Lucas Puentes

Lucas Puentes

Auto-generated profile, based on publicly available information

Lucas Puentes

Department of Mechanical Engineering, The Pennsylvania State University, University Park, PA 16802

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
9 June 2021 in Journal of Mechanical Design
Verified
Journal Article
Open access
PDF file
ARAyush Raina
Ayush Raina
LPLucas Puentes
Lucas Puentes
JCJonathan Cagan
Jonathan Cagan
CMChristopher McComb
Christopher McComb
Engineering design problems often involve large state and action spaces along with highly sparse rewards. Since an exhaustive search of those spaces is not feasible, humans utilize relevant domain knowledge to condense the search space. Deep learning agents (DLAgents) were previously introduced to use visual imitation learning to model design domain knowledge. This note builds on DLAgents and integrates them with one-step lookahead search to develop goal-directed agents capable of enhancing learned strategies for sequentially generating designs. Goal-directed DLAgents can employ human strategies learned from data along with optimizing an objective function. The visual imitation network from DLAgents is composed of a convolutional encoder–decoder network, acting as a rough planning step that is agnostic to feedback. Meanwhile, the lookahead search identifies the fine-tuned design action guided by an objective. These design agents are trained on an unconstrained truss design problem modeled as a sequential, action-based configuration design problem. The agents are then evaluated on two versions of the problem: the original version used for training and an unseen constrained version with an obstructed construction space. The goal-directed agents outperform the human designers used to train the network as well as the previous feedback-agnostic versions of the agent in both scenarios. This illustrates a design agent framework that can efficiently use feedback to not only enhance learned design strategies but also adapt to unseen design problems.

21 October 2019 in Journal of Computing and Information Science in Engineering
Verified
Journal Article
Closed access
PDF file
LPLucas Puentes
Lucas Puentes
JCJonathan Cagan
Jonathan Cagan
CMChristopher McComb
Christopher McComb

14 October 2020 in Journal of Computing and Information Science in Engineering
Verified
Journal Article
Closed access
PDF file
LPLucas Puentes
Lucas Puentes
JCJonathan Cagan
Jonathan Cagan
CMChristopher McComb
Christopher McComb

16 July 2021
Verified
Preprint
Open access
GWGlen Williams
Glen Williams
LPLucas Puentes
Lucas Puentes
JNJacob Nelson
Jacob Nelson
JMJessica Menold
Jessica Menold
CTConrad Tucker
Conrad Tucker
CMChristopher McComb
Christopher McComb
*Online data repositories provide designers and engineers with a convenient source of data. Over time, the wealth and type of readily-available data within online repositories has greatly expanded. This data increase permits new uses for machine learning approaches which rely on large, high-dimensional datasets. However, a comparison of the efficacies of attribute-based data, which

16 July 2021
Verified
Preprint
Open access
LPLucas Puentes
Lucas Puentes
JCJonathan Cagan
Jonathan Cagan
CMChristopher McComb
Christopher McComb
Through experience, designers develop guiding principles, or heuristics, to aid decision-making in familiar design domains. Generalized versions of common design heuristics have been identified across multiple domains and applied by novices to design problems. Previous work leveraged a sample of these common heuristics to assist in an agent-based design process, which typically