SEARCH · Search NASA

Results for “Reasoning AI”

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.

SourceResearch connectionFrom yearThrough year

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

At least 145 records · Page 8

All sources NASA NTRS DOE OSTI

NASA NTRS1986

The NASA automation and robotics technology program

The development and objectives of the NASA automation and robotics technology program are reviewed. The objectives of the program are to utilize AI and robotics to increase the probability of mission success; decrease the cost of ground control; and increase the capability and flexibility of space operations. There is a need for real-time computational capability; an effective man-machine interface; and techniques to validate automated systems. Current programs in the areas of sensing and perception, task planning and reasoning, control execution, operator interface, and system architecture and integration are described. Programs aimed at demonstrating the capabilities of telerobotics and system autonomy are discussed.

Holcomb, Lee B.↗

DOE OSTI2024

Assurance of Reasoning Enabled Systems (ARES)

ARES was in part motivated by the determination of President’s Council of Advisors on Science and Technology (PCAST) on May 13th, 2023 that published a set of inquiries: In an era in which convincing images, audio, and text can be generated with ease on a massive scale, how can we ensure reliable access to verifiable, trustworthy information? How can we be certain that a particular piece of media is genuinely from the claimed source? What technologies, policies, and infrastructure can be developed to detect and counter AI-generated disinformation? In an effort to automatically analyze and patch/optimize code the work in this report describes various neural Machine Learning (ML) analysis engine implementations to assist in situations where source code is deficient or completely lacking to decompile (lift) binary code to ’C’. The goal is to gradually reduce human intervention. To this end, two Large Language Model (LLM) variants (Code LLama 2, LLama 3.1 and Starcoder1, Starcoder 2) where finetuned with ’before/after’ code pairs on the OpenBLAS library. LLama trained on the lowering process, Starcoder trained on the lifting process with National Security Agency’s (NSA) open-source Ghidra decompiler assist. The inferencing test results indicate correctness for only very short sequences for Starcoder 2. Moving forward, the experiments conclude with a set of recommendations of required resources and technologies

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1990

Applications of artificial intelligence to mission planning

The scheduling problem facing NASA-Marshall mission planning is extremely difficult for several reasons. The most critical factor is the computational complexity involved in developing a schedule. The size of the search space is large along some dimensions and infinite along others. It is because of this and other difficulties that many of the conventional operation research techniques are not feasible or inadequate to solve the problems by themselves. Therefore, the purpose is to examine various artificial intelligence (AI) techniques to assist conventional techniques or to replace them. The specific tasks performed were as follows: (1) to identify mission planning applications for object oriented and rule based programming; (2) to investigate interfacing AI dedicated hardware (Lisp machines) to VAX hardware; (3) to demonstrate how Lisp may be called from within FORTRAN programs; (4) to investigate and report on programming techniques used in some commercial AI shells, such as Knowledge Engineering Environment (KEE); and (5) to study and report on algorithmic methods to reduce complexity as related to AI techniques.

Ford, Donnie R.↗

NASA NTRS1987

Raising the AIQ of the Space Station

Expert systems and robotics technologies are to be significantly advanced during the Space Station program. Artificial intelligence systems (AI) on the Station will include 'scars', which will permit upgrading the AI capabilities as the Station evolves to autonomy. NASA-Ames is managing the development of the AI systems through a series of demonstrations, the first, controlling a single subsystem, to be performed in 1988. The capabilities being integrated into the first demonstration are described; however, machine learning and goal-driven natural language understanding will not reach a prototype stage until the mid-1990s. Steps which will be taken to endow the computer systems with the ability to move from heuristic reasoning to factual knowledge, i.e., learning from experience, are explored. It is noted that the development of Space Station expert systems depends on the development of experts in Station operations, which will not happen until the Station has been used extensively by crew members.

Lum, Henry↗

DOE OSTI2026

From natural language to control signals: a conceptual framework for semantic channel finding in complex experimental infrastructure

Modern experimental platforms such as particle accelerators, fusion devices, telescopes, and industrial process control systems expose tens to hundreds of thousands of control and diagnostic channels, accumulated over decades of hardware evolution. Operators and AI systems alike depend on informal expert knowledge, inconsistent naming conventions, and scattered documentation to locate the signals required for monitoring, troubleshooting, and automated control, creating a persistent bottleneck for reliability, scalability, and emerging language-model-driven interfaces. We formalize semantic channel finding, the task of mapping natural-language intent to concrete control-system signals, as a general problem in complex experimental infrastructure, and introduce a four-paradigm conceptual framework to guide architecture selection based on facility-specific data regimes. The paradigms span (i) direct in-context lookup over small, curated channel dictionaries, (ii) constrained hierarchical navigation through structured trees, (iii) interactive agent exploration using iterative reasoning and tool-based database queries, and (iv) ontology-grounded semantic search that decouples channel meaning from facility-specific naming conventions. We demonstrate the practical feasibility of each paradigm through proof-of-concept implementations at four operational facilities spanning two orders of magnitude in scale: from compact free-electron lasers to large synchrotron light sources, operating under diverse control-system architectures ranging from clean hierarchical naming schemes to legacy environments with decades of heterogeneous conventions. Where evaluated against expert-curated operational queries, these instantiations achieve 90%–97% accuracy, validating the framework’s applicability across real-world deployment scenarios. To accelerate adoption across the broader scientific and industrial control-system community, we release open-source, plug-and-play implementations of all three interactive paradigms-direct lookup, hierarchical navigation, and middle-layer exploration-within the Osprey framework, together with tools for channel database generation, interactive testing, and minimal-configuration deployment. This work establishes semantic channel finding as a foundational capability for human-centric and agentic AI interfaces at large-scale facilities, providing both a systematic framework for architecture design and practical resources to enable adoption without building custom infrastructure from scratch.

channel finding↗

DOE OSTI2026

TETA Autoresearch [SWR-26-089]

TETA Autoresearch is a template repository based on github.com/karpathy/autoresearch for AI-assisted research science in the TETA group in the Center for Integrated Mobility Sciences (CIMS) center at the National Laboratory of the Rockies. This software is a template for running autonomous research experiments that iteratively improve an ML model for a single optimization objective. Two execution modes share one harness: LLM mode - an agent (e.g. Claude Code) edits a scaffold train.py one change at a time, tagging each experiment, logging reasoning, and pushing results. Defined by program.md. Optimizer mode - an Optuna-backed driver (TPE / CMA-ES / Random) iterates over a domain-defined search space. Defined by optimizers/. RouteE (vehicle energy prediction) is the reference domain under domains/routee/. Adding a new domain is mechanical - see EXTENDING.md.

Reinicke, Nicholas [National Laboratory of the Roc↗

NASA NTRS1987

A framework for qualitative reasoning about solid objects

Predicting the behavior of a qualitatively described system of solid objects requires a combination of geometrical, temporal, and physical reasoning. Methods based upon formulating and solving differential equations are not adequate for robust prediction, since the behavior of a system over extended time may be much simpler than its behavior over local time. A first-order logic, in which one can state simple physical problems and derive their solution deductively, without recourse to solving the differential equations, is discussed. This logic is substantially more expressive and powerful than any previous AI representational system in this domain.

Davis, E.↗

DOE OSTI2023

Combining data and theory for derivable scientific discovery with AI-Descartes

Abstract Scientists aim to discover meaningful formulae that accurately describe experimental data. Mathematical models of natural phenomena can be manually created from domain knowledge and fitted to data, or, in contrast, created automatically from large datasets with machine-learning algorithms. The problem of incorporating prior knowledge expressed as constraints on the functional form of a learned model has been studied before, while finding models that are consistent with prior knowledge expressed via general logical axioms is an open problem. We develop a method to enable principled derivations of models of natural phenomena from axiomatic knowledge and experimental data by combining logical reasoning with symbolic regression. We demonstrate these concepts for Kepler’s third law of planetary motion, Einstein’s relativistic time-dilation law, and Langmuir’s theory of adsorption. We show we can discover governing laws from few data points when logical reasoning is used to distinguish between candidate formulae having similar error on the data.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS1985

A knowledge based software engineering environment testbed

The Carnegie Group Incorporated and Boeing Computer Services Company are developing a testbed which will provide a framework for integrating conventional software engineering tools with Artifical Intelligence (AI) tools to promote automation and productivity. The emphasis is on the transfer of AI technology to the software development process. Experiments relate to AI issues such as scaling up, inference, and knowledge representation. In its first year, the project has created a model of software development by representing software activities; developed a module representation formalism to specify the behavior and structure of software objects; integrated the model with the formalism to identify shared representation and inheritance mechanisms; demonstrated object programming by writing procedures and applying them to software objects; used data-directed and goal-directed reasoning to, respectively, infer the cause of bugs and evaluate the appropriateness of a configuration; and demonstrated knowledge-based graphics. Future plans include introduction of knowledge-based systems for rapid prototyping or rescheduling; natural language interfaces; blackboard architecture; and distributed processing

Gill, C.↗

NASA NTRS1993

A review of European applications of artificial intelligence to space

The purpose is to describe the applications of Artificial Intelligence (AI) to the European Space program that are being developed or have been developed. The results of a study sponsored by the Artificial Intelligence Research and Development program of NASA's Office of Advanced Concepts and Technology (OACT) are described. The report is divided into two sections. The first consists of site reports, which are descriptions of the AI applications seen at each place visited. The second section consists of two summaries which synthesize the information in the site reports by organizing this information in two different ways. The first organizes the material in terms of the type of application, e.g., data analysis, planning and scheduling, and procedure management. The second organizes the material in terms of the component technologies of Artificial Intelligence which the applications used, e.g., knowledge based systems, model based reasoning, procedural reasoning, etc.

Drummond, Mark↗

DOE OSTI2026

A multimodal large language model for materials science

Understanding and predicting the properties of inorganic materials is crucial for accelerating advancements in materials science and driving applications in energy, electronics and beyond. Integrating material structure data with language-based information through multimodal large language models (LLMs) offers great potential to support these efforts by enhancing human–artificial intelligence interaction. However, a key challenge lies in integrating atomic structures at full resolution into LLMs. In this work, we introduce MatterChat, a versatile structure-aware multimodal LLM that unifies material structural data and textual inputs into a single cohesive model. MatterChat uses a bridging module to effectively align a pretrained universal machine learning interatomic potential with a pretrained LLM, reducing training costs and enhancing flexibility. Our results demonstrate that MatterChat greatly improves performance in material property prediction and human–artificial intelligence interaction, surpassing general-purpose LLMs such as GPT-4. We also demonstrate its usefulness in applications such as more advanced scientific reasoning and step-by-step material synthesis.

Tang, Yingheng [Lawrence Berkeley National Laborat↗

DOE OSTI2022

MetaText: Compositional Generalization in Deep Language Models

Compositional generalization, the ability to infer and reason using novel combinations of previously encountered entities and structures, is a trait of great utility across a variety of deep learning tasks; it implies a certain ability to reason using consistent and therefore interpretable rules, to understand the constituent parts of its input at multiple levels of granularity, to be robust to spurious differences between semantically equivalent inputs, and more. We seek to understand the degree to which existing natural language models achieve or fall short of compositional generalization across a variety of tasks, whether there are distinct types of compositional generalization in practice, and to identify avenues of intervention through which we can improve models’ compositional generalization ability. After exploring these questions using an array of models, tasks, and potential interventions we see that large, pretrained language models have encountered sufficient training data to account for a variety of fine-grained compositional behavior but still struggle to reason at the level of phrases or larger language structures. Non-intrusive, data-based interventions in the form of augmenting individual sequences with compressed versions of themselves or deriving new examples from induced grammars prove insufficient to encourage greater levels of compositional reasoning, indicating that future work might benefit most from focusing on changes to a model’s inductive bias at the architectural or loss level, or by integrating compositionality-boosting data interventions into the large-scale pretraining process itself.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1990

Artificial intelligence and design: Opportunities, research problems and directions

The issues of industrial productivity and economic competitiveness are of major significance in the U.S. at present. By advancing the science of design, and by creating a broad computer-based methodology for automating the design of artifacts and of industrial processes, we can attain dramatic improvements in productivity. It is our thesis that developments in computer science, especially in Artificial Intelligence (AI) and in related areas of advanced computing, provide us with a unique opportunity to push beyond the present level of computer aided automation technology and to attain substantial advances in the understanding and mechanization of design processes. To attain these goals, we need to build on top of the present state of AI, and to accelerate research and development in areas that are especially relevant to design problems of realistic complexity. We propose an approach to the special challenges in this area, which combines 'core work' in AI with the development of systems for handling significant design tasks. We discuss the general nature of design problems, the scientific issues involved in studying them with the help of AI approaches, and the methodological/technical issues that one must face in developing AI systems for handling advanced design tasks. Looking at basic work in AI from the perspective of design automation, we identify a number of research problems that need special attention. These include finding solution methods for handling multiple interacting goals, formation problems, problem decompositions, and redesign problems; choosing representations for design problems with emphasis on the concept of a design record; and developing approaches for the acquisition and structuring of domain knowledge with emphasis on finding useful approximations to domain theories. Progress in handling these research problems will have major impact both on our understanding of design processes and their automation, and also on several fundamental questions that are of intrinsic concern to AI. We present examples of current AI work on specific design tasks, and discuss new directions of research, both as extensions of current work and in the context of new design tasks where domain knowledge is either intractable or incomplete. The domains discussed include Digital Circuit Design, Mechanical Design of Rotational Transmissions, Design of Computer Architectures, Marine Design, Aircraft Design, and Design of Chemical Processes and Materials. Work in these domains is significant on technical grounds, and it is also important for economic and policy reasons.

Amarel, Saul↗

NASA NTRS1994

Diagnosing anomalies of spacecraft for space maintenance and servicing

Very often servicing of satellites is necessary to replace components which are responsible for anomalous behavior of satellite operations due to adverse interactions with the natural space environment. A major difficulty with this diagnosis is that those responsible for diagnosing these anomalies do not have the tools to assess the role of the space environment causing the anomaly. To address this issue, we have under development a new rule-based, expert system for diagnosing spacecraft anomalies. The knowledge base consists of over two-hundred rules and provides links to historical and environmental databases. Environmental causes considered are bulk charging, single event upsets (SEU), surface charging, and total radiation dose. The system's driver translates forward chaining rules into a backward chaining sequence, prompting the user for information pertinent to the causes considered. When the user selects the novice mode, the system automatically gives detailed explanations and descriptions of terms and reasoning as the session progresses, in a sense teaching the user. As such it is an effective tutoring tool. The use of heuristics frees the user from searching through large amounts of irrelevant information and allows the user to input partial information (varying degrees of confidence in an answer) or 'unknown' to any question. The system is available on-line and uses C Language Integrated Production System (CLIPS), an expert shell developed by the NASA Johnson Space Center AI Laboratory in Houston.

Lauriente, Michael↗

NASA NTRS1993

An Intelligent computer-aided tutoring system for diagnosing anomalies of spacecraft in operation

A new rule-based, expert system for diagnosing spacecraft anomalies is under development. The knowledge base consists of over two-hundred (200) rules and provides links to historical and environmental databases. Environmental causes considered are bulk charging, single event upsets (SEU), surface charging, and total radiation dose. The system's driver translates forward chaining rules into a backward chaining sequence, prompting the user for information pertinent to the causes considered. When the user selects the novice mode, the system automatically gives detailed explanations and descriptions of terms and reasoning as the session progresses, in a sense teaching the user. As such it is an effective tutoring tool. The use of heuristics frees the user from searching through large amounts of irrelevant information and allows the user to input partial information (varying degrees of confidence in an answer) or 'unknown' to any question. The system is available on-line and uses C Language Integrated Production System (CLIPS), an expert shell developed by the NASA Johnson Space Center AI Laboratory in Houston.

Rolincik, Mark↗

DOE OSTI2023

BUILD: Binary Understanding and Integration Logic for Dependencies (Final Report)

Increasingly diverse mission needs, the emergence of AI and cloud, and increasing hardware diversity are driving HPC software to be more complex. Modern codes are built from hundreds of small, complex components, and much of the software development process involves integrating these components rather than developing new components from scratch. The goal of the BUILD project was to ease the task of software integration for developers across LLNL’s programs. The project focused on (1) modeling software compatibility, (2) modeling ABI compatibility with binary analysis, (3) developing solver techniques to reason about compatibility, and (4) developing AI/ML models to fill gaps in our understanding of software compatibility. The project has developed several key technologies that help developers—by accelerating development workflows, removing the need for rebuilds, and enabling faster, automatic, and less error-prone code sharing. These technologies are used in LLNL codes and will be ready for the new El Capitan Exascale system in Livermore Computing. Many of these technologies have been hardened and integrated with LLNL’s Spack package manager, and they are already in use by production code teams. Results of BUILD have laid the groundwork for future advances in software integration—the ML and binary modification techniques developed in this project still need to be operationalized but have great potential to further speed up software integration.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

An Agenda for Multimodal Foundation Models for Earth Observation

Archives of remote sensing (RS) data are increasing swiftly as new sensing modalities with enhanced spatiotemporal resolution become operational. While promising new breakthroughs, the sheer volume of RS archives stretches the limits of human analysts and existing AI tools, as most models are: i) limited to single data modalities; ii) task-specific; iii) heavily reliant on labeled data. The emerging Foundation Models (FMs) have the potential to address these limitations. Trained on vast unlabeled datasets through self-supervised learning, FMs enable generic feature extraction that facilitate specialization to a wide variety of downstream tasks. This paper describes a vision towards an FM for multimodal Earth Observation data (FM4EO), discussing key building blocks and open challenges. We put particular emphasis on multimodal reasoning, a topic underexplored in EO. Our ultimate goal is a practical path toward FM4EO with capacity to unlock breakthroughs in few-shot learning scenarios, multimodal geographic knowledge integration, synthesis, and hypothesis generation.

Ambrozio Dias, Philipe↗

NASA NTRS1985

AIS Investigation of Agricultural Monocultures

Airborne Imaging Spectrometer (AIS) data were acquired over an agricultural area in eastern San Joaquin County, California in July, 1984. Cover type information was subsequently collected for all fields along this flight line. The lack of detailed ground data on individual fields, however, limited AIS data analysis to a qualitative comparison of the spectral reflectance curves for a total of nine cover types. Based on this analysis, it appears that cover types with a positive slope in the 1550 to 1700 nm region have a higher spectral response in the 1200 to 1300 nm region compared to those cover types with a negative slope in the 1550 to 1700 nm region. Within cover type, spectral variability was also found to be greater than that between cover types. Given the lack of additional field data, the reason for these differences is a matter of speculation.

Wood, B. L.↗