation of EquiVox

The case-based reasoning (CBR) approach consists in retrieving solutions from similar past problems and adapting them to new ones. Interpolation tools can easily be used as adaptation tools in CBR systems. The accuracies of interpolated results depend on the set of known solved problems with which the interpolation tools have previously been trained. To be sufficiently accurate, an interpolation tool must be trained with a large number of known cases. However, CBR systems are also relevant if the number of known cases is restricted. In addition, the training of interpolation tools is generally seen by users as a black box. This paper presents a generic method to optimise CBR adaptations driven by trained interpolation tools and also takes into account remarks made by users about known solution accuracy. This method was applied to the CBR system called EquiVox which retrieves, reuses (interpolates), revises and retains three-dimensional numerical representations of organ contours and thus enhances its own performance.

INTRODUCTION OF A MULTIAGENT PARADIGM TO OPTIMIZE A CASE-BASED REASONING SYSTEM DESIGNED TO PERSONALIZE THREE-DIMENSIONAL NUMERICAL REPRESENTATIONS OF HUMAN ORGANS

The case-based reasoning (CBR) approach consists of retrieving solutions from similar past problems and adapting them to new problems. Interpolation tools can easily be used as adaptation tools in CBR systems. The accuracies of interpolated results depend on the set of known solved problems with which the interpolation tools are previously trained. EquiVox is a CBR-based system designed for retrieving and adapting three-dimensional numerical representations of human organs called phantoms. EquiVox uses an interpolation tool as an adaptation process. These phantoms are used by radiation protection experts to establish dosimetric reports in case of accidental overexposure to radiation. In addition, medical physicians need these phantoms to compute and control exposure to radiation used to treat diseases such as cancerous tumors in hospitals. The present work aims at proposing a distributed architecture for EquiVox so that a user may find a solution as quickly as possible based on the most recent available knowledge of a given community. We have designed a distributed architecture based on a multiagent paradigm and studied the theoretical performance of the new version. The ability of the new system to quickly provide and adapt solutions using the most up-to-date knowledge has been analyzed from a probabilistic angle. In the case of limited and accidental exposure to radiation, the proposed parallel processing system improves the previous and sequential version of EquiVox. Improvements are also obtained in some cases of massive exposure to radiation.

EquiVox: an example of adaptation using an artificial neural network on a case-based reasoning platform

In case of a radiological emergency situation involving accidental human exposure, a dosimetry evaluation must be established as soon as possible. In most cases, this evaluation is based on numerical representations and models of victims. Unfortunately, personalised and realistic human representations are often unavailable for the exposed subjects. However, accuracy of treatment depends on the similarity of the phantom to the victim. The EquiVox platform (Research of Equivalent Voxel phantom) developed in this study uses Case-Based Reasoning (CBR) principles to retrieve and adapt, from among a set of existing phantoms, the one to represent the victim. This paper introduces the EquiVox platform and the Artificial Neural Network (ANN) developed to interpolate the victims 3D lung contours. The results obtained for the choice and construction of the contours are presented and discussed.

PRELIMINARY STUDY OF A NEW CBR-BASED APPLICATION FOR VOXELISED PHANTOM CREATION: REEPH

In the domain of radiation protection it is not always possible to perform an additional examination such as scanners or Magnetic Resonance Imaging (MRI) for a patient who has been accidentally radiated. However a medical diagnostic must be made as soon as possible to calculate the dosimetric balance. Currently this incidental calculation is based on the available voxelised phantoms which are 3D numerical reconstructions of the human body with the internal organs. The Case-Based Reasoning (CBR) is seen on the one hand like a problem solving method and on the other hand like a technology for the conception of intelligent systems. The ReEPh project (Research of Equivalent Phantom) strikes a new path in the field of problem solving methods in the radiation protection and uses the approach of the CBR to retrieve the set of the phantoms the most adapted to the irradiated victim. For this first version of ReEPh, the retrieval phase uses a Knn Algorithm (K nearest neighbours). We propose a measure of similarity and a confidence index to take into account the uncertainty implied by the possible missing characteristics of the victim. We have developed a graphic interface to view the cases retrieval, and to visually illustrate the combination of measures of similarity and confidence index.

Artificial Intelligence-Virtual Trainer: An educative system based on artificial intelligence and designed to produce varied and consistent training lessons

AI-Virtual Trainer is an educative system using Artificial Intelligence to propose varied lessons to trainers. The agents of this multi-agent system apply case-based reasoning to build solutions by analogy. However, as required by the field, Artificial Intelligence-Virtual Trainer never proposes the same lesson twice, whereas the same objective may be set many times consecutively. The adaptation process of Artificial Intelligence-Virtual Trainer delivers an ordered set of exercises adapted to the objectives and sub-objectives chosen by trainers. This process has been enriched by including the notion of distance between exercises: the proposed tasks are not only appropriate but are hierarchically ordered. With this new version of the system, students are guided towards their objectives via an underlying theme. Finally, the agents responsible for the different parts of lessons collaborate with each other according to a dedicated protocol and decision-making policy since no exercise must appear more than once in the same lesson. The results prove that Artificial Intelligence-Virtual Trainer, however perfectible, meets the requirements of this field.