University of Birmingham > Talks@bham > Artificial Intelligence and Natural Computation seminars > Relaxed AI Planning with Time and the TechLauncher Program: A Talk in Two Parts

Relaxed AI Planning with Time and the TechLauncher Program: A Talk in Two Parts

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Hector Basevi.

Host: Prof Jeremy Wyatt

Speaker’s website: http://users.cecs.anu.edu.au/~charlesg/

Abstract: The talk shall proceed in two parts, treating first some new research results in AI planning, and then a new teaching program for running group projects at university. The research part is about a new heuristic for deterministic planning with durative actions and time constraints. That research is due to be presented at the International Conference on Automated Planning and Scheduling in late June. It is joint research with Dr Patrik Haslum, and with PhD candidate Tony Allard as first author. The ANU TechLauncher teaching program shall then be presented. This is our innovative approach to group project work, where the process is driven by meaningful community, industry, and government engagements.

Research Topic Detail:

We consider planning problems with time windows, in which availability of discrete resources is time constrained. We develop a novel heuristic that addresses specifically the difficulty of coordinating actions within time windows. The heuristic is based on solving a temporally relaxed problem and measuring the magnitude by which the relaxed solution violates the time window constraints. Applied in a state-space search planner, the heuristic reduces the number of dead-ends encountered during search, and improves planner coverage.

Teaching Topic Detail:

TechLauncher is a program devised by the Research School of Science at the Australian National University, and is convened by the speaker. An interdisciplinary endeavour, TechLauncher combines many project-based courses, enabling mature undergraduate and post-graduate students to develop and showcase research and professional skills. The majority of the 320 students enrolled in TechLauncher courses are from our graduate and undergraduate computing and engineering programs. We also have participation from our School of Music, the Crawford School of Public Policy, the Research School of Earth Sciences, and others. Industry engagement is a critical aspect of TechLauncher, with more than half our groups having industry and/or government clients. Here, Project IP is licenced to the university for assessment purposes. Individual and group outputs are evaluated by student peers, industry mentors, project sponsors, and academic examiners. TechLauncher also supports student-led innovation, with a number of companies having started in TechLauncher. We advocate this model as a means for research and teaching institutions to improve student (employment) outcomes and supercharge industry and government engagement.

About the Speaker:

Dr Charles Gretton is a Senior Lecturer at the Australian National University. He convenes the TechLauncher program at the Research School of Computer Science at the Australian National University. Working closely with a number of collaborators, highlights of Dr Gretton’s recent research contributions include: (i) computing proxies for cost allocations in transportation games, (ii) compositional upper bounding concepts for Bounded Model Checking and AI Planning, with associated mechanisation of results in HOL4 , and (iii) applications oriented work developing an adaptive large neighbourhood search heuristic for computing economical and visually attractive routes. In addition to contributing to the research literature Dr Gretton co-founded HIVERY , a company backed by Data61 and The Coca-Cola Founders platform. HIVERY provides retail-AI solutions in the USA , Japan, Australia and NewZealand.

This talk is part of the Artificial Intelligence and Natural Computation seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

Talks@bham, University of Birmingham. Contact Us | Help and Documentation | Privacy and Publicity.
talks@bham is based on talks.cam from the University of Cambridge.