Improving Object Detectors by Exploiting Bounding Boxes for Augmentation Design
Not Available
97 MATHEMATICS AND COMPUTING↗
SEARCH · Engineering Papers
Results for “design augmentation”
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.
Machine learning for endoleak detection after endovascular aortic repair
Diagnosis of endoleak following endovascular aortic repair (EVAR) relies on manual review of multi-slice CT angiography (CTA) by physicians which is a tedious and time-consuming process that is susceptible to error. We evaluate the use of a deep neural network for the detection of endoleak on CTA for post-EVAR patients using a novel data efficient training approach. 50 CTAs and 20 CTAs with and without endoleak respectively were identified based on gold standard interpretation by a cardiovascular subspecialty radiologist. The Endoleak Augmentor, a custom designed augmentation method, provided robust training for the machine learning (ML) model. Predicted segmentation maps underwent post-processing to determine the presence of endoleak. The model was tested against 3 blinded general radiologists and 1 blinded subspecialist using a held-out subset (10 positive endoleak CTAs, 10 control CTAs). Model accuracy, precision and recall for endoleak diagnosis were 95%, 90% and 100% relative to reference subspecialist interpretation (AUC = 0.99). Accuracy, precision and recall was 70/70/70% for generalist1, 50/50/90% for generalist2, and 90/83/100% for generalist3. The blinded subspecialist had concordant interpretations for all test cases compared with the reference. In conclusion, our ML-based approach has similar performance for endoleak diagnosis relative to subspecialists and superior performance compared with generalists.
60 APPLIED LIFE SCIENCES↗
Input-response space-filling designs
Traditional space-filling designs are a convenient way to explore throughout an input space of flexible dimension and have design points close to any region where future predictions might be of interest. In some applications, there may be a model connecting the input factors to the response(s), which provides an opportunity to consider the spacing not only in the input space but also in the response space. In this paper, we present an approach for leveraging current understanding of the relationship between inputs and responses to generate designs that allow the experimenter to flexibly balance the spacing in these two regions to find an appropriate design for the experimental goals. Applications where good spacing of the observed response values include calibration problems where the goal is to demonstrate the adequacy of the model across the range of the responses, sensitivity studies where the outputs from a submodel may be used as inputs for subsequent models, and inverse problems where the outputs of a process will be used in the inverse prediction for the unknown inputs. Here, we use the multi-objective optimization method of Pareto fronts to generate multiple non-dominated designs with different emphases on the input and response space-filling criteria from which the experimenter can choose. The methods are illustrated through several examples and a chemical engineering case study.
42 ENGINEERING↗
A new perspective on density and strength loss profiles at the surface of thermally oxidized nuclear graphite
Oxidation of graphite components could influence their designed life in a high-temperature nuclear reactor. The oxidized regions could potentially lower the allowed stress capacity. The American Society of Mechanical Engineers rules for the design and construction of graphite-moderated reactors recommend that subsurface regions that might become excessively damaged by oxidation during reactor operation be identified and excluded from geometry and stress calculations. Identification of oxidation-affected regions is possible, in principle, through complex modeling exercises of reactor behavior during hypothetical accident scenarios coupled with graphite oxidation models, but this procedure may not have the precision needed for informed decisions. Here, this paper proposes an alternate method, based on interpretation of a series of well-designed oxidation experiments, which could augment the designer's tools. The procedure is illustrated by data on oxidation by air of several graphite grades (NBG-18, PCEA, IG-110, R4-650) that are corroborated with independent literature information, when available. The Wichner model for graphite oxidation used for this analysis provides conservative results that could be quickly implemented in the design process.
36 MATERIALS SCIENCE↗
Machine-learning for designing nanoarchitectured materials by dealloying
Machine learning-augmented materials design is an emerging method for rapidly developing new materials. It is especially useful for designing new nanoarchitectured materials, whose design parameter space is often large and complex. Metal-agent dealloying, a materials design method for fabricating nanoporous or nanocomposite from a wide range of elements, has attracted significant interest. Here, a machine learning approach is introduced to explore metal-agent dealloying, leading to the prediction of 132 plausible ternary dealloying systems. A machine learning-augmented framework is tested, including predicting dealloying systems and characterizing combinatorial thin films via automated and autonomous machine learning-driven synchrotron techniques. This work demonstrates the potential to utilize machine learning-augmented methods for creating nanoarchitectured thin films.
36 MATERIALS SCIENCE↗
Artificial Intelligence and Machine Learning for Bioenergy Research: Opportunities and Challenges
The integration of artificial intelligence and machine learning (AI/ML) with automated experimentation, genomics, biosystems design, and bioprocessing technologies is poised to revolutionize scientific investigation and, particularly, bioenergy research. To identify the opportunities and challenges in this emerging research area, the U.S. Department of Energy’s (DOE) Biological and Environmental Research program (BER) and Bioenergy Technologies Office (BETO) held a joint virtual workshop on AI/ML for Bioenergy Research (AMBER) on August 23–25, 2022. These interests have since been amplified in a September 2022 Executive Order, “Advancing Biotechnology and Biomanufacturing Innovation for a Sustainable, Safe, and Secure U.S. Bioeconomy,” to promote a whole-of government approach to biotechnology development (White House 2022). Approximately 50 scientists with various backgrounds and expertise from academia, industry, and DOE national laboratories met to discuss the opportunities and challenges of AI/ML for bioenergy research. Workshop participants were tasked with assessing the potential for AI/ML and laboratory automation to advance biological understanding and engineering in general. They particularly examined how integrating AI/ML tools with laboratory automation could accelerate biosystems design and optimize biomanufacturing. Discussions included the data and computational infrastructure needed to augment biosystems design applications and the expertise and workforce development efforts urgently required to shift integrated systems toward bioenergy research more broadly. Participants discussed many existing and future applications of AI/ML for biosystems design ranging from enzymes to plants and microbes, microbiomes, and bioprocess development. They also identified three key categories of scientific and technical opportunities and challenges: high-quality data, AI/ML algorithms, and laboratory automation. Several main takeaways emerged from the workshop: 1. Numerous AI/ML and automated experimentation applications exist for a variety of DOE mission needs in energy and the environment; 2. Exemplary research grand challenges for which AI/ML could provide solutions include: building microbes and microbial communities to specifications, developing closed-loop autonomous design and control for biosystems design, and advancing scale-up and automation; 3. Lack of sufficient high-quality, annotated data hinders the development of AI/ML applications; 4. New and improved AI/ML tools are needed, particularly those meeting the specific needs of the BER and BETO research communities; 5. Trade-offs in performance, cost, and reliability exist between deploying commercially available versus building custom-developed instrumentation and software for automated or autonomous experimentation; translation of manual to automated or autonomous methods is often a nontrivial endeavor; 6. Training a new generation of young scientists who can develop and apply AI/ML tools is needed to solve long-standing scientific challenges in bioenergy research. The integration of AI/ML tools and automated experimentation represents a new data-driven research paradigm complementary to the traditional hypothesis-driven research paradigm. This paradigm accelerates design and optimization of biological systems and processes for a variety of DOE mission needs in energy and the environment. The AMBER workshop broadly explored the potential of this new paradigm for bioenergy research, of particular interest to BER and BETO, and identified key challenges and opportunities that DOE can address in the coming years by leveraging its unique capabilities and resources.
59 BASIC BIOLOGICAL SCIENCES↗
A detailed study of interpretability of deep neural network based top taggers
Abstract Recent developments in the methods of explainable artificial intelligence (XAI) allow researchers to explore the inner workings of deep neural networks (DNNs), revealing crucial information about input–output relationships and realizing how data connects with machine learning models. In this paper we explore interpretability of DNN models designed to identify jets coming from top quark decay in high energy proton–proton collisions at the Large Hadron Collider. We review a subset of existing top tagger models and explore different quantitative methods to identify which features play the most important roles in identifying the top jets. We also investigate how and why feature importance varies across different XAI metrics, how correlations among features impact their explainability, and how latent space representations encode information as well as correlate with physically meaningful quantities. Our studies uncover some major pitfalls of existing XAI methods and illustrate how they can be overcome to obtain consistent and meaningful interpretation of these models. We additionally illustrate the activity of hidden layers as neural activation pattern diagrams and demonstrate how they can be used to understand how DNNs relay information across the layers and how this understanding can help to make such models significantly simpler by allowing effective model reoptimization and hyperparameter tuning. These studies not only facilitate a methodological approach to interpreting models but also unveil new insights about what these models learn. Incorporating these observations into augmented model design, we propose the particle flow interaction network model and demonstrate how interpretability-inspired model augmentation can improve top tagging performance.
97 MATHEMATICS AND COMPUTING↗
Dynamic machine learning-based optimization algorithm to improve boiler efficiency
With decreasing computational costs, improvement in algorithms, and the aggregation of large industrial and commercial datasets, machine learning is becoming a ubiquitous tool for process and business innovations. Machine learning is still lacking applications in the field of dynamic optimization for real-time control. This work presents a novel framework for performing constrained dynamic optimization using a recurrent neural network model combined with a metaheuristic optimizer. The framework is designed to augment an existing control system and is purely data-driven, like most industrial Model Predictive Control applications. Several recurrent neural network models are compared as well as several metaheuristic optimizers. Hyperparameters and optimizer parameters are tuned with parameter sweeps, and the resulting values are reported. Further, the best parameters for each optimizer and model combination are demonstrated in closed-loop control of a dynamic simulation, and several recommendations are made for generalizing this framework to other systems. Up to 0.953% improvement is realized over the non-optimized case for a simulated coal-fired boiler. While this is not a large improvement in percentage, the total economic impact is $991,000 per year, and this study builds a foundation for future machine learning with dynamic optimization.
32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
An approach for collaborative development of a federated biomedical knowledge graph-based question-answering system: Question-of-the-Month challenges
Knowledge graphs have become a common approach for knowledge representation. Yet, the application of graph methodology is elusive due to the sheer number and complexity of knowledge sources. In addition, semantic incompatibilities hinder efforts to harmonize and integrate across these diverse sources. As part of The Biomedical Translator Consortium, we have developed a knowledge graph–based question-answering system designed to augment human reasoning and accelerate translational scientific discovery: the Translator system. We have applied the Translator system to answer biomedical questions in the context of a broad array of diseases and syndromes, including Fanconi anemia, primary ciliary dyskinesia, multiple sclerosis, and others. A variety of collaborative approaches have been used to research and develop the Translator system. One recent approach involved the establishment of a monthly “Question-of-the-Month (QotM) Challenge” series. Herein, we describe the structure of the QotM Challenge; the six challenges that have been conducted to date on drug-induced liver injury, cannabidiol toxicity, coronavirus infection, diabetes, psoriatic arthritis, and -related phenotypes; the scientific insights that have been gleaned during the challenges; and the technical issues that were identified over the course of the challenges and that can now be addressed to foster further development of the prototype Translator system. We close with a discussion on Large Language Models such as ChatGPT and highlight differences between those models and the Translator system.
60 APPLIED LIFE SCIENCES↗
CAHS: Context-Aware Homology Search
Protein homology search is foundational to bioinformatics: it supports annotation transfer, structure/function inference, and evolutionary analysis over rapidly expanding sequence repositories (e.g., UniProtKB). Profile hidden Markov models (pHMMs), as implemented in HMMER, remain the most widely trusted approach because they provide statistically calibrated E-values; however, their gap behavior is fixed once a profile is trained, despite biological evidence that insertion/deletion tolerance varies across flexible loops and intrinsically disordered regions. We present CAHS (Context-Aware Homology Search), a lightweight query-time adapter for pHMM search that incorporates learned and biologically motivated signals without changing HMMER's downstream search pipeline or its calibrated E-value reporting. Given a query sequence, CAHS computes per-residue representations from a protein language model and a disorder predictor, maps these to profile coordinates, and modulates only match-state transition rows (gap-open and gap-extension probabilities) while preserving Plan7 constraints. We comprehensively evaluate CAHS across six structurally diverse protein families and multi-domain architectures against a 570k-sequence target corpus. CAHS expands detection capability, retrieving thousands of additional remote homologs at relaxed thresholds by maintaining alignment quality through flexible regions. For multi-domain proteins, context-aware modulation resolves 94% of fragmented alignments. Crucially, CAHS preserves hit-set invariance at stringent operating points (E<10-10), demonstrating increased statistical confidence without inflating false positives. Furthermore, sharper statistical distinction between homologs and background noise during early filter stages yields up to a 3.87× acceleration in end-to-end wall-clock time on high-performance computing clusters. Overall, CAHS illustrates a practical AI-for-science design pattern: augmenting a trusted probabilistic model with query-specific learned signals to improve interpretable, reproducible inference in data-rich biology.
Bhattaram, Swethasree [Georgia Institute of Techno↗
Enhanced Multi-Dimensional Inversion Through Target-Specific Inversion Parameter Bounds With an Application to Crosswell Electromagnetic for Sequestration Monitoring
In geophysical inversions, lower and upper model parameter bounds are a means of solution stabilization. Further, constraints that intend to let only geologically plausible inverse solutions pass are amenable to lower and upper bounds. Reliable prior information is paramount to construct such bound constraints. It is common practice to narrow and widen bound intervals for regions of, respectively, more and less certain prior information. Contrary to this practice, we experiment with widened bound intervals in zones that are poorly resolved by a given survey configuration but where prior information would suggest structural anomalies of interest. The purpose of enlarged parameter bounds that correlate spatially with predefined targets is to let the inversion explore a larger solution space, thus increasing the potential to resolve otherwise hidden anomalies. Application of the method is based on a carbon-sequestration baseline (pre-injection) crosswell electromagnetic (EM) field survey at the Containment and Monitoring Institute Field Research Station (Alberta, Canada), where impeded measurements led to generally reduced sensitivities for the interwell region. Synthetic-data proofs of concept use augmented bounds designed to boost the resolution of artificial plume targets, indicating an enhanced illumination compared to constant bounds. Comparative field data inversions with spatially variable bounds constructed from prior resistivity and velocity information highlight non-horizontal baseline structures.
3-D electrical resistivity imaging↗
SHOTEAM: Superalloy Heat exchangers Optimized for Temperature Extremes and Additive Manufacturability (Final Technical Report)
UCLA developed an extreme-condition heat exchanger technology targeted to ultra-high efficiency hybrid aviation power cycles. The heat exchanger targeted operation at 50 kW (thermal) at supercritical CO 2 pressures of 80 and 250 bar (1160 and 3626 psi) in hot and cold streams, respectively and at a hot-stream inlet temperature of 800°C (1472°F). A metallic superalloy capable of withstanding high temperature and pressure was used to fabricate a microtube shell-and-tube-based design supplemented with 3D-printed tube augmentations. The optimized design enhanced overall heat transfer while maintaining a small overall form factor and low weight. This class of heat exchanger could dramatically improve efficiency, power density, and cost effectiveness for new hybrid aviation power cycles, thus enabling economically feasible routes to air vehicle propulsion with substantially less CO 2 emissions to the environment.
32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
HydrogenAR: Interactive Data-Driven Presentation of Dispenser Reliability: Preprint
When delivering presentations to a co-located audience, we typically use slides with text and 2D graphics to complement the spoken narrative. Though presentations have largely been explored on 2D media, augmented reality (AR) allows presentation designers to add data and augmentations to existing physical infrastructure on display. This could provide a more engaging experience to the audience and support comprehension. With HydrogenAR, we present a novel application that leverages the benefits of data-driven storytelling with those of AR to explain the unique challenges of hydrogen dis- penser reliability. Utilizing physical props, situated data, and virtual augmentations and animations, HydrogenAR serves as a unique presentation tool, particularly critical for stakeholders, tour groups, and VIPs. HydrogenAR is a product of multiple collaborative de- sign iterations with a local Hydrogen Fuel research team and is evaluated through interviews with team members and a user study with end-users to evaluate the usability and quality of the interactive AR experience. Through this work, we provide design considera- tions for AR data-driven presentations and discuss how AR could be used for innovative content delivery beyond traditional slide-based presentations.
augmented reality↗
Design and performance of parallel-channel nanocryotrons in magnetic fields
We introduce a design modification to conventional geometry of the cryogenic three-terminal switch, the nanocryotron (nTron). The conventional geometry of nTrons is modified by including parallel current-carrying channels, an approach aimed at enhancing the device's performance in magnetic field environments. The common challenge in nTron technology is to maintain efficient operation under varying magnetic field conditions. Here, we show that the adaptation of parallel channel configurations leads to an enhanced gate signal sensitivity, an increase in operational gain, and a reduction in the impact of superconducting vortices on nTron operation within magnetic fields up to 1 T. Contrary to traditional designs that are constrained by their effective channel width, the parallel nanowire channels permits larger nTron cross sections, further bolstering the device's magnetic field resilience while improving electro-thermal recovery times due to reduced local inductance. This advancement in nTron design not only augments its functionality in magnetic fields but also broadens its applicability in technological environments, offering a simple design alternative to existing nTron devices.
75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
Heat transfer augmentation of natural convection heat sink through notched fin design
A novel notched fin heat sink with a central opening is proposed in the present study to augment the heat transfer performance of the natural convection heat sink under horizontal configuration. Thermal and airflow behavior in the heat sinks were studied using both experimental and numerical methods. The results showed that the performance of the conventional heat sink can be enhanced by optimizing its fin design. The basic heat sink model provided a 17% reduction in thermal resistance when its fin height and fin spacing were doubled, and it was accompanied by a 21% increase in surface area. The heat sink with a notched fin design with a less surface area offered a 13% lower thermal resistance compared to the basic model by entraining more airflow into the interior fin region. Providing a central opening in the notched fin heat sink can further reduce the thermal resistance by about 13%. Further, this study also analyzed the influence of emissivity of the heat sink on natural convection conditions. The findings demonstrated that the heat sink with an emissivity of 0.9 can reduce the thermal resistance by nearly 23% compared to one with an emissivity of 0.15.
42 ENGINEERING↗
Synergistic Porosity and Charge Effects in a Supramolecular Porphyrin Cage Promote Efficient Photocatalytic CO 2 Reduction**
Abstract We present a supramolecular approach to catalyzing photochemical CO 2 reduction through second‐sphere porosity and charge effects. An iron porphyrin box ( PB ) bearing 24 cationic groups, FePB‐2(P) , was made via post‐synthetic modification of an alkyne‐functionalized supramolecular synthon. FePB‐2(P) promotes the photochemical CO 2 reduction reaction (CO 2 RR) with 97 % selectivity for CO product, achieving turnover numbers (TON) exceeding 7000 and initial turnover frequencies (TOF max ) reaching 1400 min −1 . The cooperativity between porosity and charge results in a 41‐fold increase in activity relative to the parent Fe tetraphenylporphyrin ( FeTPP ) catalyst, which is far greater than analogs that augment catalysis through porosity ( FePB‐3(N ), 4‐fold increase) or charge (Fe p ‐tetramethylanilinium porphyrin ( Fe‐ p ‐TMA ), 6‐fold increase) alone. This work establishes that synergistic pendants in the secondary coordination sphere can be leveraged as a design element to augment catalysis at primary active sites within confined spaces.
An, Lun↗
Inverse design of anisotropic microstructures using physics-augmented neural networks
Explore the source record for details and available documents.
Jadoon, Asghar [University of Texas at Austin]↗
Synergistic Porosity and Charge Effects in a Supramolecular Porphyrin Cage Promote Efficient Photocatalytic CO 2 Reduction**
We present a supramolecular approach to catalyzing photochemical CO 2 reduction through second-sphere porosity and charge effects. An iron porphyrin box (PB) bearing 24 cationic groups, FePB-2(P), was made via post-synthetic modification of an alkyne-functionalized supramolecular synthon. FePB-2(P) promotes the photochemical CO 2 reduction reaction (CO 2 RR) with 97 % selectivity for CO product, achieving turnover numbers (TON) exceeding 7000 and initial turnover frequencies (TOF max ) reaching 1400 min -1 . The cooperativity between porosity and charge results in a 41-fold increase in activity relative to the parent Fe tetraphenylporphyrin (FeTPP) catalyst, which is far greater than analogs that augment catalysis through porosity (FePB-3(N), 4-fold increase) or charge (Fe p-tetramethylanilinium porphyrin (Fe-p-TMA), 6-fold increase) alone. In conclusion, this work establishes that synergistic pendants in the secondary coordination sphere can be leveraged as a design element to augment catalysis at primary active sites within confined spaces.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗