ARAyush Raina

Ayush Raina

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Ayush Raina has a total of 8 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Artificial Intelligence Impersonating a Human: The Impact of Design Facilitator Identity on Human Designers

    by Guanglu Zhang, Ayush Raina, Ethan Brownell, Jonathan Cagan

    Advances in artificial intelligence (AI) offer new opportunities for human–AI cooperation in engineering design. Human trust in AI is a crucial factor in ensuring an effective human–AI cooperation, and several approaches to enhance human trust in AI have been explored in prior studies. However, it remains an open question in engineering design whether human designers have more trust in an AI and achieve better joint performance when they are deceived into thinking they are working with another human designer. This research assesses the impact of design facilitator identity (“human” versus AI) on human designers through a human subjects study, where participants work with the same AI design facilitator and they can adopt their AI facilitator’s design anytime during the study. Half of the participants are told that they work with an AI, and the other half of the participants are told that they work with another human participant but in fact they work with the AI design facilitator. The results demonstrate that, for this study, human designers adopt their facilitator’s design less often on average when they are deceived about the identity of the AI design facilitator as another human designer. However, design facilitator identity does not have a significant impact on human designers’ average performance, perceived workload, and perceived competency and helpfulness of their design facilitator in the study. These results caution against deceiving human designers about the identity of an AI design facilitator in engineering design.


  • Learning to Design Without Prior Data: Discovering Generalizable Design Strategies Using Deep Learning and Tree Search

    by Ayush Raina, Jonathan Cagan, Christopher McComb

    Building an Artificial Intelligence (AI) agent that can design on its own has been a goal since the 1980s. Recently, deep learning has shown the ability to learn from large-scale data, enabling significant advances in data-driven design. However, learning over prior data limits us only to solve problems that have been solved before and biases data-driven learning toward existing solutions. The ultimate goal for a design agent is the ability to learn generalizable design behavior in a problem space without having seen it before. We introduce a self-learning agent framework in this work that achieves this goal. This framework integrates a deep policy network with a novel tree search algorithm, where the tree search explores the problem space, and the deep policy network leverages self-generated experience to guide the search further. This framework first demonstrates an ability to discover high-performing generative strategies without any prior data, and second, it illustrates a zero-shot generalization of generative strategies across various unseen boundary conditions. This work evaluates the effectiveness and versatility of the framework by solving multiple versions of two engineering design problems without retraining. Overall, this paper presents a methodology to self-learn high-performing and generalizable problem-solving behavior in an arbitrary problem space, circumventing the need for expert data, existing solutions, and problem-specific learning.


  • The Evolution and Impact of Human Confidence in Artificial Intelligence and in Themselves on AI-Assisted Decision-Making in Design

    by Leah Chong, Ayush Raina, Kosa Goucher-Lambert, Kenneth Kotovsky, Jonathan Cagan

    Decision-making assistance by artificial intelligence (AI) during design is only effective when human designers properly utilize the AI input. However, designers often misjudge the AI’s and/or their own ability, leading to erroneous reliance on AI and therefore bad designs occur. To avoid such outcomes, it is crucial to understand the evolution of designers’ confidence in both their AI teammate(s) and themselves during AI-assisted decision-making. Therefore, this work conducts a cognitive study to explore how to experience various and changing (without notice) AI performance levels and feedback affects these confidences and consequently the decisions to accept or reject AI suggestions. The results first reveal that designers’ confidence in an AI agent changes with poor, but not with good, AI performance in this work. Interestingly, designers’ self-confidence initially remains unaffected by AI accuracy; however, when the accuracy changes, self-confidence decreases regardless of the direction of the change. Moreover, this work finds that designers tend to infer flawed information from feedback, resulting in inappropriate levels of confidence in both the AI and themselves. Confidence in AI and self-confidence are also shown to affect designers’ probability of accepting AI input in opposite directions in this study. Finally, results that are uniquely applicable to design are identified by comparing the findings from this work to those from a similar study conducted with a non-design task. Overall, this work offers valuable insights that may enable the detection of designers’ dynamic confidence and their consequent misuse of AI input in the design.


  • Open Access

    Design Strategy Network: A deep hierarchical framework to represent generative design strategies in complex action spaces

    by Ayush Raina, Jonathan Cagan, Christopher McComb

    Generative design problems often encompass complex action spaces that may be divergent over time, contain state-dependent constraints, or involve hybrid (discrete and continuous) domains. To address those challenges, this work introduces Design Strategy Network (DSN), a data-driven deep hierarchical framework that can learn strategies over these arbitrary complex action spaces. The hierarchical architecture decomposes every action decision into first predicting a preferred spatial region in the design space and then outputting a probability distribution over a set of possible actions from that region. This framework comprises a convolutional encoder to work with image-based design state representations, a multi-layer perceptron to predict a spatial region, and a weight-sharing network to generate a probability distribution over unordered set-based inputs of feasible actions. Applied to a truss design study, the framework learns to predict the actions of human designers in the study, capturing their truss generation strategies in the process. Results show that DSNs significantly outperform non-hierarchical methods of policy representation, demonstrating their superiority in complex action space problems.


  • Open Access

    Goal-Directed Design Agents: Integrating Visual Imitation With One-Step Lookahead Optimization for Generative Design

    by Ayush Raina, Lucas Puentes, Jonathan Cagan, 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.