Search NASA⌕ Search

SEARCH · Search NASA

Results for “decision graph”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

PySIDT: Subgraph Isomorphic Decision Trees for Molecular Property Prediction

Accurate molecular property prediction is important across all fields of chemistry. Deep neural networks (DNNs) have become increasingly popular due to their ability to train automatically, avoiding the incredibly tedious process of constructing and extending traditional property estimation schemes. However, DNNs require large amounts of training data, are challenging to interpret, require large amounts of memory to load even during inference, and have severe difficulties incorporating qualitative chemical knowledge, which are often desired for molecular property prediction tasks. Here, in this study, we present PySIDT (https://github.com/zadorlab/PySIDT), a software for training and running inference on Subgraph Isomorphic Decision Trees (SIDTs). SIDTs are graph-based decision trees made of nodes associated with molecular substructures. Inference is done by descending target molecular structures down the decision tree to nodes with matching subgraph isomorphic substructures and making predictions based on the final (most specific) nodes matched. SIDTs scale down well to dataset sizes much smaller than is feasible for DNNs. As trees of molecular substructures, SIDTs are inherently readable and easy to visualize, making them easy to analyze. They are also straightforward to extend and retrain, facilitate uncertainty estimation, and enable easy integration of expert knowledge. We demonstrate the SIDT approach discussing its application to a diverse range of molecular prediction tasks: rate coefficient estimation, diffusion coefficient estimation, thermochemistry estimation, transition state bond stretch prediction, p K a prediction, stability of molecular structures, stability of surface structures, and prediction of surface lateral interaction energetics. Additionally, we demonstrate the power of the SIDT algorithms in two direct learning curve vanilla comparisons with the popular DNN-based software Chemprop and the popular gradient boosted trees-based software XGBoost on enthalpy of formation and rate coefficient prediction tasks. In particular, in the enthalpy of formation case, vanilla PySIDT is able to outperform vanilla Chemprop and XGBoost across the full range of training/validation set sizes out to 11,560 data points.

Johnson, Matthew Sean [Sandia National Laboratorie↗
DOE OSTI2022

graphenv: a Python library for reinforcement learning on graph search spaces

Many important and challenging problems in combinatorial optimization (CO) can be expressed as graph search problems, in which graph vertices represent full or partial solutions and edges represent decisions that connect them. Graph structure not only introduces strong relational inductive biases for learning (Battaglia et al., 2018) - in this context, by providing a way to explicitly model the value of transitioning (along edges) between one search state (vertex) and the next - but lends itself to problems both with and without clearly defined algebraic structure. For example, classic CO problems on graphs such as the Traveling Salesman Problem (TSP) can be expressed as either pure graph search or integer programs. Other problems, however, such as molecular optimization, do no have concise algebraic formulations and yet are readily implemented as a graph search (V. et al., 2022; Zhou et al., 2019). Such "model-free" problems constitute a large fraction of modern reinforcement learning (RL) research owing to the fact that it is often much easier to write a forward simulation that expresses all of the state transitions and rewards, than to write down the precise mathematical expression of the full optimization problem. In the case of molecular optimization, for example, one can use domain knowledge alongside existing software libraries to model the effect of adding a single bond or atom to an existing but incomplete molecule, and let the RL algorithm build a model of how good a given decision is by "experiencing" the simulated environment many times through. In contrast, a model-based mathematical formulation that fully expresses all the chemical and physical constraints is intractable. In recent years, RL has emerged as an effective paradigm for optimizing searches over graphs and led to state-of-the-art heuristics for games like Go and chess, as well as for classical CO problems such as the TSP. This combination of graph search and RL, while powerful, requires non-trivial software to execute, especially when combining advanced state representations such as Graph Neural Networks (GNN) with scalable RL algorithms.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Multi-Agent Graph-Attention Deep Reinforcement Learning for Post-Contingency Grid Emergency Voltage Control

Grid emergency voltage control (GEVC) is paramount in electric power systems to improve voltage stability and prevent cascading outages and blackouts in case of contingencies. While most deep reinforcement learning (DRL)-based paradigms perform single agents in a static environment, real-world agents for GEVC are expected to cooperate in a dynamically shifting grid. Moreover, due to high uncertainties from combinatory natures of various contingencies and load consumption, along with the complexity of dynamic grid operation, the data efficiency and control performance of the existing DRL-based methods are challenged. To address these limitations, we propose a multi-agent graph-attention (GATT)-based DRL algorithm for GEVC in multi-area power systems. Here, we develop graph convolutional network (GCN)-based agents for feature representation of the graph-structured voltages to improve the decision accuracy in a data-efficient manner. Furthermore, a cutting-edge attention mechanism concentrates on effective information sharing among multiple agents, synergizing different-sized subnetworks in the grid for cooperative learning. We address several key challenges in the existing DRL-based GEVC approaches, including low scalability and poor stability against high uncertainties. Test results in the IEEE benchmark system verify the advantages of the proposed method over several recent multi-agent DRL-based algorithms.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Comparing Interaction Graphs on Cascading Outages under Different Loading Conditions

Interaction graphs on cascading outages of power systems provide valuable insights into how cascading outages evolve and propagate, and which components and links are critical to the propagation of cascading outages, enabling further development of mitigation strategies to support decision-making. However, the sensitivity of the interaction graph’s topology to the system’s loading condition has not been studied sufficiently. This paper compared interaction graphs under various loading conditions on the Northeastern Power Coordinating Council 140-bus system, and discovers the strong relationships between the graph topology, the cascade size distribution, and the load condition. Accordingly, three representative interaction graphs are constructed and illustrated.

Guo, Zhenping↗
DOE OSTI2023

Near-Optimal Distributed Linear-Quadratic Regulator for Networked Systems

This paper studies the trade-off between the degree of decentralization and the performance of a distributed controller in a linear-quadratic control setting. We study a system of interconnected agents over a graph and a distributed controller, called k-distributed control, which lets the agents make control decisions based on the state information within distance k on the underlying graph. This controller can tune its degree of decentralization using the parameter k and thus allows a characterization of the relationship between decentralization and performance. We show that under mild assumptions, including stabilizability, detectability, and a subexponentially growing graph condition, the performance difference between k-distributed control and centralized optimal control becomes exponentially small in k. Finally, this result reveals that distributed control can achieve near-optimal performance with a moderate degree of decentralization, and thus it is an effective controller architecture for large-scale networked systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

ACES-GNN: can graph neural network learn to explain activity cliffs?

Graph Neural Networks (GNNs) have revolutionized molecular property prediction by leveraging graph-based representations, yet their opaque decision-making processes hinder broader adoption in drug discovery. This study introduces the Activity-Cliff-Explanation-Supervised GNN (ACES-GNN) framework, designed to simultaneously improve predictive accuracy and interpretability by integrating explanation supervision for activity cliffs (ACs) into GNN training. ACs, defined by structurally similar molecules with significant potency differences, pose challenges for traditional models due to their reliance on shared structural features. By aligning model attributions with chemist-friendly interpretations, the ACES-GNN framework bridges the gap between prediction and explanation. Validated across 30 pharmacological targets, ACES-GNN consistently enhances both predictive accuracy and attribution quality for ACs compared to unsupervised GNNs. Our results demonstrate a positive correlation between improved predictions and accurate explanations, offering a robust and adaptable framework to better understand and interpret ACs. This work underscores the potential of explanation-guided learning to advance interpretable artificial intelligence in molecular modeling and drug discovery.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Quantitative Risk Management - Risk Informed Decision Making on High-Risk Waste Remediation Projects - 19022

For every $1 billion invested in projects, $97 million is wasted due to poor project performance. That means 9.7% of money spent on projects is wasted due to poor risk and uncertainty management. High-risk projects, similar to those in waste remediation, have the most to lose. A solid quantitative risk management program is one of the keys to increasing the likelihood that a project will succeed. A quantitative risk program solves two major problems that challenge the effective management of risk and uncertainty: 1. Risks arise at all levels of an organization; however, not all organizational levels have an effective way to identify and resolve them. This can lead to risks that, even if identified, are not reported. 2. Decision-makers are often forced to operate solely on intuition when implementing mitigation strategies, not fully understanding the effects that certain decisions have on the program's cost and schedule. A quantitative risk program uses a risk register and builds it into the project/program integrated schedule to assess the impact of these risks, typically at the 80% confidence contingency level for specific milestones and/or project completion. Implementing a quantitative risk program drives a forward-thinking, proactive mentality to problem solving into all levels of the organization. A healthy risk program for regular reporting and communication ensures that all members of an organization are thinking about risks and using their communication channels to resolve them. It gives a method of communication and incentive for all levels in the organization and helps to ensure that risks will be reported and resolved. Pacific Northwest National Laboratory (PNNL) has developed tools and methods that provide a level of realism for analyzing different scenarios and estimating contingency that does not exist in the standard industry risk tool kit. These tools and methods provide decision-makers with actionable, quantitatively based information to guide their decision making. The first tool is a probabilistic branching risk model for mitigation strategies. Current quantitative risk analysis tools in the industry simulate two cases: fully mitigated and un-mitigated. In the fully mitigated case, the user assumes that every planned mitigation activity is successful. In the un-mitigated case, the user assumes that none are successful. The truth lies between these two extremes. The PNNL team has developed a probabilistic risk model for mitigation scenarios that includes a probability that a mitigation activity will succeed and a probability that a mitigation action will be funded. These probabilities are elicited from subject matter experts and ensure that the contingency calculations at the specified confidence are approaching reality. The second tool developed at PNNL is the Risk Box, which is a tool that integrates with different scenarios in the risk register. For example, say the user is faced with two different and mutually exclusive mitigation actions to reduce a risk. The Risk Box methodology performs the contingency calculations and then displays them on a time-phased graph that shows how each scenario reduces the risk contingency needed on the project/program. This graph then provides quantitative information to show decision-makers which mitigation action provides the greatest reduction (cost and schedule) in risk contingency for the program. Implementing these and other tools within a structured risk program increases the likelihood that a project will come in on schedule and budget, ensures that all personnel are proactively thinking about problems and resolutions, creates communication channels for all members to submit and resolve potential problems, produces a prioritized list of risk rankings to focus leadership on issues that may cause the biggest problems, and implements processes that drive updates to the risk program as the risk profile changes. PNNL has been developing these methods and using them on projects for over 20 years. They have been successfully applied on waste management projects, high-tech research and development projects, and construction projects, and they continue to be used on a number of U.S. Department of Energy projects today. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2025

InterGraph-CPS: A Graph-Theoretic Approach to Characterize Cross-Domain Cyber-Physical Interdependencies and Uncertainties in Electric Grid Systems for Improved Decision-Making in Operation and Response

Critical infrastructure systems such as the electric grid are increasingly cyber-physical; yet, despite the cyber-physical characteristics of critical infrastructure systems, the physical process system and communication/control network system are traditionally analyzed in siloes. As these systems become more cyber-physical, it is crucial that models and methods are available to assess the cyber physical system (CPS) interdependencies, characteristics, and event propagation for improved planning, operation, and response. Thus, we proposed an integrated structural and temporal CPS interdependency analysis framework, InterGraph-CPS, that provides insight into the CPS function during normal operation as well as disturbances. This integrated structural and temporal interdependency framework is uniquely designed for assessing CPSs by account for the challenges of analyzing cyber and physical data streams together due to data availability, data type, and time scale differences. By leveraging both structural (e.g., graph analysis) and temporal (e.g., data analytics) techniques, different CPS behaviors and configurations can be accounted for.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Improving Trustworthiness of Data-Driven Power Grid Contingency Analysis With Bayesian Residual Graph Neural Networks

The evolving energy landscape requires novel tools to efficiently perform contingency analysis and reliability assessment of power grids, potentially in real-time. The high computational cost of traditional power flow solvers limits their applicability in practice. Machine learning (ML) surrogates such as deep neural networks (NNs) accelerate power flow solvers computations, enabling high-order contingency analysis and real-time decision-making by learning highly nonlinear functions and integrating grid topology via graph architectures. However, (graph) NNs lack predictive power away from training data and do not provide predictive confidence estimates. Here, we present a Bayesian residual graph NN that integrates knowledge from low-fidelity data via residual training and embeds granular quantification of uncertainties, improving trustworthiness critical for high-consequence decision-making. Applying Bayesian concepts to NNs is challenging due to the high-dimensionality of both the parameter space, complicating derivation of a meaningful prior, and the output space in large grid systems, requiring enhanced techniques to assess the predicted high-dimensional uncertainties. Our contributions include: (1) Deriving a prior for fully connected and graph NNs that leverages low-fidelity data to guide mean predictions and appropriately control prior predictive uncertainty. (2) Integrating this prior within an ensembling with anchoring scheme for efficient approximate posterior inference. (3) Deriving enhanced metrics to assess accuracy of both the mean and uncertainty predictions in high dimensions, appropriately accounting for correlations propagated through graph layers. The resulting Bayesian residual graph NN is tested on a contingency analysis task for 14-bus and 118-bus grids.

24 - POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2019

Scalable Causal Graph Learning through a Deep Neural Network

Learning the causal graph in a complex system is crucial for knowledge discovery and decision making, yet it remains a challenging problem because of the unknown nonlinear interaction among system components. Most of the existing methods either rely on predefined kernel or data distribution, or they focus simply on the causality between a single target and the remaining system. This work presents a deep neural network for scalable causal graph learning (SCGL) through low-rank approximation. The SCGL model can explore nonlinearity on both temporal and intervariable relationships without any predefined kernel or distribution assumptions. Through low-rank approximation, the noise influence is reduced, and better accuracy and high scalability are achieved. Experiments using synthetic and real-world datasets show that our SCGL algorithm outperforms existing state-of-the-art methods for causal graph learning.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Efficient graph representation framework for chemical molecule similarity tasks

Graph data has emerged in numerous scientific domains and machine learning techniques have been widely used for analysis and learning of diverse data for prediction and decision. Machine learning techniques can readily address complex problems by leveraging their structural information. But graphs cannot be directly used for existing machine learning algorithms unless encoded as vectors. The problem of efficient representation of graphs is a substantial challenge in graph machine learning. In this paper, we propose a novel two-stage framework for the representation of chemical molecule graphs based on the strengths of Graph Isomorphism Networks (GINs) and Siamese autoencoders. In the first stage, the GIN model is constructed and trained using the structural information of chemical molecule graphs. Node attributes, edge attributes, and edge indices are used as input data, while graph attributes are used as labels. The GIN model effectively captures the structural characteristics of graphs and can accurately predict graph attributes, i.e., molecular properties. It also generates Graph Embeddings, represented as vectors that encode the structural information of graphs. In the second stage, Graph Embedding vectors are further optimized for downstream similarity tasks while preserving the graph structural information. The Siamese autoencoder is constructed and trained, which reduces the dimensionality of the Graph Embedding vectors, while maximizing the preservation of structural information in the original high-dimensional vectors. The resulting low-dimensional Graph Embeddings can be effectively utilized for tasks such as approximate nearest neighbor search. The experimental results demonstrate the effectiveness of our proposed framework in accurately predicting graph similarity.

Ma, Jiaji↗
DOE OSTI2025

Model-based Hierarchical Reinforcement Learning for Improved Physical Security Design: A Prototype

Prior work in FY24 developed an adversarial AI agent aid in path analysis of physical protection systems. This agent, trained using a model-based reinforcement learning algorithm, was able to successfully learn the most vulnerable path in facilities. It was able to extend the current state of practice for physical protection design by exhibiting dynamic behavior based on current environmental conditions. Whereas PathTrace largely performs a static, graph-based analysis, the AI agent was able to make decisions based on relative position in the facility, current conditions (was the adversarial agnet discovered?), and proximity to secondary targets. The agent demonstrated some novel capabilities, but had limitations that need to be resolved before it can be used for production purposes. For example, the adversarial agent generalizes poorly and takes a relatively long time to train. Nonetheless, there is still considerable promise for developing the adversarial agent further in order to explore even richer, more dynamic behaviors (e.g., adversary motivations, environmental debris, and more). This work considers a complementary idea; development of a planning agent. The planning agent is envisioned as an auto-complete-like tool that can help accelerate security system design by human experts. The agent would respect existing barriers and sensors placed by a human expert while offering cost-effective suggestions (i.e., implicitly balancing effectiveness with cost) to improve the design. The goal is for this agent to be part of an expert’s toolbox, not to totally upend the current state-of-practice, or to displace human experts. The ultimate goal would be concurrent training of both the adversarial and planning agent together, to learn entirely through self-play. This would represent an entirely new way of performing system deign. We selected a hierarchical, model-based reinforcement learning algorithm to serve as the planning agent. This is an extension of concepts used in the prior FY24 adversarial agent work. There, we had a single agent acting an environment. Here, we have two different sub-agents (policies), working together, to form a complete agent. There is a manager policy, which can select abstract goals on slower time scales, and a worker, which performs primitive actions to reach goals selected by the manager. It is worth noting that this class of algorithm is challenging to work with. From our understanding, our work is one of the first successful uses of model-based reinforcement learning (MBRL) in nuclear energy1 , and likely the first hierarchical model-based reinforcement learning application in nuclear energy. Further, this work is one of the first known attempts to apply AI to perform a design tasks in nuclear energy. Consequently, there were significant implementation challenges and the bulk of the work was focused on successful implementation and algorithm design. The results presented here are very low technology readiness level as a consequence of the lack of related literature, but still represent a significant step forward in the pursuit of applied AI for design.

42 ENGINEERING↗
DOE OSTI2023

End-to-End Pipeline for Trigger Detection on Hit and Track Graphs

There has been a surge of interest in applying deep learning in particle and nuclear physics to replace labor-intensive offline data analysis with automated online machine learning tasks. This paper details a novel AI-enabled triggering solution for physics experiments in Relativistic Heavy Ion Collider and future Electron-Ion Collider. The triggering system consists of a comprehensive end-to-end pipeline based on Graph Neural Networks that classifies trigger events versus background events, makes online decisions to retain signal data, and enables efficient data acquisition. Here, the triggering system first starts with the coordinates of pixel hits lit up by passing particles in the detector, applies three stages of event processing (hits clustering, track reconstruction, and trigger detection), and labels all processed events with the binary tag of trigger versus background events. By switching among different objective functions, we train the Graph Neural Networks in the pipeline to solve multiple tasks: the edge-level track reconstruction problem, the edge-level track adjacency matrix prediction, and the graph-level trigger detection problem. We propose a novel method to treat the events as track-graphs instead of hit-graphs. This method focuses on intertrack relations and is driven by underlying physics processing. As a result, it attains a solid performance (around 72% accuracy) for trigger detection and outperforms the baseline method using hit-graphs by 2% higher accuracy.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

CPES-QSM: A Quantitative Method Towards the Secure Operation of Cyber-Physical Energy Systems

Power systems are evolving into cyber-physical energy systems (CPES) mainly due to the integration of modern communication and Internet-of-Things (IoT) devices. CPES security evaluation is challenging since the physical and cyber layers are often not considered holistically. Existing literature focuses on only optimizing the operation of either the physical or cyber layer while ignoring the interactions between them. This paper proposes a metric, the Cyber-Physical Energy System Quantitative Security Metric (CPES-QSM), that quantifies the interaction between the cyber and physical layers across three domains: electrical, cyber-risk, and network topology. A method for incorporating the proposed cyber-metric into operational decisions is also proposed by formulating a cyber-constrained AC optimal power flow (C-ACOPF) that considers the status of all the CPES layers. The C-ACOPF considers the vulnerabilities of physical and cyber networks by incorporating factors such as voltage stability, contingencies, graph-theory, and IoT cyber risks, while using a multi-criteria decision-making technique. We note that simulation studies are conducted using standard IEEE test systems to evaluate the effectiveness of the proposed metric and the C-ACOPF formulation.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

System Engineers and Decisions: It?s All about Knowledge

In order to guarantee that a system meets adequate levels of reliability and availability, system performances are continuously monitored and analyzed thanks to the technological advancements driving the Industry 4.0 revolution. An Industry 4.0 approach is typically based on advanced statistical, big data mining, machine learning, and internet-of-things methods designed to detect anomalies in the behavior of system, detect the most likely failure modes, and provide indications to system engineers on when maintenance activities should be performed before system performance are deemed unacceptable (which can be generated by diagnostic and prognostic methods). However, these analyses, which are designed to automatize and increase the efficacy of the system maintenance program, require large amount of data which can come in various forms: numeric, textual, images, sounds etc. Such data constitutes the historic knowledge benchmark to track system performances and support system engineer decisions. Here we claim that data is not sufficient to support this kind of analyses when applied to systems characterized by complex architectures and behaviors. Robust system engineer decisions require the ability to understand the system operational context that lies behind the observed data elements. In this respect, system models are in fact necessary to “put data in context” and capture relationships between data elements. Industry 4.0 methods require in fact contextual knowledge as a basis upon which hypotheses can be generated and assumptions tested. In our view, for complex systems, model-based system engineering (MBSE) models can afford this contextual knowledge, as they are typically used to describe systems architecture and dynamic behaviors. System knowledge is here intended as the blending of collected data and system architecture which takes the form of a “knowledge graph”. A knowledge graph is a database which consists of a large set of nodes (in our case an entity can be either a data or an MBSE element) which are linked to each other. The types of nodes and links follow a pre-defined topology, sometimes also refers as an ontology, that is designed to fit the actual decisions that needs to be performed. We show here how a knowledge graph can be defined to support system engineer maintenance decisions and how the same graph can be built based on system MBSE models and pre-processed data from numeric (through anomaly detections and diagnostic methods) and textual elements (through technical language processing TLP).

97 - MATHEMATICS AND COMPUTING↗
DOE OSTI2024

GraMeR: Gra ph Me ta R einforcement learning for multi-objective influence maximization

Influence maximization (IM) is a combinatorial problem of identifying a subset of seed nodes in a network (graph), which when activated, provide a maximal spread of influence in the network for a given diffusion model and a budget for seed set size. IM has numerous applications such as viral marketing, epidemic control, sensor placement and other network-related tasks. However, its practical uses are limited due to the computational complexity of current algorithms. Recently, deep reinforcement learning has been leveraged to solve IM in order to ease the computational burden. However, there are serious limitations in current approaches, including narrow IM formulation that only consider influence via spread and ignore self-activation, low scalability to large graphs, and lack of generalizability across graph families leading to a large running time for every test network. In this work, we address these limitations through a unique approach that involves: (1) Formulating a generic IM problem as a Markov decision process that handles both intrinsic and influence activations; (2)incorporating generalizability via meta-learning across graph families. There are previous works that combine deep reinforcement learning with graph neural network, but this work solves a more realistic IM problem and incorporates generalizability across graphs via meta reinforcement learning. Extensive experiments are carried out in various standard networks to validate performance of the proposed Graph Meta Reinforcement learning (GraMeR) framework. Finally, the results indicate that GraMeR is multiple orders faster and generic than conventional approaches when applied on small to medium scale graphs.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Contradictory Ambiguous Revocable Assertion Tracker (CARAT) Encoding

How data is encoded in a knowledge graph directly influences what can be done with that data. A common problem with many encodings is that they have difficulty representing ambiguity and evolution inherent in many real-world data sets. The data encoding represented in this paper (called CARAT) is a graph-level description of our attempt to capture data that is contradictory, ambiguous and evolves over time (including deleting information). The data encoding relies on tracking assertions about subjects rather than directly tracking states. This encoding decision resolves many issues our team had experienced using other data encodings but produces a a larger graph. This is a preliminary presentation of our experience with CARAT.

Cottam, Joseph A. [BATTELLE (PACIFIC NW LAB)]↗
DOE OSTI2026

Cross-Domain Reasoning for Neuromorphic Model Design

Designing performant neuromorphic models requires reasoning across neuroscience, neuromorphic computing, and machine learning, making it a natural target for cross-domain hypothesis generation. Our primary contribution is a multi-corpus knowledge graph spanning all three domains, which we show substantially increases cross-domain retrieval novelty over single-corpus baselines. We additionally introduce NeuKReAct, an agentic reasoning framework that iteratively retrieves from this graph and synthesizes design hypotheses via a step-by-step blackboard architecture, enabling structured compartmentalization of design decisions. Lastly, we introduce an execution head that translates hypotheses into structured design documents and runnable code. We evaluate novelty using a combinatorial creativity metric that measures cross-domain retrieval distance across the citation graph. Our results confirm that corpus breadth is the dominant driver of novelty. Moreover, we highlight a concrete instance of the novelty-utility tradeoff within NeuKReAct, underscoring a need for joint creativity evaluation, balancing both novelty and utility.

Ramavarapu, Vikram [ORNL] (ORCID:0009000188757213)↗