Nina Moorman
I’m Nina Moorman, a third-year Computer Science Ph.D. student at the CORE Robotics Lab at Georgia Tech, advised by professor Matthew Gombolay.
I graduated with my B.S. in Computer Science at Georgia Tech, specializing in Theory and Intelligence.
My research interests are in human-interactive robot learning and care robotics. Through my research, I develop algorithms that enable robots to learn in-situ, from non-expert user demonstrators.
See below for examples of my recent work!
Investigating the Impact of Experience on a User's Ability to Perform Hierarchical Abstraction
Prior work has yet to show that human users can provide sufficient demonstrations in novel domains without showing the demonstrators explicit teaching strategies for each domain. Our findings demonstrate for the first time that non-expert demonstrators can transfer knowledge from a series of training experiences to novel domains without the need for explicit instruction, such that they can provide necessary and sufficient demonstrations when programming robots to complete task and motion planning problems.
RSS 2023 (Best Paper Nominee)
Negative Result for Learning from Demonstration: Challenges for End-Users Teaching Robots with Task and Motion Planning Abstractions
Prior works have not examined whether non-roboticist endusers are capable of providing hierarchical demonstrations without explicit training from a roboticist showing how to teach each task. Our findings determine the necessary conditions to teach users through hierarchy and task abstractions, and the form of instructional information or feedback that is required to support users to learn to program robots effectively to solve novel tasks.
Lancon-learn: Learning with Language to Enable Generalization in Multi-Task Manipulation
We present LanCon-Learn, a novel attention-based approach to language-conditioned multi-task learning in manipulation domains to enable learning agents to reason about relationships between skills and task objectives through natural language and interaction.
RA-L 2022, presented at ICRA 2022
Impacts of Robot Learning on User Attitude and Behavior
We examine how different learning methods influence both in-person and remote participants' perceptions of the robot. We additionally compare the impact of these factors on the caregiver population, as compared to the general population.
Mind meld: Personalized Meta-Learning for Robot-Centric Imitation Learning
We present Mutual Information-driven Meta-learning from Demonstration (MIND MELD). MIND MELD meta-learns a mapping from suboptimal and heterogeneous human feedback to optimal labels, thereby improving the learning signal for robot-centric LfD. The key to our approach is learning an informative personalized em-bedding using mutual information maximization via variational inference. The embedding then informs a mapping from human provided labels to optimal labels.
Effects of Social Factors and Team Dynamics on Adoption of Collaborative Robot Autonomy
The attitudes of workers towards automation are influenced by a variety of complex and multi-faceted factors such as intention to use and perceived usefulness. In an analog manufacturing environment, we explore how these various factors influence an individual's willingness to work with a robot over a human co-worker in a collaborative Lego building task. We explore how this willingness is affected by the level of social rapport established between the individual and his or her human co-worker, the anthropomorphic qualities of the robot, and factors including trust, fluency and personality traits.
Athletic Mobile Manipulator System for Robotic Wheelchair Tennis
We propose the first open-source, autonomous robot for playing regulation wheelchair tennis. We demonstrate the performance of our full-stack system in executing ground strokes and evaluate each of the system's hardware and software components.
RA-L 2023, presented at IROS 2023
On the Bipartiteness Constant and Expansion of Cayley Graphs
Let G be a finite, undirected, d-regular graph and A(G) its normalized adjacency matrix. It is a classical fact that the smallest eigenvalue of A(G) is equal to -1, if and only if G is bipartite. Our main result provides a quantitative separation of this eigenvalue from -1 in the case of Cayley graphs, in terms of their expansion. We tighten the bounds on this eigenvalue if G is a non-bipartite Cayley graph, constructed using a group and a symmetric generating set of size d. We exhibit graphs for which this result is tight up to a factor depending on d.