Published on January 27, 2009
The Good, The Bad, and the Ugly Experiences with Active Learning in Artificial Intelligence your name
The Module: Artificial Intelligence The study of algorithms that do intelligent things. your name
The Old Course • Evaluated through two tests, two small projects and one largish project • Very little practical experience with the algorithms and programming your name
The New Course One large semester long programming project broken into 5 parts, done by groups of three or less. your name
The Product An agent capable of playing a game of Chess your name
The Goal To win a tournament at the end of the course. The winning group receives 1 extra point on their final grade. your name
Development Plan • 5 smaller projects 1)Design of the board (rules, data model) 2)Algorithms for searching 3)Logic (define rules logical, redesign board) 4)Learning 5)Complete Agent, Presentation, Report your name
Special Rules • After the submittal of any one project, students can use the work of another group in their next projects (With proper documentation) • Students evaluate the work of other groups. This evaluation forms part of the groups grade for their project. The evaluation counts as part of the student's grade for participation. your name
Teaching Goals • More experience with programming, working in groups, reading and evaluating code • Learning the techniques of AI through hands on experience • Give the students a goal and motivation to learn and explore your name
Other notes • Each subproject designed to be more than what is needed for the next subproject. Students must decide what should go, what should stay. • Project definitions have a certain amount of ambiguity (to encourage students to gather software requirements) your name
The Good: • Some very motivated groups, reading and investigating methods to improve their agents. • Student evaluations more or less reflect the teacher's grades • Some changes of attitudes throughout the term • Better programming by the end of term. your name
The Bad • Low performance Only two of the projects were able to win against a random agent • Not much reuse of code students complained it was too hard to understand the code of others • Students unable to analogize from the work of others • Students unable to move ideas from theory to code • Much more time required by professor your name
More of The Bad • Unequal work by students within groups • Students lost sight of Artificial Intelligence aspects of course, focused more on Programming your name
The Ugly • Students expected pautas for the evaluation of other projects • No group was able to complete project 1 successfully. This delayed all further projects and caused much stress for the students and delayed their progress on the following projects. • Complaints of unclear/changing specifications for the projects • High expectations of students by the professor not matched by reality your name
The lessons • Set aside time for group meetings • Have a backup plan • Let the students make mistakes (but not too many) • Discuss group projects in class • Relate theory to code your name
The take home message Was it worth it? Yes your name
Ver el perfil profesional de Matthew Bardeen en LinkedIn. LinkedIn es la red de negocios más grande del mundo que ayuda a profesionales como Matthew ...
... vuelo de uno de los cuadricópteros armados en la Facultad de Ingeniería de la Universidad de Talca, por el profesor Matthew Bardeen, ...
26 Mayo 2015. Abrir espacios de diálogo y generar nuevas ideas y proyectos es el propósito del Consejo Universidad- Empresa (CUE), agrupación creada ...
Matthew Bardeen, Cesar Astudillo, June Verner; Experiencia. Assistant Professor Universidad de Talca. 2003 – actualidad (13 años) Lecturer, Researcher.
El profesor Matthew Bardeen nos invita a analizar y discutir sobre algoritmos genéticos y vida artificial.
Profesor Informante: Matthew Bardeen, Dphil Computer Science and Artificial Intelligence. Fecha de presentación de la Defensa de Memoria: 04/10/2012.
Matthew Bardeen. Professor at Universidad de Talca. Gina Morales Arnaiz. Ejecutiva Comercial- Pre evaluacion en Banco de Chile. Pablo Espinoza Zenteno.