Our group has immediate openings for a post-doctoral researcher and a research scientist for a new project on interactive scene recognition (from images and video). Our focus is on integrating algorithms for structure learning (e.g., Sum-Product Networks (SPNs), Stochastic Image Grammars (SIGs), Deep Learning (DL)), constrained by sources of domain knowledge (including those obtainable through interactive and active learning (AL) techniques), to automatically notify a human operator of interesting objects and events.
We have specific interest in using a constrained SPN approach, where constraints will reduce the hypothesis space of the structures that can be learned. These structures will be used with a separate knowledge base (e.g., of rules or cases) to derive inferences for the operator. Our application objective is to develop tools that can assist operators with answering situational awareness questions pertaining to a scene.
In addition to having experience with developing software prototypes and conducting focused evaluations of research hypotheses, ideal applicants should have some familiarity with computer vision and symbolic reasoning methods (e.g., Bayesian networks, case-based reasoning). However, we also welcome applicants with related experience/expertise.
Our group has immediate openings for a post-doctoral researcher, research scientist, and/or research programmer for projects pertaining to goal reasoning (GR), which is the study of agents that can autonomously select their own objectives. Our current projects on this topic pertain to the control of autonomous unmanned vehicles, both in simulation and hardware (robotics), either executing alone or in collaboration with other manned or unmanned vehicles. We also have interest in studying the use of GR techniques in interactive decision aids.
This is a particularly exciting research topic at this time because GR is a core capability of highly autonomous agents. Our group currently includes several employees, post-doctoral researchers, and contractors who are collaborating on GR-related projects involving autonomous underwater vehicles, collaborating teams of autonomous vehicles, robotic assistants to detached squads, and mixed human/unmanned simulated air combat teams.
Ideal applicants should have experience with conducting studies of focused research hypotheses, or with supporting such studies in software development roles. Familiarity with autonomous unmanned vehicles, simulations for studying them, goal reasoning and machine learning algorithms, and related AI or control methods would be welcome. However, we also welcome applicants with related experience/expertise.
If any of these openings interest you and you qualify, please contact David Aha to discuss them.