Search NASA⌕ Search

SEARCH · Search NASA

Results for “Distributed training”

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 307 records · Page 17

Search for top-quark production via flavor-changing neutral currents in W+1 jet events at CDF

We report on the first search for top-quark production via flavor-changing neutral-current (FCNC) interactions in the non-standard-model process u(c)+g {yields} t using p{bar p} collision data collected by the CDF II detector. The data set corresponds to an integrated luminosity of 2.2 fb{sup -1}. The candidate events feature the signature of semileptonic top-quark decays and are classified as signal-like or background-like by an artificial neural network trained on simulated events. The observed discriminant distribution is in good agreement with the one predicted by the standard model and provides no evidence for FCNC top-quark production, resulting in a Bayesian upper limit on the production cross section {sigma}(u(c)+g {yields} t) < 1.8 pb at the 95% confidence level. Using theoretical predictions we convert the cross-section limit to upper limits on FCNC branching ratios: {Beta}(t {yields} u + g) < 3.9 x 10{sup -4} and {Beta}(t {yields} c + g) < 5.7 x 10{sup -3}.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Investigation of Beam Loading Effects for the Neutrino Factory Muon Accelerator

The International design study (IDS) study showed that a Neutrino Factory [1] seems to be the most promising candidate for the next phase of high precision neutrino oscillation experiments. One part of the increased precision is due to the fact that in a Neutrino Factory the decay of muons produces a neutrino beam with narrow energy distribution and divergence. The effect of beam loading on the energy distribution of the muon beam in the Neutrino Factory decay rings has been investigated numerically. The simulations have been performed using the baseline accelerator design including cavities for different number of bunch trains and bunch train timing. A detailed analysis of the beam energy distribution expected is given together with a discussion of the energy spread produced by the gutter acceleration in the FFAG and the implications for the neutrino oscillation experiments will be presented.

43 PARTICLE ACCELERATORS↗

Reinforcement Learning for Distribution Grid Optimization (PyCIGAR) v0.1

PyCIGAR is a python software package that merges off-the-shelf reinforcement learning libraries (RLLib and Ray) with electric power distribution system simulation tools (OpenDSS and a custom power flow solver built by LBL). PyCIGAR enables the training of neural networks to optimize the behavior of different components in the electric distribution grid, such as control systems in photovoltaic rooftop solar inverters and electric battery storage systems. The software package has been used to train neural networks to update settings in photovoltaic rooftop solar inverter control systems to mitigate cyber attacks on other solar photovoltaic rooftop devices.

Arnold, Daniel↗

Large Scale Caching and Streaming of Training Data for Online Deep Learning

The training of deep neural network models on large data remains a difficult problem, despite progress towards scalable techniques. In particular, there is a mismatch between the random but predetermined order in which AI flows select training samples and the streaming I/O patterns for which traditional HPC data storage (e.g., parallel file systems) are designed. In addition, as more data are obtained, it is feasible neither simply to train learning models incrementally, due to catastrophic forgetting (i.e., bias towards new samples), nor to train frequently from scratch, due to prohibitive time and/or resource constraints. In this paper, we study data management techniques that combine caching and streaming with rehearsal support in order to enable efficient access to training samples in both offline training and continual learning. We revisit state-of-art streaming approaches based on data pipelines that transparently handle prefetching, caching, shuffling, and data augmentation, and discuss the challenges and opportunities that arise when combining these methods with data-parallel training techniques. We also report on preliminary experiments that evaluate the I/O overheads involved in accessing the training samples from a parallel file system (PFS) under several concurrency scenarios, highlighting the impact of the PFS on the design of the data pipelines.

data pipelines↗

A Machine‐Learning‐Assisted Stochastic Cloud Population Model as a Parameterization of Cumulus Convection

Abstract A machine‐learning‐assisted stochastic cloud population model is coupled with the Advanced Research Weather Research and Forecasting (WRF) model to represent fluctuations in the cloud‐base mass flux associated with the life cycles and interactions among cumulus convection cells. In this cloud population model, the size distribution and the associated cloud‐base mass flux of the convective cells are related to their previous state and to the change in the total convective area via a transition function. The convective area tendency in turn is assumed to depend on the cloud‐base mass flux that is resolved by the host WRF model. The transition function is represented by a single hidden‐layer neural network trained by the evolution of convective cell size distributions in a 1‐km grid‐spacing WRF simulation run over the Australian Monsoon region. At every grid point of the host model, the cloud population model predicts the cell size and cloud‐base mass flux distributions from which a random sample of cells is fed to an entraining parcel model that calculates precipitation as well as the associated liquid water potential temperature and total moisture tendencies. These tendencies are averaged over the cells and provided to the host model. Several regional simulations are performed over tropical and midlatitude domains to test this as a potential approach to scale‐aware parameterization. It is shown that such an approach could be a new promising path to simulating realistic precipitation statistics and propagation of precipitation associated with the Madden‐Julian Oscillation while maintaining realistic depictions of the diurnal cycle over both land and ocean.

54 ENVIRONMENTAL SCIENCES↗

Temperature Field Reconstruction of Surfaces Heated Through Radiative Heat Transfer Using Convolutional Neural Networks

Microreactors could play a crucial role in decarbonizing our energy portfolio. However, their development and implementation come with specific challenges, particularly regarding cost. Due to their compact size and the harsh operational environment, collecting real-time data on reactor operation can be challenging. Many probe designs are unable to withstand extreme conditions (e.g., temperature, radiation) in the reactor. In this context, using convolutional neural networks (CNNs) can pave the way for developing a nonintrusive approach that relies solely on ex-core sensors. A well-trained physics-informed CNN can reconstruct the distribution of a given physical quantity over a domain using only a few sensors, allowing us to reconstruct the desired field distribution even in a limited space or complex geometries where a large array of sensors is impractical. In this work, we present the initial steps toward developing a real-time tool for monitoring the thermal behavior of nuclear reactor pressure vessels. Based on an experimental setup, a computational model using the Multiphysics Object-Oriented Simulation Environment (moose) framework was built, where the Ray Tracing and Heat Conduction modules were used to evaluate the temperature distribution over a convex metal surface heated through radiative heat transfer. This metal surface represents a section of a heated nuclear reactor vessel wall. The model also accounts for solid mechanics physics through the moose Solid Mechanics module. In situ experimental data, acquired from a Texas A&M facility, were used to validate the computational model. Part of the data generated by the moose model was used to train the convolutional neural network to reconstruct the vessel wall's outer surface temperature. The CNN generalization was then compared against the experimental and computational data.

Aldeia Machado, Luiz Carlos↗

Machine learning assisted bayesian inference of mix and hot-spot conditions in NIF implosions

Experiments on the National Ignition Facility (NIF) have provided clear evidence of ablator material mixing into the Hot-Spot, leading to degraded performance. However, inferring the amount of mix and Hot-Spot conditions from typical experimental observations (e.g. x-ray spectra and images) is highly challenging. Here, we have developed an analysis method that utilizes machine learning assisted Bayesian inference to find the probability distributions of the Hot-Spot and mix conditions. This approach uses a neural network, trained on an idealized 2-dimensional representation of the Hot-Spot and mix distribution, and Bayesian inference to find the statistical distributions of Hot-Spot conditions that provide a match with observations. We have tested this method with synthetic data from simulations.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Transferable predictions of energetic and structural properties for refractory solid solution alloys across chemical compositions

We present a data-efficient approach to train graph neural networks (GNNs) on density functional theory (DFT) data for accurate and transferable predictions of energetic and structural properties of refractory solid solution alloys in the niobium-tantalum-vanadium (Nb-Ta-V) chemical space. We start by training the GNN model only on DFT data that describes refractory binary alloys niobium-tantalum (Nb-Ta), niobium-vanadium (Nb-V), and tantalum-vanadium (Ta-V) to predict formation enthalpy and root mean squared displacement. Once trained, the GNN predictions are tested on DFT data describing refractory ternary alloys Nb-Ta-V. While, unsurprisingly, direct transferability from binary to ternary is not sufficiently accurate, augmenting the training with only 1% of the available ternary data (uniformly distributed across the entire range of chemical compositions) improves significantly the quality of the GNN predictions. For comparison, we assess the transferability in the opposite direction by training GNN models on ternary Nb-Ta-V data and making predictions on binaries Nb-Ta, Nb-V, and Ta-V, which exhibits notably higher predictive errors. The proposed methodology, which favors transferability from lower-component to higher-component alloys, offers an efficient path towards avoiding the curse of dimensionality incurred when collecting DFT data for discovery and design of multi-component disordered alloys.

Density functional theory calculations↗

Exploring the energy landscape of RBMs: reciprocal space insights into bosons, hierarchical learning and symmetry breaking

Deep generative models have become ubiquitous due to their ability to learn and sample from complex distributions. Despite the proliferation of various frameworks, the relationships among these models remain largely unexplored, a gap that hinders the development of a unified theory of AI learning. In this work, we address two central challenges: clarifying the connections between different deep generative models and deepening our understanding of their learning mechanisms. We focus on Restricted Boltzmann Machines (RBMs), a class of generative models known for their universal approximation capabilities for discrete distributions. By introducing a reciprocal space formulation for RBMs, we reveal a connection between these models, diffusion processes, and systems of coupled bosons. Our analysis shows that at initialization, the RBM operates at a saddle point, where the local curvature is determined by the singular values of the weight matrix, whose distribution follows the Marc̆enko-Pastur law and exhibits rotational symmetry. During training, this rotational symmetry is broken due to hierarchical learning, where different degrees of freedom progressively capture features at multiple levels of abstraction. This leads to a symmetry breaking in the energy landscape, reminiscent of Landau’s theory. This symmetry breaking in the energy landscape is characterized by the singular values and the weight matrix eigenvector matrix. We derive the corresponding free energy in a mean-field approximation. We show that in the limit of infinite size RBM, the reciprocal variables are Gaussian distributed. Our findings indicate that in this regime, there will be some modes for which the diffusion process will not converge to the Boltzmann distribution. To illustrate our results, we trained replicas of RBMs with different hidden layer sizes using the MNIST dataset. Our findings not only bridge the gap between disparate generative frameworks but also shed light on the fundamental processes underpinning learning in deep generative models.

97 MATHEMATICS AND COMPUTING↗

MO-E-18C-02: Hands-On Monte Carlo Project Assignment as a Method to Teach Radiation Physics

Purpose: To present a hands-on project on Monte Carlo methods (MC) recently added to the curriculum and to discuss the students' appreciation. Methods: Since 2012, a 1.5 hour lecture dedicated to MC fundamentals follows the detailed presentation of photon and electron interactions. Students also program all sampling steps (interaction length and type, scattering angle, energy deposit) of a MC photon transport code. A handout structured in a step-by-step fashion guides student in conducting consistency checks. For extra points, students can code a fully working MC simulation, that simulates a dose distribution for 50 keV photons. A kerma approximation to dose deposition is assumed. A survey was conducted to which 10 out of the 14 attending students responded. It compared MC knowledge prior to and after the project, questioned the usefulness of radiation physics teaching through MC and surveyed possible project improvements. Results: According to the survey, 76% of students had no or a basic knowledge of MC methods before the class and 65% estimate to have a good to very good understanding of MC methods after attending the class. 80% of students feel that the MC project helped them significantly to understand simulations of dose distributions. On average, students dedicated 12.5 hours to the project and appreciated the balance between hand-holding and questions/implications. Conclusion: A lecture on MC methods with a hands-on MC programming project requiring about 14 hours was added to the graduate study curriculum since 2012. MC methods produce “gold standard” dose distributions and slowly enter routine clinical work and a fundamental understanding of MC methods should be a requirement for future students. Overall, the lecture and project helped students relate crosssections to dose depositions and presented numerical sampling methods behind the simulation of these dose distributions. Research funding from governments of Canada and Quebec. PP acknowledges partial support by the CREATE Medical Physics Research Training Network grant of the Natural Sciences and Engineering Research Council (Grant number: 432290)

07 ISOTOPE AND RADIATION SOURCES↗

Data-Driven Affinely Adjustable Robust Volt/VAr Control

Recent years have seen the increasing proliferation of distributed energy resources with intermittent power outputs, posing new challenges to the voltage management in distribution networks. To this end, this paper proposes a data-driven affinely adjustable robust Volt/VAr control (AARVVC) scheme, which modulates the smart inverter’s reactive power in an affine function of its active power, based on the voltage sensitivities with respect to real/reactive power injections. To achieve a fast and accurate estimation of voltage sensitivities, we propose a data-driven method based on deep neural network (DNN), together with a rule-based bus-selection process using the bidirectional search method. Our method only uses the operating statuses of selected buses as inputs to DNN, thus significantly improving the training efficiency and reducing information redundancy. Finally, a distributed consensus-based solution, based on the alternating direction method of multipliers (ADMM), for the AARVVC is applied to decide the inverter’s reactive power adjustment rule with respect to its active power. Only limited information exchange is required between each local agent and the central agent to obtain the slope of the reactive power adjustment rule, and there is no need for the central agent to solve any (sub)optimization problems. Finally, numerical results on the modified IEEE-123 bus system validate the effectiveness and superiority of the proposed data-driven AARVVC method.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Real-time image denoising of mixed Poisson–Gaussian noise in fluorescence microscopy images using ImageJ

Fluorescence microscopy imaging speed is fundamentally limited by the measurement signal-to-noise ratio (SNR). To improve image SNR for a given image acquisition rate, computational denoising techniques can be used to suppress noise. However, common techniques to estimate a denoised image from a single frame either are computationally expensive or rely on simple noise statistical models. These models assume Poisson or Gaussian noise statistics, which are not appropriate for many fluorescence microscopy applications that contain quantum shot noise and electronic Johnson–Nyquist noise, therefore a mixture of Poisson and Gaussian noise. In this paper, we show convolutional neural networks (CNNs) trained on mixed Poisson and Gaussian noise images to overcome the limitations of existing image denoising methods. The trained CNN is presented as an open-source ImageJ plugin that performs real-time image denoising (within tens of milliseconds) with superior performance (SNR improvement) compared to conventional fluorescence microscopy denoising methods. The method is validated on external datasets with out-of-distribution noise, contrast, structure, and imaging modalities from the training data and consistently achieves high-performance ( > <#comment/> 8 d B ) denoising in less time than other fluorescence microscopy denoising methods.

47 OTHER INSTRUMENTATION↗

Image-Based Fracture Surface Defect Characterization Methods for Additively Manufactured Ti-6Al-4V Tested in Fatigue

Abstract Fatigue initiation in additively manufactured samples/parts often occurs at processed-induced defects such as lack-of-fusion (LoF), keyhole, or other morphological/microstructural defects that have unique characteristics and measurable qualities. Attempts at identifying and minimizing such defects have utilized optimized processing conditions along with in situ and ex situ characterization that includes metallography and/or X-ray computed tomography (XCT). This paper highlights the benefits of using fracture surface analyses to detect and quantify defects that may not be detected by metallography/XCT due to sectioning and resolution limits. In addition to using manual quantification of fatigue initiating LoF and keyhole defects on fracture surfaces, image-based machine learning using convolutional neural networks such as U-Net were also used to automate the process. Statistical analyses were used to identify the extreme cases of defects that initiated and accelerated fatigue and to model the distribution of defect size and shape characteristics to distinguish the type of defect. Initial results show agreement between trained machine learning models and ground truth data in defect segmentation, and the distributions of defect characteristics are distinguishable to particular process-induced defect types.

Materials Science↗

Current distribution monitoring enables quench and damage detection in superconducting fusion magnets

Abstract Fusion magnets made from high temperature superconducting ReBCO CORC ® cables are typically protected with quench detection systems that use voltage or temperature measurements to trigger current extraction processes. Although small coils with low inductances have been demonstrated, magnet protection remains a challenge and magnets are typically operated with little knowledge of the intrinsic performance parameters. We propose a protection framework based on current distribution monitoring in fusion cables with limited inter-cable current sharing. By employing inverse Biot-Savart techniques to distributed Hall probe arrays around CORC ® Cable-In-Conduit-Conductor (CICC) terminations, individual cable currents are recreated and used to extract the parameters of a predictive model. These parameters are shown to be of value for detecting conductor damage and defining safe magnet operating limits. The trained model is then used to predict cable current distributions in real-time, and departures between predictions and inverse Biot-Savart recreated current distributions are used to generate quench triggers. The methodology shows promise for quality control, operational planning and real-time quench detection in bundled CORC ® cables for compact fusion reactors.

47 OTHER INSTRUMENTATION↗

SU-F-J-199: Predictive Models for Cone Beam CT-Based Online Verification of Pencil Beam Scanning Proton Therapy

Purpose: To utilize online CBCT scans to develop models for predicting DVH metrics in proton therapy of head and neck tumors. Methods: Nine patients with locally advanced oropharyngeal cancer were retrospectively selected in this study. Deformable image registration was applied to the simulation CT, target volumes, and organs at risk (OARs) contours onto each weekly CBCT scan. Intensity modulated proton therapy (IMPT) treatment plans were created on the simulation CT and forward calculated onto each corrected CBCT scan. Thirty six potentially predictive metrics were extracted from each corrected CBCT. These features include minimum/maximum/mean over and under-ranges at the proximal and distal surface of PTV volumes, and geometrical and water equivalent distance between PTV and each OARs. Principal component analysis (PCA) was used to reduce the dimension of the extracted features. Three principal components were found to account for over 90% of variances in those features. Datasets from eight patients were used to train a machine learning model to fit these principal components with DVH metrics (dose to 95% and 5% of PTV, mean dose or max dose to OARs) from the forward calculated dose on each corrected CBCT. The accuracy of this model was verified on the datasets from the 9th patient. Results: The predicted changes of DVH metrics from the model were in good agreement with actual values calculated on corrected CBCT images. Median differences were within 1 Gy for most DVH metrics except for larynx and constrictor mean dose. However, a large spread of the differences was observed, indicating additional training datasets and predictive features are needed to improve the model. Conclusion: Intensity corrected CBCT scans hold the potential to be used for online verification of proton therapy and prediction of delivered dose distributions.

60 APPLIED LIFE SCIENCES↗

Knowledge-based prediction of plan quality metrics in intracranial stereotactic radiosurgery

Purpose: The objective of this work was to develop a comprehensive knowledge-based methodology for predicting achievable dose–volume histograms (DVHs) and highly precise DVH-based quality metrics (QMs) in stereotactic radiosurgery/radiotherapy (SRS/SRT) plans. Accurate QM estimation can identify suboptimal treatment plans and provide target optimization objectives to standardize and improve treatment planning. Methods: Correlating observed dose as it relates to the geometric relationship of organs-at-risk (OARs) to planning target volumes (PTVs) yields mathematical models to predict achievable DVHs. In SRS, DVH-based QMs such as brain V{sub 10Gy} (volume receiving 10 Gy or more), gradient measure (GM), and conformity index (CI) are used to evaluate plan quality. This study encompasses 223 linear accelerator-based SRS/SRT treatment plans (SRS plans) using volumetric-modulated arc therapy (VMAT), representing 95% of the institution’s VMAT radiosurgery load from the past four and a half years. Unfiltered models that use all available plans for the model training were built for each category with a stratification scheme based on target and OAR characteristics determined emergently through initial modeling process. Model predictive accuracy is measured by the mean and standard deviation of the difference between clinical and predicted QMs, δQM = QM{sub clin} − QM{sub pred}, and a coefficient of determination, R{sup 2}. For categories with a large number of plans, refined models are constructed by automatic elimination of suspected suboptimal plans from the training set. Using the refined model as a presumed achievable standard, potentially suboptimal plans are identified. Predictions of QM improvement are validated via standardized replanning of 20 suspected suboptimal plans based on dosimetric predictions. The significance of the QM improvement is evaluated using the Wilcoxon signed rank test. Results: The most accurate predictions are obtained when plans are stratified based on proximity to OARs and their PTV volume sizes. Volumes are categorized into small (V{sub PTV} < 2 cm{sup 3}), medium (2 cm{sup 3} < V{sub PTV} < 25 cm{sup 3}), and large (25 cm{sup 3} < V{sub PTV}). The unfiltered models demonstrate the ability to predict GMs to ∼1 mm and fractional brain V{sub 10Gy} to ∼25% for plans with large V{sub PTV} and critical OAR involvements. Increased accuracy and precision of QM predictions are obtained when high quality plans are selected for the model training. For the small and medium V{sub PTV} plans without critical OAR involvement, predictive ability was evaluated using the refined model. For training plans, the model predicted GM to an accuracy of 0.2 ± 0.3 mm and fractional brain V{sub 10Gy} to 0.04 ± 0.12, suggesting highly accurate predictive ability. For excluded plans, the average δGM was 1.1 mm and fractional brain V{sub 10Gy} was 0.20. These δQM are significantly greater than those of the model training plans (p < 0.001). For CI, predictions are close to clinical values and no significant difference was observed between the training and excluded plans (p = 0.19). Twenty outliers with δGM > 1.35 mm were identified as potentially suboptimal, and replanning these cases using predicted target objectives demonstrates significant improvements on QMs: on average, 1.1 mm reduction in GM (p < 0.001) and 23% reduction in brain V{sub 10Gy} (p < 0.001). After replanning, the difference of δGM distribution between the 20 replans and the refined model training plans was marginal. Conclusions: The results demonstrate the ability to predict SRS QMs precisely and to identify suboptimal plans. Furthermore, the knowledge-based DVH predictions were directly used as target optimization objectives and allowed a standardized planning process that bettered the clinically approved plans. Full clinical application of this methodology can improve consistency of SRS plan quality in a wide range of PTV volume and proximity to OARs and facilitate automated treatment planning for this critical treatment site.

60 APPLIED LIFE SCIENCES↗

Uncertainty quantification for neural network potential foundation models

Abstract For neural network potentials (NNPs) to gain widespread use, researchers must be able to trust model outputs. However, the blackbox nature of neural networks and their inherent stochasticity are often deterrents, especially for foundation models trained over broad swaths of chemical space. Uncertainty information provided at the time of prediction can help reduce aversion to NNPs. In this work, we detail two uncertainty quantification (UQ) methods. Readout ensembling, by finetuning the readout layers of an ensemble of foundation models, provides information about model uncertainty, while quantile regression, by replacing point predictions with distributional predictions, provides information about uncertainty within the underlying training data. We demonstrate our approach with the MACE-MP-0 model, applying UQ to the foundation model and a series of finetuned models. The uncertainties produced by the readout ensemble and quantile methods are demonstrated to be distinct measures by which the quality of the NNP output can be judged.

36 MATERIALS SCIENCE↗

Evaluation of a Machine-Learning Algorithm for Treatment Planning in Prostate Low-Dose-Rate Brachytherapy

Purpose: This work presents the application of a machine learning (ML) algorithm to automatically generate high-quality, prostate low-dose-rate (LDR) brachytherapy treatment plans. The ML algorithm can mimic characteristics of preoperative treatment plans deemed clinically acceptable by brachytherapists. The planning efficiency, dosimetry, and quality (as assessed by experts) of preoperative plans generated with an ML planning approach was retrospectively evaluated in this study. Methods and Materials: Preimplantation and postimplantation treatment plans were extracted from 100 high-quality LDR treatments and stored within a training database. The ML training algorithm matches similar features from a new LDR case to those within the training database to rapidly obtain an initial seed distribution; plans were then further fine-tuned using stochastic optimization. Preimplantation treatment plans generated by the ML algorithm were compared with brachytherapist (BT) treatment plans in terms of planning time (Wilcoxon rank sum, α = 0.05) and dosimetry (1-way analysis of variance, α = 0.05). Qualitative preimplantation plan quality was evaluated by expert LDR radiation oncologists using a Likert scale questionnaire. Results: The average planning time for the ML approach was 0.84 ± 0.57 minutes, compared with 17.88 ± 8.76 minutes for the expert planner (P=.020). Preimplantation plans were dosimetrically equivalent to the BT plans; the average prostate V150% was 4% lower for ML plans (P=.002), although the difference was not clinically significant. Respondents ranked the ML-generated plans as equivalent to expert BT treatment plans in terms of target coverage, normal tissue avoidance, implant confidence, and the need for plan modifications. Respondents had difficulty differentiating between plans generated by a human or those generated by the ML algorithm. Conclusions: Prostate LDR preimplantation treatment plans that have equivalent quality to plans created by brachytherapists can be rapidly generated using ML. The adoption of ML in the brachytherapy workflow is expected to improve LDR treatment plan uniformity while reducing planning time and resources.

62 RADIOLOGY AND NUCLEAR MEDICINE↗