skip to main content


Title: Extracting generalizable skills from a single plan execution using abstraction-critical state detection
Robotic task planning is computationally challenging. To reduce planning cost and support life-long operation, we must leverage prior planning experience. To this end, we address the problem of extracting reusable and generalizable abstract skills from successful plan executions. In previous work, we introduced a supporting framework, allowing us, theoretically, to extract an abstract skill from a single execution and later automatically adapt it and reuse it in new domains. We also proved that, given a library of such skills, we can significantly reduce the planning effort for new problems. Nevertheless, until now, abstract-skill extraction could only be performed manually. In this paper, we finally close the automation loop and explain how abstract skills can be practically and automatically extracted. We start by analyzing the desired qualities of an abstract skill and formulate skill extraction as an optimization problem. We then develop two extraction algorithms, based on the novel concept of abstraction-critical state detection. As we show experimentally, the approach is independent of any planning domain.  more » « less
Award ID(s):
1830549
NSF-PAR ID:
10483224
Author(s) / Creator(s):
; ;
Publisher / Repository:
IEEE
Date Published:
Journal Name:
2023 IEEE International Conference on Robotics and Automation (ICRA)
Page Range / eLocation ID:
5772 to 5778
Format(s):
Medium: X
Location:
London, United Kingdom
Sponsoring Org:
National Science Foundation
More Like this
  1. Billard, A ; Asfour, T ; Khatib, O. (Ed.)
    Task planning is the problem of finding a discrete sequence of actions to achieve a goal. Unfortunately, task planning in robotic domains is computationally challenging. To address this, in our prior work, we explained how knowledge from a successful task solution can be cached for later use, as an “abstract skill.” Such a skill is represented as a trace of states (“road map”) in an abstract space and can be matched with new tasks on-demand. This paper explains how one can use a library of abstract skills, derived from past planning experience, to reduce the computational cost of solving new task planning problems. As we explain, matching a skill to a task allows us to decompose it into independent sub-tasks, which can be quickly solved in parallel. This can be done automatically and dynamically during planning. We begin by formulating this problem of “planning with skills” as a constraint satisfaction problem. We then provide a hierarchical solution algorithm, which integrates with any standard task planner. Finally, we experimentally demonstrate the computational benefits of the approach for reach-avoid tasks. 
    more » « less
  2. A model that maps the requisite skills, or knowledge components, to the contents of an online course is necessary to implement many adaptive learning technologies. However, developing a skill model and tagging courseware contents with individual skills can be expensive and error prone. We propose a technology to automatically identify latent skills from instructional text on existing online courseware called Smart (Skill Model mining with Automated detection of Resemblance among Texts). Smart is capable of mining, labeling, and mapping skills without using an existing skill model or student learning (aka response) data. The goal of our proposed approach is to mine latent skills from assessment items included in existing courseware, provide discovered skills with human-friendly labels, and map didactic paragraph texts with skills. This way, mapping between assessment items and paragraph texts is formed. In doing so, automated skill models produced by Smart will reduce the workload of courseware developers while enabling adaptive online content at the launch of the course. In our evaluation study, we applied Smart to two existing authentic online courses. We then compared machine-generated skill models and human-crafted skill models in terms of the accuracy of predicting students’ learning. We also evaluated the similarity between machine-generated and human-crafted skill models. The results show that student models based on Smart-generated skill models were equally predictive of students’ learning as those based on human-crafted skill models— as validated on two OLI (Open Learning Initiative) courses. Also, Smart can generate skill models that are highly similar to human-crafted models as evidenced by the normalized mutual information (NMI) values. 
    more » « less
  3. Background: Project management and other professional skill training is often lacking in graduate student education, typically as a result of limited resources, lack of faculty buy-in, and narrow focus on thesis research. To address this need and with support from NSF, we are developing the Graduates for Advancing Professional Skills (GAPS) program at Iowa State University. To aid the initial development of this program, we conducted a literature review to understand the current context of the development and implementation of professional skills in higher education curricula, with specific interest in STEM fields. Purpose: The purpose of our study was to identify best practices related to implementing professional development skills into an academic curriculum. The goal was to utilize this information in the development, planning, implementation, and assessment of our GAPS program. Design: We engaged in a systematic literature review. We focused on the curricular and pedagogical approaches to implementing these skills, results of the initiatives, and methodologies used to assess their effectiveness. Results: Our literature review uncovered the “messiness” of teaching and learning of skills such as project management. There is often not one approach or definition of project management – it may change based on scope of project and context. Successful implementation requires adaptability, mentorship, problem solving, creativity, and communication. Additionally, project management has been referred to as a “threshold concept” and requires a certain level of intuition that cannot necessarily be gained through traditional classroom education. Conclusions: There appears to be an agreement on the importance of implementing project management skills at the postsecondary level. Our work illustrates the difficulty associated with undertaking this endeavor and provides guidance on approaches that can make these initiatives more beneficial. Although this literature was conducted to aid in the planning for our specific project, the synthesis of the extant works can inform other faculty and industry leaders who are interested in teaching and applying project management techniques in their courses or companies. 
    more » « less
  4. Metacognition is the understanding of your own knowledge including what knowledge you do not have and what knowledge you do have. This includes knowledge of strategies and regulation of one’s own cognition. Studying metacognition is important because higher-order thinking is commonly used, and problem-solving skills are positively correlated with metacognition. A positive previous disposition to metacognition can improve problem-solving skills. Metacognition is a key skill in design and manufacturing, as teams of engineers must solve complex problems. Moreover, metacognition increases individual and team performance and can lead to more original ideas. This study discusses the assessment of metacognitive skills in engineering students by having the students participate in hands-on and virtual reality activities related to design and manufacturing. The study is guided by two research questions: (1) do the proposed activities affect students’ metacognition in terms of monitoring, awareness, planning, self-checking, or strategy selection, and (2) are there other components of metacognition that are affected by the design and manufacturing activities? The hypothesis is that the participation in the proposed activities will improve problem-solving skills and metacognitive awareness of the engineering students. A total of 34 undergraduate students participated in the study. Of these, 32 were male and 2 were female students. All students stated that they were interested in pursuing a career in engineering. The students were divided into two groups with the first group being the initial pilot run of the data. In this first group there were 24 students, in the second group there were 10 students. The groups’ demographics were nearly identical to each other. Analysis of the collected data indicated that problem-solving skills contribute to metacognitive skills and may develop first in students before larger metacognitive constructs of awareness, monitoring, planning, self-checking, and strategy selection. Based on this, we recommend that the problem-solving skills and expertise in solving engineering problems should be developed in students before other skills emerge or can be measured. While we are sure that the students who participated in our study have awareness as well as the other metacognitive skills in reading, writing, science, and math, they are still developing in relation to engineering problems. 
    more » « less
  5. Robots deployed in many real-world settings need to be able to acquire new skills and solve new tasks over time. Prior works on planning with skills often make assumptions on the structure of skills and tasks, such as subgoal skills, shared skill implementations, or task-specific plan skeletons, which limit adaptation to new skills and tasks. By contrast, we propose doing task planning by jointly searching in the space of parameterized skills using high-level skill effect models learned in simulation. We use an iterative training procedure to efficiently generate relevant data to train such models. Our approach allows flexible skill parameterizations and task specifications to facilitate lifelong learning in general-purpose domains. Experiments demonstrate the ability of our planner to integrate new skills in a lifelong manner, finding new task strategies with lower costs in both train and test tasks. We additionally show that our method can transfer to the real world without further fine-tuning. 
    more » « less