This article is a position paper on strategic argumentation in multi-agent systems. We study argumentation persuasive dialogues between two agents, one persuader and one persuadee. We analyze a real case study in the field of marketing, not studied before in the automated argumentation literature.
Daniel Gibert, Javier Béjar, Carles Mateu, Jordi Planes, Daniel Solis, Ramon Vicens
221 - 226
Traditional signature-based methods have started becoming inadequnate to deal with next generation malware which utilize sophisticated obfuscation (polymorphic and metamorphic) techniques to evade detection. Recently, research efforts have been conducted on malware detection and classification by applying machine learning techniques. Despite them, most methods are build on shallow learning architectures and rely on the extraction of hand-crafted features. In this paper, based on assembly language code extracted from disassembled binary files and embedded into vectors, we present a convolutional neural network architecture to learn a set of discriminative patterns able to cluster malware files amongst families. To demonstrate the suitability of our approach we evaluated our model on the data provided by Microsoft for the BigData Innovators Gathering 2015 Anti-Malware Prediction Challenge. Experiments show that the method achieves competitive results without relying on the manual extraction of features and is resilient to the most common obfuscation techniques.
In type 1 diabetes management, mobile health applications are becoming a cornerstone to empower people to self-manage their disease. There are many applications addressed to calculate insulin doses based on the current information (e.g. carbohydrates intake) and a few of them are accompanied by modules able to supervise postprandial conditions and recommend corrective actions if the user falls in an abnormal state (i.e. hyperglycaemia or hypoglycaemia). On the other hand, mobile apps favour the gathering of historical data from which machine learning techniques can be used to predict if user conditions will worsen.
This work presents the application of k-nearest neighbour on the historical data gathered on patients, so that given the information related to a sequence of meals, the method is able to predict if the patient will fall in an abnormal condition. The experimentation has been carried out with the UVA-Padova type 1 diabetes simulator over eleven adult profiles. Results corroborate that the use of sequential data improve significantly the prediction outcome when forecasts distinguish the type of meal (breakfast, lunch and dinner).
Nicola Covallero, David Martínez, Guillem Alenyà, Carme Torras
233 - 238
Manipulation planning of cluttered objects involves a mixture of symbolic and geometric constraints which makes such planning very time consuming and often unsuitable for real applications. We propose to divide the geometric restrictions in two groups. The ones in the first group are used to generate a set of symbolic states used for planning. The evaluation of the ones in the second group is delayed after planning, and only relevant ones are evaluated when necessary. We demonstrate our proposal in a simple but effective implementation using pushing and grasping actions.
Recommender Systems have become a fundamental part of various applications supporting users when searching for items they could be interested in, at a given moment. However, the majority of Recommender Systems generate isolate item recommendations based mainly on user-item interactions, without taking into account other important information about the recommendation moment, able to deliver users a more complete experience. In this paper, a hybrid Case-based Reasoning model generating recommendations of sets of music items, based on the underlying structures found in previous playlists, is proposed. Furthermore, the described system takes into account the similarity of the basic contextual information of the current and the past recommendation moments. The initial evaluation shows that the proposed approach may deliver recommendations of equal and higher accuracy than some of the widely used techniques.
Joan P. Company-Corcoles, Alberto Ortiz, Pep Lluis Negre-Carrasco
245 - 250
Recent advances in technology have allowed robotic platforms to implement some tasks currently accomplished by humans. Of particular relevance is the fact that this technology is allowing to perform tasks in hazardous environments. Underwater environments fall into this category, specially when carried out at high depth areas, naturally risky for divers. An important capability, lately affordable by autonomous underwater robots, is to grasp objects from the sea floor. To accomplish this task, the target has to be found into the scene and its pose estimated to guide the robot manipulator. In this work, we present an object detection and pose estimation pipeline which guides the target grasping during an intervention operation. The proposed algorithm is based on two stages: detection and tracking. The detection process takes care of detecting the object pose for the first time, or when it is lost by the second stage. On the other hand, the tracking process, less computationally demanding, is responsible for correcting some small odometry inaccuracies or for adjusting the current pose to align better with the object. To conclude, we report on the results of a number of experiments, carried out by an underwater robot operating in various environments, a water tank and in the sea, and using a stereo camera as the input sensor.
Martí Sánchez-Fibla, Clément Moulin-Frier, Xerxes Arsiwalla, Paul Verschure
251 - 256
A first step to reach Theory of Mind (ToM) abilities (attribution of beliefs to others) in synthetic agents through sensorimotor interactions, would be to tag sensory data with agent typology and action intentions: autonomous agent X moved an object under the box. We propose a dual arm robotic setup in which ToM could be probed. We then discuss what measures can be extracted from sensorimotor interaction data (based on a correlation analysis) in the proposed setup that allow to distinguish self than other and other/inanimate from other/active with intentions.
We finally discuss what elements are missing in current cognitive architectures to be able to acquire ToM abilities in synthetic agents from sensorimotor interactions, bottom-up from reactive agent interaction behaviors and top-down from the optimization of social behaviour and cooperation.
Josep L. de la Rosa, Andres El-Fakdi, Victor Torres, Xesca Amengual
257 - 262
We propose a novel method to detect, locate and classify logos in images, based on consensus to enable Blockchain implementations. An incremental learning algorithm is proposed to detect logos by just using a synthetic image template, without the need of annotating a training set. Then, a crowdsourced solution is generated to carry out the consensus of the executions of the incremental learning.
Pablo Almajano, Dolça Tellols, Inmaculada Rodriguez, Maite Lopez-Sanchez
263 - 268
This paper proposes the usage of an Embodied Conversational Agent (ECA) as user interface for achieving structured tasks seamlessly. Specifically, we present the so called METO Agent, a Motivated and Emotional Task-Oriented ECA that is represented in a 3D Virtual Scene, is able to engage in a task oriented conversation with a person in natural language and by expressing different facial expressions and gestures. Our ECA is endowed with personality traits and has the need of getting attention from the user. Along the paper we refer to Virtual Pitonisa, an ECA we have deployed as a METO Agent that represents a real fortune-teller.
Gilberto Rivera, Jorge Rodas-Osollo, Pedro Bañuelos, Marcela Quiroz, Mario Lopez
269 - 274
A good management of operating rooms is an important concern for hospitals. Being aware of that fact, the public hospital “Hospital General de Cd. Juárez” has specially announced it as a priority objective to reach. After analyzing the scheduling process at that hospital, we proposed to the hospital Administration to apply a genetic algorithm to schedule its surgical procedures. Our implementation is ad-hoc, satisfying the particular needs of that institution. To validate our proposal, we considered the current institutional heuristic for scheduling by the hospital personnel. The experimental results give evidence in favor of the several benefits obtained by this kind of heuristic.
Jorge Rodas-Osollo, Karla Olmos-Sanchez, Yadira Ortiz-Chow, Alberto Ochoa-Ortiz
275 - 280
Objective – To present the Knowledge Management on a Strategy for Requirements Engineering (KMoS-RE) as a systematic way to elicit, structure and explicit Specialized Knowledge as convenient requirements of Informally Structured Domains that have Distributed Tacit Knowledge often associated with an ad hoc Collaborative Network.
Method – A summary of the KMoS-RE strategy applied to a project of HVAC industrial design is presented.
Results – An Integral AIntelligent Solution was provided.
Conclusion – KMoS-RE is a strategy that allows working with Distributed Tacit Knowledge of an ad hoc Collaborative Network for obtaining Explicit Knowledge that has been capitalizated in an Integral AIntelligent Solution whose implementation achieved the complete client satisfaction.
Miller Stiven Espinosa, Javier Antich, Alberto Ortiz
281 - 286
Navigation is an essential task for any mobile robot, whose primary focus is to guide the robot from its initial position to a specific target position while avoiding collisions with obstacles along the way. When talking about reactive navigation, we are referring to the same task described above, but now some restrictions are added to the way in which that task can be performed. Basically, a robot which navigates reactively makes decisions using only the information that, at that moment, its sensors collect. Acting that way, these robots are able to react quickly to any unexpected event (for instance, an obstacle). Unfortunately, the reactive navigation paradigm also has limitations, the most important being its inability to make robots avoid certain obstacles. In this work, we take as our starting point one of the best algorithms to date to make a robot navigate reactively; specifically, this algorithm is the well-known Nearness Diagram (ND). We modify ND to provide it with the ability to overcome complex obstacles, i.e. obstacles of great size and with intricate shapes. Moreover, this modification is carried out without losing the reactive nature of the ND algorithm. improved Nearness Diagram (iND) is the name given to the algorithm resulting from this modification. We test iND under simulation in a set of environments of increasing complexity, and we compare its results with those obtained by two other reactive navigation algorithms: namely, the Virtual Force Field (VFF) algorithm and the original ND algorithm.
Cristóbal Raya, Francisco J. Ruiz, Cecilio Angulo, Albert Samà, Núria Agell
287 - 292
In this paper we use the idea of conceptual space introduced by Boden and redefine some properties such appropriateness and relevance that facilitate the computational implementation of the transformational creativity mechanism. While appropriateness can only be evaluated by an expert, relevance can be objectively measured for any spectator. Computational creativity is based on the relationship between appropriateness and relevance of a concept, and therefore a computational system can be used to support this task. The paper analyses this relationship in the field of music in order to obtain a computer tool to support the musical composition task.