Abstract: The temporal dimension is highly present in almost all critical medical scenarios (e.g. intensive care units, burn units or cardiology departments), considering that the temporal evolution of the patient is a key factor in providing an effective healthcare. While the irruption of new technologies in hospital services provides large amounts of data, knowledge‐based systems are not often considered by clinicians, as part of the clinical information process. Case‐based reasoning (CBR) is a field of artificial intelligence that tackles new problems by referring to analogous problems that have already been solved in the past. Therefore, CBR seems to be an effective approach in medical domains since cases refer directly to patient episodes within the health records. Despite the importance of the temporal dimension, no in‐depth study of the impact of time on the CBR systems has been carried out in critical medical domains. This work focuses on the development of case retrieval systems based on their temporal similarity in the medical domain. We present T‐CARE, a temporal case retrieval system that combines classical and non‐classical approaches to measure temporal similarity of cases which are composed of temporal sequences of time point events and intervals. Finally, we show its practical implementation in an intensive care burn unit.

The early diagnosis and the correct therapy for generalized infections is an important factor for patient survival in intensive care burn units (ICBUs). Due to the number of pathologies involved, there is not a specific etiology and, therefore, it is difficult for physicians to quantify the patient severity to state the diagnosis. In this scenario, CBR finds problems to obtain a reliable solution when retrieved cases are highly similar. For example, in ICBU patients slight variations of monitored parameters have a deep impact on the patient's severity evaluation. Therefore, it seems necessary to extend the system outcome in order to indicate the reliance of the solution obtained. Main efforts in the literature for CBR evaluation focus on case retrieval (i.e. similarity) or on a retrospective analysis. However, these approaches do not seem to suffice when cases are very close. In this work, we propose and implement a CBR system to state the chance of a patient to survive. The system has been tested using a database of 89 patients from an ICBU, obtaining about 76% accuracy. Furthermore, in order to evaluate the behaviour of the CBR system in this kind of scenarios, we propose three techniques to obtain a reliance solution degree, one based on case retrieval and two based on case reuse.

In this work we address an approach for solving the problem of building a temporal constraint network from the set of frequent sequences obtained after a temporal data mining process. In particular, the temporal data mining algorithm used is TSET [7], an algorithm based on the inter-transactional framework that uses a unique tree-based structure to discover frequent sequences from datasets. The model of temporal network is the proposed by Hadjali, Dubois and Prade [8] where each constraint is formed by three possibility values expressing the relative plausibility of each basic relations between two point-based events, that is, "before", "at the same time" and "after". We propose the use of the Shafer Theory for computing the possibility values of the temporal relations involved in the network from the calculated probability masses of the sequences. The final goal is to obtain a more understandable and useful sort of knowledge from a huge volume of temporal associations resulting after the data mining process.

For those situations in which the user wants to interact with the system by using, for example, voice commands, it would be convenient to refer to the objects by their names (e.g., "cube") instead of other types of interactions (e.g., "grasp object 1"). Thus, automatic object recognition is the first step in order to acquire a higher level of interaction between the user and the robot. Nevertheless, applying object recognition techniques when the camera images are being transmitted through the web is not an easy task. In this situation, images cannot have a very high resolution, which affects enormously the recognition process due to the inclusion of more errors while digitalizing the real image. Some experiments with the Universitat Jaume I Online Robot evaluate the performance of different neural-network implementations, comparing it to that of some distance-based object recognition algorithms. Results will show which combination of object features, and algorithms (both statistical and neural networks) is more appropriate to our purpose in terms of both effectiveness and computing time.

Within the context of online robots, a considerable amount of research has traditionally focused on the global system functionality, including the way of interaction between the user and the robot. Recent results in different robotics areas have demonstrated the potential of a number of techniques from the Pattern Recognition and Machine Learning domains, although very few work has been specifically addressed to online robots, where the object recognition is directly performed by the user. In this paper, we investigate the feasibility of using a neural network approach to object recognition in the context of online robots, and discuss the main advantages over the application of statistical learning methods. Some experiments with the UJI (Universitat Jaume I) online robot evaluate the performance of different neural network implementations, comparing it to that of some distance-based object recognition algorithms.

We introduce a new model, MPDN, for quantitative temporal reasoning with points and durations, that supposes an extension of the TCSP formalism and previous point-duration network models. The problem of deciding consistency for a MPDN is shown to be NP-complete. So, we identify a tractable fragment, named simple MPDN, that subsumes the STP model and allows for duration reasoning. Necessary and sufficient conditions for deciding consistency of a simple MPDN are used to design an algorithm for consistency checking, whose time complexity is cubic in the number of variables. This is a significant improvement, not only in computational complexity but also in simplicity, over previous non-specific algorithms that can be applied to solve the consistency problem.

This paper describes a web-based telerobotic system with two basic capabilities: object recognition and incremental or continual learning. Within this context, the present paper investigates the feasibility of applying several distance-based classifiers to remote object recognition tasks. On the other hand, incremental learning enables the system to maintain a representative set of past training examples that are used together with new data to appropriately modify the currently held knowledge. This can constitute a very significant advantage because it allows a system to continually increase its knowledge and accordingly, to enhance its performance in recognition as it learns. And also, because this capability allows to detect an incorrect or anomalous employment of this particular remote system. The performance of the system is evaluated by means of an empirical analysis addressed to assess two competing goals: computing time and recognition accuracy.

The system consists of a multirobot architecture that gives access to an educational and an industrial robot, through the Internet, by using advanced multimedia and distributed programming tools like Java, Java3D and CORBA. The main effort is focused on the user interface that allows specification of a task into a 3D model and sending commands to the real robot once the operator approves them. This kind of interaction has the ability to consume very little bandwidth in order to perform a remote manipulation via the Internet. The user interface offers augmented and virtual reality possibilities to the users, as well as object recognition in order to specify commands to the robot in a natural way. This high-level task specification would not be possible without the grasping determination module, which suggests manipulation characteristics for every scene's object.

The disjunctive fuzzy temporal constraint network (DFTCN) model is a framework that allows reasoning with fuzzy qualitative and quantitative complex temporal constraints. However, its general complexity is exponential. In this work, we review methods that can reduce this complexity, based on graph topology and preprocessing. The final goal is to integrate all these methods into a general tool for temporal reasoning with DFTCN.

Generic reasoning models facilitate the construction of knowledge-based systems for solving complex problems, such as therapy planning in an agricultural context. The basic features of these models are problem-solving methods, which are proposed to carry out the tasks ordinarily done by experts in solving a specific problem. This article describes an expertise model in the domain of plant health, which was obtained applying the CommonKADS methodology. The most important result is the abductive method proposed for supplying a solution to treatment problems in domains where there is no protocol for therapy planning tasks. The description of the method is based on the algebraic formulation of a domain-independent general treatment. Therefore, the abductive method proposed is sufficiently generalized to be applied to different domains.

In this work we present the design proposal for a therapeutic specialist TS in the domain of the intensive coronary care unit ICCU for situations of acute myocardial infarct AMI . This TS can be placed within a more general framework of the intelligent supervision of patients interned in an ICCU, and it carries out two basic functions: 1 recommending to the physician that therapy should be administered according to the state and response of the treatment in progress; and 2 to continually monitor the evolution of the patient with the therapy in progress, advising the physician of necessary therapeutic modifications associations with other drugs, suspension of treatment, etc. in order to compensate for the evolution of the patient and aid his recovery. Thus we present a specialist agent architecture designed to carry out both functions, which is based on the identification, design, and implementation of therapeutic tasks, on the explicit separation between knowledge of the domain and the metaknowledge necessary in order for the coordinated accomplishment of these tasks, and on mimicking the methodology followed by the physician when evaluating the monitoring of the patient's condition.