Antonio A. Sánchez-Ruiz

Personality aware recommendations to groups

In this article we introduce a novel method of making recommendations to groups based on existing techniques of collaborative filtering and taking into account the group personality composition. We have tested our method in the movie recommendation domain and we have experimentally evaluated its behavior under heterogeneous groups according to the group personality composition.

Least Common Subsumer Trees for Plan Retrieval

This paper presents a new hierarchical case retrieval method called Least Common Subsumer Trees (LCS trees). LCS trees perform a hierarchical clustering of the cases in the case base by iteratively computing the least-common subsumer of pairs of cases. We show that LCS trees offer two main advantages: First, they can enhance the accuracy of the CBR system by capturing regularities in the case base that are not captured by the similarity measure. Second, they can reduce retrieval time by filtering the set of cases that need to be considered for retrieval. We present and evaluate LCS trees in the context of plan retrieval for plan recognition, and present procedures for both assessing similarity and computing the least common subsumer of plans using refinement operators.

Refinement-Based Similarity Measure over DL Conjunctive Queries

Similarity assessment is a key operation in case-based reasoning and other areas of artificial intelligence. This paper focuses on measuring similarity in the context of Description Logics (DL), and specifically on similarity between individuals. The main contribution of this paper is a novel approach based on measuring similarity in the space of Conjunctive Queries, rather than in the space of concepts. The advantage of this approach is two fold. On the one hand it is independent of the underlying DL, and thus, there is no need to design similarity measures for different DL, and on the other hand, the approach is computationally more efficient than searching in the space of concepts.

Abstraction in Knowledge-Rich Models for Case-Based Planning

Abstraction in case-based planning is a mechanism for plan retrieval and adaptation. An abstract case is a generalization of a concrete case that can be reused in different situations to that where the original case was obtained. Additional knowledge is also required to instantiate an abstract case for a new concrete solution. In this paper, we show how the cases built by a generative planner, that uses Description Logics to represent knowledge-rich models of the state of the world, can be automatically abstracted by using the same knowledge model. An algorithm for case abstraction is presented, along with the conditions that a new problem must fulfill for being solvable by an abstract case.

Adaptation through Planning in Knowledge Intensive CBR

Adaptation is probably the most difficult task in Case-Based Reasoning (CBR) systems. Most techniques for adaptation propose ad-hoc solutions that require an effort on knowledge acquisition beyond typical CBR standards. In this paper we demonstrate the applicability of domain-independent planning techniques that exploit the knowledge already acquired in many knowledge-rich approaches to CBR. Those techniques are exemplified in a case-based training system that generates a 3D scenario from a declarative description of the training case.

Narrative Balance Management in an Intelligent Biosafety Training Application for Improving User Performance

The use of three-dimensional virtual environments in training applications supports the simulation of complex scenarios and realistic object behaviour. While these environments have the potential to provide an advanced training experience to students, it is difficult to design and manage a training session in real time due to the number of parameters to pay attention to: timing of events, difficulty, user’s actions and their consequences or eventualities are some examples. For that purpose, we have extended our virtual Bio-safety Laboratory application used for training biohazard procedures with a Narrative Manager. The Narrative Manager controls the simulation deciding which events will take place in the simulation, and when, by controlling the narrative balance of the session. Our hypothesis is that the Narrative Manager allows us to increase the number of tasks for the user to solve and, due to balancing difficulty and intensity, it keeps the user interested in training. When evaluating our system we observed that the Narrative Manager effectively introduces more tasks for the user to solve, and despite that, is accepted by the users as more interesting and not harder than an identical system without a Narrative Manager. Also, a knowledge test demonstrated better results in users’ interest and learning output in the narrative condition.

Similarity Metrics from Social Network Analysis for Content Recommender Systems

Online judges are online systems that test programs in programming contests and practice sessions. They tend to become big problem live archives, with hundreds, or even thousands, of problems. This wide problem statement availability becomes a challenge for new users who want to choose the next problem to solve depending on their knowledge. This is due to the fact that online judges usually lack of meta information about the problems and the users do not express their own preferences either. Nevertheless, online judges collect a rich information about which problems have been attempted, and solved, by which users. In this paper we consider all this information as a social network, and use social network analysis techniques for creating similarity metrics between problems that can be then used for recommendation.

Betfunding: A Distributed Bounty-Based Crowdfunding Platform over Ethereum

Blockchain, the technology behind Bitcoin, is a permissionless distributed database which allows distributed storage and computation over a large network of nodes. This technology has been applied recently to many other fields besides e-currencies (Bitcoin 2.0 projects). In this paper we present Betfunding: a blockchain-based decentralized crowdfunding platform. On the contrary of regular crowdfunding platforms, our system does not require a central and reliable organization. In Betfutding users bet whether the project will or will not be implemented in a given time frame, increasing the bounty and incentive for potential developers to carry it out.

Authoring Behaviour for Characters in Games Reusing Abstracted Plan Traces

Authoring the AI for non-player characters (NPCs) in modern video games is an increasingly complex task. Designers and programmers must collaborate to resolve a tension between believable agents with emergent behaviours and scripted story lines. Behaviour trees (BTs) have been proposed as an expressive mechanism that let designers create complex behaviours along the lines of the story they want to tell. However, BTs are still too complex for non-programmers. In this paper, we propose the use of plan traces to assist designers when building BTs. In order to make this approach feasible within state-of-the-art video game technology, we generate the planning domain through an extension of the component-based approach, a widely used technique for representing entities in commercial video games.

Case-Based Recommendation for Online Judges Using Learning Itineraries.

Online judges are online repositories with hundreds or thousands of programming exercises or problems. They are very interesting tools for learning programming concepts, but novice users tend to feel overwhelmed by the large number of problems available. Traditional recommendation techniques based on content or collaborative filtering do not work well in these systems due to the lack of user ratings or semantic descriptions of the problems. In this work, we propose a recommendation approach based on learning itineraries, i.e., the sequences of problems that the users tried to solve. Our experiments reveal that interesting learning paths can emerge from previous user experiences and we can use those learning paths to recommend interesting problems to new users. We also show that the recommendation can be improved if we consider not only the problems but also the order in which they were solved.