Dar a conocer temáticas relacionadas con investigaciones del área de la informática, es el fin del ciclo de charlas dictadas por académicos del Instituto de Informática de la Facultad de Ciencias de la Ingeniería UACh.
“Queremos socializar las investigaciones que realizan los académicos del Instituto e investigadores colaboradores invitados”, explicó Eliana Scheihing, Directora del Instituto de Informática UACh.
Las charlas se realizarán todos los lunes de mayo y junio, entre las 15:50 y 17:20 en el Edificio 9.000 del Campus Miraflores UACh. Cabe destacar que estas presentaciones están abiertas a toda la comunidad universitaria.
Programa
Lunes 2 de mayo: Diego Sáez“Perspectivas de investigación y ejercicio profesional en data science”.
«Conceptos como data science y big data han cobrado gran relevancia en los últimos años. La demanda por profesionales en estas áreas tanto en la industria como la academia es constante. Pero ¿qué es un data scientist? ¿Cuáles son los conocimientos y habilidades necesarias para realizar este trabajo? En la actualidad aún no existe formación de pregrado para este trabajo, haciendo que profesionales de distintas formaciones (matemáticos, estadísticos e ingenieros de diversas áreas) deban reinventarse para abordar este nuevo ámbito de trabajo.
«En esta charla, Diego compartirá su experiencia de los últimos 8 años estudiando y trabajando en temas relacionados con data science en universidades y empresas internacionales tales como la Universidad Pompeu Fabra, Universidad de Cambridge y Yahoo Labs, comentando el presente y futuro de esta área».
Lunes 16 de mayo: Eliana Scheihing “Classifying discourse in a CSCL platform to evaluate correlations with teacher participation and progress”.
«In computer-supported learning, monitoring and engaging a group of learners is a complex task for teachers, especially when learners are working collaboratively: Are my students motivated? What kind of progress are they making? Should I intervene? Is my communication and the didactic design adapted to my students? Our hypothesis is that the analysis of natural language interactions between students, and between students and teachers, provide very valuable information and could be used to produce qualitative indicators to help teachers’ decisions. We develop an automatic approach in three steps (1) to explore the discursive functions of messages in a CSCL platform, (2) to classify the messages automatically and (3) to evaluate correlations between discursive attitudes and other variables linked to the learning activity. Results tend to show that some types of discourse are correlated with a notion of progress on the learning activities and the importance of emotive participation from the teacher».
Lunes 30 de mayo: Luis Álvarez “Designing questions to teach Python with a classroom response system”.
«To do an effective class using a classroom response system, two important questions appear: What questions to ask?, and How to ask them?, This paper tries to contribute with answers for these questions for an introductory course of Python language. For the first question, the Question Driven Instruction (QDI) methodology is introduced. In addition, for the second one, the questions are divided in content questions and process questions according to the content objectives and process objectives. Some examples questions are shown. The Classroom Response System (CRS) called ‘Mobile QTI’ is used. ‘Mobile QTI’ was designed for mobile devices (tablets and smartphones), supports eight types of question, and is under IMS Question & Test Interoperability Specification (QTI). Finally, an experiment is presented in order to validate the methodology. The experiment shows a relationship between Grade Point Average (GPA) and the use of the QDI and the CRS Mobile QTI».
Lunes 13 de junio: Cristobal Navarro “Adaptive multi-GPU Exchange Monte Carlo for the 3D Random Field Ising Model”.
«This work presents an adaptive multi-GPU Exchange Monte Carlo approach for the simulation of the 3D Random Field Ising Model (RFIM). The design follows a two level parallelization. The first level, spin-level parallelism, maps the parallel computation as optimal 3D thread-blocks that simulate blocks of spins in shared memory with minimal halo surface, assuming a constant block volume. The second level, replica-level parallelism, uses multi-GPU computation to handle the simulation of an ensemble of replicas. CUDA’s concurrent kernel execution feature is used in order to fill the occupancy of each GPU with many replicas, providing a performance boost that is more notorious at the smallest values of L. In addition to the two-level parallel design, the work proposes an adaptive multi-GPU approach that dynamically builds a proper temperature set free of exchange bottlenecks. The strategy is based on mid-point insertions at the temperature gaps where the exchange rate is most compromised. The extra work generated by the insertions is balanced across the GPUs independently of where the mid-point insertions were performed. Performance results show that spin-level performance is approximately two orders of magnitude faster than a single core CPU version and one order of magnitude faster than a parallel multi-core CPU version running on 16-cores. Multi-GPU performance is highly convenient under a weak scaling setting, reaching up to 99% efficiency as long as the number of GPUs and L increase together».