Search NASA⌕ Search

SEARCH · Search NASA

Results for “hypothesis generation”

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 37 records · Page 2

Synthetic Scientific Image Generation with VAE, GAN, and Diffusion Model Architectures

Generative AI (genAI) has emerged as a powerful tool for synthesizing diverse and complex image data, offering new possibilities for scientific imaging applications. This review presents a comprehensive comparative analysis of leading generative architectures, ranging from Variational Autoencoders (VAEs) to Generative Adversarial Networks (GANs) on through to Diffusion Models, in the context of scientific image synthesis. We examine each model's foundational principles, recent architectural advancements, and practical trade-offs. Our evaluation, conducted on domain-specific datasets including microCT scans of rocks and composite fibers, as well as high-resolution images of plant roots, integrates both quantitative metrics (SSIM, LPIPS, FID, CLIPScore) and expert-driven qualitative assessments. Results show that GANs, particularly StyleGAN, produce images with high perceptual quality and structural coherence. Diffusion-based models for inpainting and image variation, such as DALL-E 2, delivered high realism and semantic alignment but generally struggled in balancing visual fidelity with scientific accuracy. Importantly, our findings reveal limitations of standard quantitative metrics in capturing scientific relevance, underscoring the need for domain-expert validation. We conclude by discussing key challenges such as model interpretability, computational cost, and verification protocols, and discuss future directions where generative AI can drive innovation in data augmentation, simulation, and hypothesis generation in scientific research.

Generative Adversarial Networks↗

Idea2Data: Toward a New Paradigm for Drug Discovery

Increasing the success rate and throughput of drug discovery will require efficiency improvements throughout the process that is currently used in the pharmaceutical community, including the crucial step of identifying hit compounds to act as drivers for subsequent optimization. Hit identification can be carried out through large compound collection screening and frequently involves the generation and testing of many hypotheses based on available knowledge. In practice, hypothesis generation can involve the selection of promising chemical structures from compound collections using predictive models built from previous screening/ assay results. Available physical collections, typically used during hit identification, are of the order of 10 6 compounds but represent only a small fraction of the small molecule drug-like chemical space. In an effort to survey a larger portion of chemical space and eliminate inefficiencies during hit identification, we introduce a new process, termed Idea2Data (I2D) that tightly integrates computational and experimental components of the drug discovery process. I2D provides the ability to connect a vast virtual collection of compounds readily synthesizable on automated synthesis systems with computational predictive models for the identification of promising structures. This new paradigm enables researchers to process billions of virtual molecules and select structures that can be prepared on automated systems and made available for biological testing, allowing for timely hypothesis testing and follow-up. Since its introduction, I2D has positively impacted several portfolio efforts through identification of new chemical scaffolds and functionalization of existing scaffolds. In this Innovations paper, we describe the I2D process and present an application for the discovery of new ULK inhibitors.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

SCIMON: Scientific Inspiration Machines Optimized for Novelty

We explore and enhance the ability of neu- ral language models to generate novel scien- tific directions grounded in literature. Work on literature-based hypothesis generation has traditionally focused on binary link prediction— severely limiting the expressivity of hypothe- ses. This line of work also does not focus on optimizing novelty. We take a dramatic depar- ture with a novel setting in which models use as input background contexts (e.g., problems, experimental settings, goals), and output natu- ral language ideas grounded in literature. We present SCIMON, a modeling framework that uses retrieval of “inspirations” from past scien- tific papers, and explicitly optimizes for novelty by iteratively comparing to prior papers and up- dating idea suggestions until sufficient novelty is achieved. Comprehensive evaluations reveal that GPT-4 tends to generate ideas with over- all low technical depth and novelty, while our methods partially mitigate this issue. Our work represents a first step toward evaluating and developing language models that generate new ideas derived from the scientific literature.

Ji, Heng↗

Shape matching utilizing indexed hypotheses generation and testing

An indexing mechanism is developed as part of an overall scheme called SMITH (shape matching utilizing indexed hypothesis generation and testing) for two-dimensional model-based object recognition. The approach is based on a dynamic programming implementation of attributed string matching, is computationally efficient, and works effectively for both nonoccluded and occluded shapes. Another advantage of this technique is that models may be inserted or deleted with relatively little cost.

Mehrotra, Rajiv↗

Developing an Interactive Data Visualization Tool to Assess the Impact of Decision Support on Clinical Operations

Due to mandates from recent legislation, clinical decision support (CDS) software is being adopted by radiology practices across the country. This software provides imaging study decision support for referring providers at the point of order entry. CDS systems produce a large volume of data, providing opportunities for research and quality improvement. In order to better visualize and analyze trends in this data, an interactive data visualization dashboard was created using a commercially available data visualization platform. Following the integration of a commercially available clinical decision support product into the electronic health record, a dashboard was created using a commercially available data visualization platform (Tableau, Seattle, WA). Data generated by the CDS were exported from the data warehouse, where they were stored, into the platform. This allowed for real-time visualization of the data generated by the decision support software. The creation of the dashboard allowed the output from the CDS platform to be more easily analyzed and facilitated hypothesis generation. Integrating data visualization tools into clinical decision support tools allows for easier data analysis and can streamline research and quality improvement efforts.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Mitochondria Dysfunction Central in Driving Health Risks Associated with Spaceflight

Background: Determining the biological impact of spaceflight through novel approaches is essential to reduce the health risks to astronauts for long-term space missions. The current established health risks due to spaceflight are only reflecting known symptomatic and physiologic responses and do not reflect early onset of other potential diseases. There are many unknown variables which still need to be identified to fully understand the health impacts due to the environmental factors in space. Materials and Methods: One method to uncover potential novel biological mechanisms responsible for health risks in astronauts is by utilizing NASA’s GeneLab platform (genelab.nasa.gov). GeneLab is public repository that hosts multiple omics datasets generated from space biology experiments that include experiments flown in space, simulated cosmic radiation experiments, and simulated microgravity experiments. This presentation will provide an example of analysis and novel hypothesis generation that is being produced with GeneLab datasets. A comprehensive multi-omics approach was implemented correlating transcriptomics, proteomics, metabolomics, and methylation analysis. Results: We found that cells have stronger overall biological response than the tissues to spaceflight, with mitochondrial activity and innate immunity pathways being heavily impacted. NASA Twin Study results are consistent with a specific alteration in mitochondrial ATP production. Our results indicate that the space environment can directly induce mitochondrial damage, with mitochondrial dysfunctions being a cause for chronic inflammation and both being involved in the development of metabolic disorders that cause changes in lipid metabolism. We also found biological changes occurring during spaceflight with cell cycle, circadian rhythm and olfactory activity pathways can also influence and be influenced by alterations on mitochondrial activity

Afshin Beheshti↗

An Algorithm for Constructing and Searching Spaces of Alternative Hypotheses

In this paper we develop techniques for automated hypothesis generation over data sets that maycontain contradictions. To do so, we make use of the equivalence between two formulations: those offirst order predicate logic with prefix modal quantifiers under the finite model hypothesis and those ofMixed Integer Linear Programming (MILP) problems. The equivalence between integer programming and logical satisfiability has been known since Karp'sseminal work in NP-completeness. Other authors have made use of this equivalence to explore efficientmethods of solving the satisfiability problem in the propositional calculus for specific problem types.The work presented here differentiates itself from previous work in that we do not assume that alllogical assertions are true without doubt. Instead we look for alternative hypotheses about the validityof the claims by identifying alternative optimal solutions to a corresponding MILP. We use a collectionof slack variables in the derived linear constraints to indicate the presence of contradictory data orassumptions. The objective is to minimize contradictions between data and assertions represented bythe presence of non-zero slack in the set of linear constraints.

99 GENERAL AND MISCELLANEOUS↗

Aging matrix visualizes complexity of battery aging across hundreds of cycling protocols

To reliably deploy lithium-ion batteries, a fundamental understanding of cycling aging behavior is critical. Battery aging consists of complex and highly coupled phenomena, making it challenging to develop a holistic interpretation. In this work, we generate a diverse battery cycling dataset with a broad range of degradation trajectories, consisting of 359 high energy density commercial Li(Ni,Co,Al)O 2 /graphite + SiO x cylindrical 21 700 cells cycled across 207 unique cycling protocols. We consolidate aging via 16 mechanistic state-of-health (SOH) metrics, including cell-level performance metrics, electrode-specific capacities/state-of-charges (SOCs), and aging trajectory metrics. We develop a framework using interpretable machine learning and explainable features to generate an aging matrix that visually deconvolutes the complex battery degradation behavior. This generalizable data-driven mechanistic framework simplifies the complex interplay between cycling conditions, degradation modes, and SOH, acting as a hypothesis-generation tool to aid battery users in identifying key degradation regimes for further study and experimentation.

25 ENERGY STORAGE↗

Reverse engineering systems models of regulation: discovery, prediction and mechanisms

Biological systems can now be understood in comprehensive and quantitative detail using systems biology approaches. Putative genome-scale models can be built rapidly based upon biological inventories and strategic system-wide molecular measurements. Furthermore, current models combine statistical associations, causative abstractions, and known molecular mechanisms to explain and predict quantitative and complex phenotypes. This top-down ‘reverse engineering’ approach generates useful organism-scale models despite noise and incompleteness in data and knowledge. Here we review and discuss the reverse engineering of biological systems using top-down data-driven approaches, in order to improve discovery, hypothesis generation, and the inference of biological properties.

59 BASIC BIOLOGICAL SCIENCES↗

Hypothesis Learning in Automated Experiment: Application to Combinatorial Materials Libraries

Machine learning is rapidly becoming an integral part of experimental physical discovery via automated and high-throughput synthesis, and active experiments in scattering and electron/probe microscopy. This, in turn, necessitates the development of active learning methods capable of exploring relevant parameter spaces with the smallest number of steps. In this work, an active learning approach based on conavigation of the hypothesis and experimental spaces is introduced. This is realized by combining the structured Gaussian processes containing probabilistic models of the possible system's behaviors (hypotheses) with reinforcement learning policy refinement (discovery). This approach closely resembles classical human-driven physical discovery, when several alternative hypotheses realized via models with adjustable parameters are tested during an experiment. This approach is demonstrated for exploring concentration-induced phase transitions in combinatorial libraries of Sm-doped BiFeO 3 using piezoresponse force microscopy, but it is straightforward to extend it to higher-dimensional parameter spaces and more complex physical problems once the experimental workflow and hypothesis generation are available.

36 MATERIALS SCIENCE↗

LENS: Learning Enabled Network Synthesis

RTRC and UMD have developed novel machine learning based methods under the ARPA-E DIFFERENTIATE program for rapid acceleration of hypothesis generation in complex architecture design spaces involving both discrete choices of component inclusion and interconnection and continuous parametric decisions. The project named Learning Enabled Network Synthesis (LENS) further demonstrated the developed methods on challenging electrical power converter design problems by identifying the most suitable circuit topologies and simultaneously selecting the most appropriate components to achieve optimized design of power converter with improved performances. We demonstrated that LENS could enable exploration of very large design space of circuit topologies and components by addressing the limitations of conventional design process in non-linear, high switching speed, multi-dimensional power converter design and optimization. The key innovation developed in LENS is the seamless integration of statistical learning and logical reasoning techniques and building on the individual strengths of these techniques for rapid hypothesis discovery. The main component of LENS comprises of: 1) Graph Reasoning Engine (GRE) to enforce composition rules that rapidly reject all discrete architectures that are composed incorrectly and generates an adaptive database of feasible designs which can be used by ML modules, 2) Graph Generative Learning module which is a deep neural network based generative model for graph architectures which can enable design space exploration beyond the dataset generated by the GRE, 3) Graph Reduced Order Model (ROM) for graph domains for accelerating computation of output metrics, and 4) Active learning and Rule Discovery module for sample efficient learning and extracting logical rules from the learned ML models which will be integrated in the GRE to enhance the filtering effectiveness. LENS approach can be applied to any design domains where designs can be represented as multi-attribute graphs. The LENS team integrated the various technical innovations listed above into an optimization pipeline and exercised the optimization pipeline on the converter design problem. The LENS project demonstrated that the developed AI/ML technologies can be used to generate novel converter circuits >45x faster than experts on chosen use-cases. This can enable faster design space exploration and identification of new designs which are not considered by experts due to the increasing design space complexity. This has significant potential impact on the public and energy needs of the country. It is currently estimated that 30% of all electrical powers generated passes through power converters. The future estimate is that 80% of all power generated would be passing through converters. LENS fills a critical gap in this space since by accelerating the design process the designers would be able to generate more efficient converters which can lead to significant energy savings for the country.

42 ENGINEERING↗

Ice Crystal Icing Physics Study using a NACA 0012 Airfoil at the National Research Council of Canada's Research Altitude Test Facility

This paper presents results from a study of the fundamental physics of ice-crystal ice accretion using a NACA 0012 airfoil at the National Research Council of Canada (NRC) Research Altitude Test Facility in August 2017. These tests were a continuation of work which began in 2010 as part of a joint collaboration between NASA and NRC. The research seeks to generate icing conditions representative of those that occur inside a jet engine when ingesting ice crystals. In this test, an airfoil was exposed to mixed-phase icing conditions and the resulting ice accretions were recorded and analyzed. This paper details the specific objectives, procedures, and measurements which included the aero-thermal and cloud measurements. The objectives were built upon observations and hypothesis generated from several previous test campaigns regarding mixed-phase ice-crystal icing. The specific objectives included (A) ice accretions under different wet-bulb temperatures, (B) investigations of steady-state ice shapes previously reported in the literature, (C) total water content variations in search of a threshold for accretion, and (D) probe characterization related to measuring melt fraction which is important to characterize the mixed-phase condition. The resulting ice accretions and conditions leading to such accretions are intended to help extend NASA’s predictive ice-accretion codes to include conditions occurring in engine ice-crystal icing.

Icing↗

Exploring Hydrogenotrophic Methanogenesis: a Genome Scale Metabolic Reconstruction of Methanococcus maripaludis

ABSTRACT Hydrogenotrophic methanogenesis occurs in multiple environments, ranging from the intestinal tracts of animals to anaerobic sediments and hot springs. Energy conservation in hydrogenotrophic methanogens was long a mystery; only within the last decade was it reported that net energy conservation for growth depends on electron bifurcation. In this work, we focus on Methanococcus maripaludis , a well-studied hydrogenotrophic marine methanogen. To better understand hydrogenotrophic methanogenesis and compare it with methylotrophic methanogenesis that utilizes oxidative phosphorylation rather than electron bifurcation, we have built iMR539, a genome scale metabolic reconstruction that accounts for 539 of the 1,722 protein-coding genes of M. maripaludis strain S2. Our reconstructed metabolic network uses recent literature to not only represent the central electron bifurcation reaction but also incorporate vital biosynthesis and assimilation pathways, including unique cofactor and coenzyme syntheses. We show that our model accurately predicts experimental growth and gene knockout data, with 93% accuracy and a Matthews correlation coefficient of 0.78. Furthermore, we use our metabolic network reconstruction to probe the implications of electron bifurcation by showing its essentiality, as well as investigating the infeasibility of aceticlastic methanogenesis in the network. Additionally, we demonstrate a method of applying thermodynamic constraints to a metabolic model to quickly estimate overall free-energy changes between what comes in and out of the cell. Finally, we describe a novel reconstruction-specific computational toolbox we created to improve usability. Together, our results provide a computational network for exploring hydrogenotrophic methanogenesis and confirm the importance of electron bifurcation in this process. IMPORTANCE Understanding and applying hydrogenotrophic methanogenesis is a promising avenue for developing new bioenergy technologies around methane gas. Although a significant portion of biological methane is generated through this environmentally ubiquitous pathway, existing methanogen models portray the more traditional energy conservation mechanisms that are found in other methanogens. We have constructed a genome scale metabolic network of Methanococcus maripaludis that explicitly accounts for all major reactions involved in hydrogenotrophic methanogenesis. Our reconstruction demonstrates the importance of electron bifurcation in central metabolism, providing both a window into hydrogenotrophic methanogenesis and a hypothesis-generating platform to fuel metabolic engineering efforts.

59 BASIC BIOLOGICAL SCIENCES↗

DNA Repair Biomarkers Predict Response to Neoadjuvant Chemoradiotherapy in Esophageal Cancer

Purpose: The addition of neoadjuvant chemoradiotherapy prior to surgical resection for esophageal cancer has improved clinical outcomes in some trials. Pathologic complete response (pCR) following neoadjuvant therapy is associated with better clinical outcome in these patients, but only 22% to 40% of patients achieve pCR. Because both chemotherapy and radiotherapy act by inducing DNA damage, we analyzed proteins selected from multiple DNA repair pathways, using quantitative immunohistochemistry coupled with a digital pathology platform, as possible biomarkers of treatment response and clinical outcome. Methods and Materials: We identified 79 patients diagnosed with esophageal cancer between October 1994 and September 2002, with biopsy tissue available, who underwent neoadjuvant chemoradiotherapy prior to surgery at the Massachusetts General Hospital and used their archived, formalin-fixed, paraffin-embedded biopsy samples to create tissue microarrays (TMA). TMA sections were stained using antibodies against proteins in various DNA repair pathways including XPF, FANCD2, PAR, MLH1, PARP1, and phosphorylated MAPKAP kinase 2 (pMK2). Stained TMA slides were evaluated using machine-based image analysis, and scoring incorporated both the intensity and the quantity of positive tumor nuclei. Biomarker scores and clinical data were assessed for correlations with clinical outcome. Results: Higher scores for MLH1 (p = 0.018) and lower scores for FANCD2 (p = 0.037) were associated with pathologic response to neoadjuvant chemoradiation on multivariable analysis. Staining of MLH1, PARP1, XPF, and PAR was associated with recurrence-free survival, and staining of PARP1 and FANCD2 was associated with overall survival on multivariable analysis. Conclusions: DNA repair proteins analyzed by immunohistochemistry may be useful as predictive markers for response to neoadjuvant chemoradiotherapy in patients with esophageal cancer. These results are hypothesis generating and need confirmation in an independent data set.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

From soil to sequence: filling the critical gap in genome-resolved metagenomics is essential to the future of soil microbial ecology

Abstract Soil microbiomes are heterogeneous, complex microbial communities. Metagenomic analysis is generating vast amounts of data, creating immense challenges in sequence assembly and analysis. Although advances in technology have resulted in the ability to easily collect large amounts of sequence data, soil samples containing thousands of unique taxa are often poorly characterized. These challenges reduce the usefulness of genome-resolved metagenomic (GRM) analysis seen in other fields of microbiology, such as the creation of high quality metagenomic assembled genomes and the adoption of genome scale modeling approaches. The absence of these resources restricts the scale of future research, limiting hypothesis generation and the predictive modeling of microbial communities. Creating publicly available databases of soil MAGs, similar to databases produced for other microbiomes, has the potential to transform scientific insights about soil microbiomes without requiring the computational resources and domain expertise for assembly and binning.

59 BASIC BIOLOGICAL SCIENCES↗

On three dimensional object recognition and pose-determination: An abstraction based approach

A method of computing reliable Gaussian and mean curvature sign-map descriptors from the polynomial approximation of surfaces was demonstrated. Such descriptors which are invariant under perspective variation are suitable for hypothesis generation. A means for determining the pose of constructed geometric forms whose algebraic surface descriptors are nonlinear in terms of their orienting parameters was developed. This was done by means of linear functions which are capable of approximating nonlinear forms and determining their parameters. It was shown that biquadratic surfaces are suitable companion linear forms for cylindrical approximation and parameter estimation. The estimates provided the initial parametric approximations necessary for a nonlinear regression stage to fine tune the estimates by fitting the actual nonlinear form to the data. A hypothesis-based split-merge algorithm for extraction and pose determination of cylinders and planes which merge smoothly into other surfaces was developed. It was shown that all split-merge algorithms are hypothesis-based. A finite-state algorithm for the extraction of the boundaries of run-length regions was developed. The computation takes advantage of the run list topology and boundary direction constraints implicit in the run-length encoding.

Quek, Kok How Francis↗

Neurochemical Evidence of Potential Neurotoxicity After Prophylactic Cranial Irradiation

Purpose: To examine whether cerebrospinal fluid biomarkers for neuroaxonal damage, neuroglial activation, and amyloid β–related processes could characterize the neurochemical response to cranial radiation. Methods and Materials: Before prophylactic cranial irradiation (PCI) of patients with small cell lung cancer, each patient underwent magnetic resonance imaging of the brain, lumbar puncture, and Mini-Mental State Examination of cognitive function. These examinations were repeated at approximately 3 and 12 months after radiation. Results: The major findings were as follows. (1) Cerebrospinal fluid markers for neuronal and neuroglial injury were elevated during the subacute phase after PCI. Neurofilament and T-tau increased 120% and 50%, respectively, after PCI (P<.05). The same was seen for the neuroglial markers YKL-40 and glial fibrillary acidic protein, which increased 144% and 106%, respectively, after PCI (P<.05). (2) The levels of secreted amyloid precursor protein-α and -β were reduced 44% and 46%, respectively, 3 months after PCI, and the levels continued to decrease as long as 1 year after treatment (P<.05). (3) Mini-Mental State Examination did not reveal any cognitive decline, indicating that a more sensitive test should be used in future studies. Conclusion: In conclusion, we were able to detect radiation therapy–induced changes in several markers reflecting neuronal injury, inflammatory/astroglial activation, and altered amyloid precursor protein/amyloid β metabolism, despite the low number of patients and quite moderate radiation doses (20-30 Gy). These changes are hypothesis generating and could potentially be used to assess the individual risk of developing long-term symptoms of chronic encephalopathy after PCI. This has to be evaluated in large studies with extended clinical follow-up and more detailed neurocognitive assessments.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Moving Toward Focal Therapy in Prostate Cancer: Dual-Isotope Permanent Seed Implants as a Possible Solution

Purpose: To compare the ability of single- and dual-isotope prostate seed implants to escalate biologically effective dose (BED) to foci of disease while reducing prescription dose to the prostate. Methods and Materials: Nine plans, using {sup 125}I, {sup 103}Pd, and {sup 131}Cs alone and in combination were created retrospectively for 2 patients. Ultrasound and MRI/MRS datasets were used for treatment planning. Voxel-by-voxel BED was calculated for single- and dual-isotope plans. Equivalent uniform BED (EUBED) was used to compare plans. The MRS-positive planning target volumes (PTV{sub i}) were delineated along with PTV (prostate + 5 mm), rectum, and urethra. Single-isotope implants, prescribed to conventional doses, were generated to achieve good PTV coverage. The PTV{sub i} were prospectively used to generate implants using mixtures of isotopes. For mixed-radioisotope implants, we also explored the impact on EUBED of lowering prescription doses by 15%. Results: The EUBED of PTV{sub i} in the setting of primary {sup 125}I implant increased 20-66% when {sup 103}Pd and {sup 131}Cs were used compared with {sup 125}I boost. Decreasing prescription dose by 15% in mixed-isotope implants results in a potential 10% reduction in urethral EUBED with preservation of PTV coverage while still boosting PTV{sub i} (up to 80%). When radiobiologic parameters corresponding to more-aggressive disease are assigned to foci, faster-decaying isotopes used in mixed implants have the potential to preserve the equivalent biological effect of mono-isotope implants considering less-aggressive disease distributed in the entire prostate. Conclusions: This is a hypothesis-generating study proposing a treatment paradigm that could be the middle ground between whole-gland irradiation and focal-only treatment. The use of two isotopes concurrent with decreasing the minimal peripheral dose is shown to increase EUBED of selected subvolumes while preserving the therapeutic effect at the level of the gland.

62 RADIOLOGY AND NUCLEAR MEDICINE↗