The 4E Cognition Group has a new coordinator: Dr. Melina Gastelum!
Tom Froese will still be involved in the group as an external affiliate, and as the external supervisor of some student members. Starting this month he is the group leader of the new Embodied Cognitive Science Unit at OIST.
This workshop will be held at IIMAS, UNAM, Mexico City
Date: June 26, 2019
For further information, visit: https://www.facebook.com/events/481838489227218/
Abstracts:
Relevancia del Contexto y la Tarea para la Flexibilidad Conductual: una Arquitectura Computacional para la Robótica Cognitiva
Speaker: Diana Valenzo Macías
El presente proyecto busca inspirarse en el marco explicativo del procesamiento predictivo (PP) para modelar una arquitectura computacional cognitiva que codifique el contexto y la tarea para la ejecución de acciones en un agente artificial. El proyecto ha sido motivado por el PP debido a su interesante propuesta de la inversión del sentido en el que se pensaba que la información sensorial es procesada y a sus estrategias basadas en conocimiento estructurado dependiente del contexto y la tarea a realizar. Esta teoría propone que el cerebro se encuentra constantemente anticipándose a las entradas sensoriales que recibe mediante la generación de predicciones, enfatizando la importancia del flujo de información descendente o top-down de la percepción y la acción. El PP implica la minimización del error predictivo, mismo que se genera de la comparación de las predicciones con las entradas sensoriales y que se procesa vía ascendente o bottom-up (Clark, 2015). Este proyecto está enmarcado dentro de la robótica cognitiva, un área de investigación que se inspira en teorıas y modelos de disciplinas como la psicología y las neurociencias para investigar modelos computacionales biológicamente plausibles en agentes artificiales. Para realizar esta investigación se tomará como base el aprendizaje de máquina, en especial el aprendizaje profundo. Además, se retomarán los principios propuestos por la arquitectura cognitiva computacional denominada SOIMA (Self-Organized Internal Models Architecture, por sus siglas en inglés). La SOIMA integra los modelos internos -inverso y directo- en una sola arquitectura mediante el uso de mapas autoorganizables (SOMs, por sus siglas en inglés Self-Organizing Maps) que generan clusters o grupos de información de distintas modalidades sensoriales (visual, propioceptiva, táctil) o motoras. De esta forma, esta arquitectura permite modelar esquemas sensorimotrices multimodales en agentes (Escobar-Juárez et al., 2016). Mediante este proyecto se desea investigar e implementar el contexto y la tarea en un agente artificial. Esto se logrará basándose en la SOIMA y otras herramientas de aprendizaje profundo, buscando estudiar la importancia del contexto y la tarea en la planeación y ejecución de una acción.
Self-optimization in a Hopfield neural network based on the C. elegans connectome
Authors: Alejandro Morales and Tom Froese Speaker: Alejandro Morales
It has recently been demonstrated that a Hopfield neural network that learns its own attractor configurations, for instance by repeatedly resetting the network to an arbitrary state and applying Hebbian learning after convergence, is able to form an associative memory of its attractors and thereby facilitate future convergences on better attractors. This process of structural self-optimization has so far only been demonstrated on relatively small artificial neural networks with random or highly regular and constrained topologies, and it remains an open question to what extent it can be generalized to more biologically realistic topologies. In this work, we therefore test this process by running it on the connectome of the widely studied nematode worm, C. elegans, the only living being whose neural system has been mapped in its entirety. Our results demonstrate, for the first time, that the self-optimization process can be generalized to bigger and biologically plausible networks. We conclude by speculating that the reset-convergence mechanism could find a biological equivalent in the sleep-wake cycle in C. elegans.
Análisis de la temporalidad en modelos internos
Speaker: Eduardo Raymundo Rojas Salazar
Durante los últimos 50 años, numerosas investigaciones realizadas principalmente en el campo de las neurociencias (Wolpert y Ghahramani, 2000; Jordan y Rumelhart, 1992; Flanagan y Wing, 1997) han sustentado la hipótesis de que el sistema nervioso central (SNC) genera modelos internos, estos se construyen a través de la realización de asociaciones entre información sensorial y motora. Dichos modelos dotan al SNC con la capacidad para ensamblar y ejecutar planes motores prediciendo las consecuencias sensoriales de una acción incluso sin la ejecución de esta. Dentro de la robótica cognitiva, los modelos internos han cobrado relevancia debido que representan una alternativa para lograr que los agentes artificiales se adapten a un medio no controlado y cambiante, generando conductas motoras satisfactorias ante situaciones novedosas. Sin embargo, aún quedan preguntas pendientes fundamentales acerca de funcionamiento de los modelos internos relacionadas principalmente con las características temporales del procesamiento, las cuales podrían estar determinadas por diferencias de procesamiento entre las distintas modalidades, la interacción entre estas, así como la magnitud del error en el control motor. El presente trabajo busca profundizar en los aspectos temporales de las predicciones mediante la implementación de una tarea de levantamiento de objetos con la intención de proponer un modelo más amplio sobre el funcionamiento de los modelos internos que permita mejorar las implementaciones en agentes artificiales.
From embodied interaction to compositional referential communication: A minimal agent-based model without dedicated communication channels
Authors: Jorge I. Campos and Tom Froese
Speaker: Jorge I. Campos
Referential communication is a “representation-hungry” behavior, and the bee waggle dance is a classical example of referential communication in nature. We used an evolutionary robotics approach to create a simulation model of a minimalist example of this situation. Two structurally identical agents engage in embodied interaction such that one of them can find a distant target in 2D space that only the other could perceive. This is a challenging task: during their interaction the agents must disambiguate translational and communicative movements, allocate distinct behavioral roles (sender versus receiver), and switch behaviors from communicative to target seeking behavior. We found an evolutionary convention with compositionality akin to the waggle dance, correlating duration and angle of interaction with distance and angle to target, respectively. We propose that this behavior is more appropriately described as interactive mindshaping, rather than as the transfer of informational content.
El Sentido de Agencia en el marco de los Modelos Internos
Speaker: Dadaí Alejandra Astorga Méndez
Bajo la premisa que tener un cuerpo es necesario para la cognición y, por lo tanto, un requisito previo para el comportamiento inteligente (Pfeifer y Bongard, 2006; Smith y Gasser, 2005), la Robótica Cognitiva Corporizada, centra su atención en el diseño de agentes artificiales capaces de realizar tareas cognitivas de forma autónoma. Un tema central en ello, consiste en estudiar, inspirados en modelos provenientes de las ciencias cognitivas, el proceso mediante el cual los agentes aprenden a través de la interacción con su entorno, (Lara, Astorga, Mendoza-Bock, Pardo, Escobar y Ciria, 2018). Tomando en cuenta lo anterior, posiblemente, uno de los principales retos a resolver, es lograr que un agente artificial pueda interactuar de manera efectiva con el ambiente, de modo que, logre construir conexiones causales entre su sistema y procesos internos y el mundo externo (Ziemke, 2001), que le permitan tener la capacidad de diferenciar entre las consecuencias que tienen sus acciones en el mundo de aquellas que son producidas por alguna causa externa, es decir, agentes artificiales que presenten formas básicas de autoconciencia corporal, como lo es el Sentido de Agencia (SoA). El SoA es un componente clave de la autoconciencia corporal y refiere a la experiencia que informa a un agente sobre su influencia causal en el mundo (Braun, 2017). Permitiéndole separar los movimientos propios de los inducidos por el entorno o por otros agentes (Kannape y Blanke, 2012). Se ha establecido que el SoA se deriva de la coherencia entre nuestra intención de actuar, la ejecución voluntaria de comandos motrices específicos y los efectos derivados de ello en el mundo. La presente investigación tiene como objetivo, mediante la realización de experimentos con agentes naturales, la construcción de un modelo coherente que permita comprender la contribución de las señales motrices, perceptuales y contextuales en la emergencia del SoA.
Self-optimization in networks using unsupervised learning
Speaker: Raúl González-Cruz
Many natural dynamical systems have behaviors that can be understood, for example, as the local minimization of energy. If a dynamical system with multiple point attractors is released from an arbitrary initial condition, it will relax into a configuration that locally resolves the constraints of interdependent state variables. However, when there are many conflicting interdependencies between variables, this method will take many attempts for finding a configuration that globally optimizes these constraints. In this article, we show that a simple distributed mechanism can alter a dynamical system such that it finds lower energy configurations, more reliably and more quickly, applying an unsupervised learning technique known as Hebbian learning. If it is applied to the connections of a simple dynamical system undergoing repeated relaxation, the system will develop an associative memory that modifies the dynamics of the system such that its ability to find configurations that minimize total system energy and globally resolve conflicts between interdependent variables. We investigate the interaction of two well-known properties of complex systems that have each been independently well studied in the Hopfield network: The energy minimization behavior of dynamical systems, which can be interpreted as a local optimization of constraints; and Hebbian learning with its capacity to implement associative memory. The model is analogous in some circumstances to the behavior of multiple autonomous agents in a complex system, such as servers in a grid computing system or people in a social network, so we propose the idea of implementing associative memory in different distributed complex adaptive systems and discuss its effects on system behavior.
Evolución de señalización y altruismo en agentes artificiales Evolution of signaling and altruism in artificial agents
Speaker: José Manuel Pardo Cruz
En este trabajo intentaremos mostrar si la robótica evolutiva se presenta como una herramienta eficaz, para el estudio de los procesos evolutivos que conducen al surgimiento de habilidades cognitivas en organismos biológicos, para ello modelaremos dichos procesos utilizando algoritmos evolutivos y redes neuronales artificiales en agentes autónomos simulados.
Nos centraremos en casos que han resultado particularmente difíciles de tratar para la robótica evolutiva, como lo son el surgimiento del altruismo y la comunicación; intentaremos encontrar las condiciones mínimas necesarias para que las conductas de señalización y altruismo surjan, modelando procesos propios de la evolución en agentes artificiales simulados, lo cual podría a largo plazo permitirnos plantear una analogía con su surgimiento en agentes biológicos.
Applying Social Network Analysis to Agent-Based Models: A Case Study of Task Allocation in Swarm Robotics Inspired by Ant Foraging Behavior
Authors: Georgina Montserrat Reséndiz-Benhumea, Tom Froese, Gabriel Ramos- Fernández and Sandra E. Smith-Aguilar
Speaker: Georgina Montserrat Reséndiz-Benhumea
Social network analysis and agent-based modeling are two approaches used to study biological and artificial multi-agent systems. However, so far there is little work integrating these two approaches. Here we present a first step toward integration. We developed a novel approach that allows the creation of a social network on the basis of measures of interactions in an agent-based model for purposes of social network analysis. We illustrate this approach by applying it to a minimalist case study in swarm robotics loosely inspired by ant foraging behavior. For simplicity, we measured a network’s inter-agent connection weights as the total number of interactions between mobile agents. This measure allowed us to construct weighted directed networks from the simulation results. We then applied standard methods from social network analysis, specifically focusing on node centralities, to find out which are the most influential nodes in the network. This revealed that task allocation emerges and induces two classes of agents, namely foragers and loafers, and that their relative frequency depends on food availability. This finding is consistent with the behavioral analysis, thereby showing the compatibility of these two approaches.
City: Donostia-San Sebastián (Spain).
Date: 12, 13 & 14 July, 2019.
Venue: Ignacio María Barriola Building (Elhuyar Square, 1), University of the Basque Country.
This workshop aims to gather researchers in Evolution and 4E Cognition in order to evaluate which are the complementarities and tensions between these two approaches. The topics of the workshop will include (although will not be restricted to) the following ones:
Unfortunately, there is no extra place for speakers, but do not hesitate to contact them for joining as attendants. There will be a great deal of discussion and debate, and it will be great to have an audience willing to engage into these topics. If you plan to go, please, confirm your assistance by emailing them in the contact section of the website (provided below). The workshop is organized by Manuel Heras-Escribano and Ezequiel Di Paolo (IAS Research Centre for Life, Mind, and Society, EHU-UPV) and generously funded by the BBVA Foundation through the 2018 Leonardo Grant for Researchers and Cultural Creators entitled “La filosofía de las affordances: los orígenes ecológicos, evolutivos y sociales de la cognición [AFFORDEVOCOG]”
For more information, visit: https://evolutionandtheembodiedmind.wordpress.com/
City: Donostia-San Sebastián (Spain).
Date: 9 & 10 July, 2019.
Venue: Ignacio María Barriola Building (Elhuyar Square, 1), University of the Basque Country.
This workshop aims to gather researchers in Enaction and Ecological Psychology in order to evaluate which are the complementarities, tensions, and overlaps between these two approaches.The topics of the workshop will be divided in three main sections:
Unfortunately, there is no extra place for speakers, but do not hesitate to contact them for joining as attendants. There will be a great deal of discussion and debate, and it will be great to have an audience willing to engage into these topics. If you plan to go, please, confirm your assistance by emailing them in the contact section of the website (below). The workshop is organized by Manuel Heras-Escribano and Ezequiel Di Paolo (IAS Research Centre for Life, Mind, and Society, EHU-UPV) and generously funded by the BBVA Foundation through the 2018 Leonardo Grant for Researchers and Cultural Creators entitled “La filosofía de las affordances: los orígenes ecológicos, evolutivos y sociales de la cognición [AFFORDEVOCOG]”
For more information, visit: https://enactionandecologicalpsychology.wordpress.com/
This one-day workshop will be held on July 2nd, in Heidelberg University (Jaspers-Bibliothek des Zentrums für Psychosoziale Medizin Voss-Straße 4 (2. OG), 69115 Heidelberg).
Organizers:
Valeria Bizzari & Oren Bader
Keynote speakers:
Thomas Fuchs (Heidelberg), Joel Krueger (Exeter), Alessandro Salice (Cork), Anna Bortolan (Aberdeen).
Attendance is free of charge, but registration is required! To register, please, write to valeria.bizzari@libero.it, or oren.bader@med.uni-heidelberg.de.
For more information, visit https://cespa.uconn.edu/SelfOrgSummer2019#
December 6-9, 2019, Xiamen, China.
IEEE ALIFE 2019 brings together researchers working on the emerging areas of Artificial Life and Complex Adaptive Systems, aiming to understand and synthesize life-like systems and applying bio-inspired synthetic methods to other science/engineering disciplines, including Biology, Robotics, Social Sciences, among others.
Artificial Life is the study of the simulation and synthesis of living systems. In particular, this science of generalized living and life-like systems provides engineering with billions of years of design expertise to learn from and exploit through the example of the evolution of organic life on earth. Increased understanding of the massively successful design diversity, complexity, and adaptability of life is rapidly making inroads into all areas of engineering and the Sciences of the Artificial. Numerous applications of ideas from nature and their generalizations from life-as-we-know-it to life-as-it-could-be continually find their way into engineering and science.
Best Paper/Best Student Paper Awards will be sponsored by Wolfram Research, Inc.
Important dates
Paper Submissions: July 10, 2019
Notification to Authors: Sep. 1, 2019
Final Submission: Oct. 1, 2019
Early Registration: Oct. 1, 2019
Publications
Accepted papers after peer-review will be published in the IEEE SSCI conference proceedings. Submissions will be made via the main IEEE SSCI website.
Topics
We invite submissions of high-quality contributions on a wide variety of topics relevant to the wide research areas of Artificial Life. Some sample topics of interest include, but are not limited to, the following aspects of Artificial Life:
Organizing Committee
Hiroki Sayama – Binghamton University, USA (chair) – sayama@binghamton.edu
Chrystopher Nehaniv – University of Waterloo, Canada
Joseph Lizier – The University of Sydney, Australia
Stefano Nichele – Oslo Metropolitan University, Norway
Terry Bossomaier – Charles Sturt University, Australia
For more information, visit http://ssci2019.org/alife.html
This year’s Australasian Association of Philosophy (AAP) conference will be hosted by the University of Wollongong on July 7-11, 2019.
The conference is designed to give professional philosophers and philosophy postgraduate students the opportunity to present and discuss papers in all areas of philosophy. Each year it attracts around 300 philosophers worldwide.
AAP 2019 welcomes papers in all areas of philosophy. In addition to regular streams on topics such as Epistemology, Ethics, Philosophy of Mind, and Political Philosophy (among numerous others), we are organising the following Special Streams on more specific topics:
For more information, visit https://aap.org.au/conference2019
¿Te has preguntado qué es la mente? ¿Qué nos hace conscientes? y ¿Cuáles son las bases de nuestras decisiones? ¿Sabes si la Inteligencia Artificial puede replicar nuestras capacidades mentales? ¡Entonces ésta Escuela de Verano es para ti! Se llevará a cabo del 3 al 7 de junio en Ciudad Universitaria, UNAM.
El objetivo de la primera escuela de verano en ciencia cognitiva es difundir y promover el estudio interdisciplinar en ciencia cognitiva en la comunidad universitaria. Para ello durante una semana contaremos con diversas actividades dirigidas tanto a estudiantes, como a profesores e investigadores.
¡Apúntate y participa en esta experiencia única!
El Centro de Investigación de Ciencias Cognitivas (CINCCO) invita al 5º Coloquio Internacional de Ciencias Cognitivas, que tiene por objetivos:
Visita el sitio web del evento para conocer el programa y registrar tu asistencia.
¡Te esperamos!
4E Cognition Group 