Search NASA⌕ Search

SEARCH · Search NASA

Results for “Decoder Input”

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 55 records · Page 3

A comparison of model-based VQ compression with other VQ approaches

In our previous work on Model-Based Vector Quantization (MVQ), we presented some performance comparisons (both rate distortion and decompression time) with VQ and JPEG/DCT. In this paper, we compare the MVQ's rate distortion performance with Mean Removed Vector Quantization (MRVQ) and include our previous comparison with VQ. MVQ is similar to MRVQ in many ways. Both of these techniques extract means of the vectors (raster-scanned image blocks) and reduce them to mean removed residuals by subtracting block means from the elements of the vectors. In the case of MRVQ, a codebook of residual vectors is generated using a training set. For every vector from the input image, the block mean and address of the codevector from the codebook that matches the input vector closest are transmitted to the decoder. The codebook is generated using generalized Lloyd algorithm on training set of residual vectors. For MVQ the pairs consist of vector means and address of the closest matching vector from codebook generated by models based on statistical properties of the residuals and Human Visual System (HVS). In our experiments, we found that MVQ performance in rate distortion sense is almost always better than VQ and is comparable to MRVQ. Further, MVQ is much easier to use than either VQ or MRVQ, since the training and managing of explicit codebooks is not required.

Manohar, Mareboyana↗

A high speed sequential decoder

The performance and theory of operation for the High Speed Hard Decision Sequential Decoder are delineated. The decoder is a forward error correction system which is capable of accepting data from binary-phase-shift-keyed and quadriphase-shift-keyed modems at input data rates up to 30 megabits per second. Test results show that the decoder is capable of maintaining a composite error rate of 0.00001 at an input E sub b/N sub o of 5.6 db. This performance has been obtained with minimum circuit complexity.

Lum, H., Jr.↗

A Variational Autoencoder Model Toward Molecular Structure Representation Learning of Fuels

Here, in this work, a Variational Autoencoder (VAE)-based data-driven modeling framework is developed with the overarching goal of enabling fuel design. The VAE model is trained on a large dataset with several chemical species to learn a compressed latent space molecular representation. Chemical structure in the form of Simplified Molecular Input Line Entry System (SMILES) string is fed as input, encoded into the VAE latent space, and decoded back to the SMILES string using Long Short-Term Memory (LSTM) networks. Complexities of the VAE training loss function are thoroughly examined by varying the weightage (beta (𝜷) parameter) of the latent space regularization term, thereby assessing the balance between reconstruction accuracy and validity, and focusing on both accurate molecular structure reconstruction and latent space consistency. Two different strategies for 𝜷 variation are evaluated: linear annealing and cyclic annealing. In addition, the impact of total correlation adjustment and hierarchical priors is also studied with regard to the balance between reconstruction fidelity and latent space regularization, and potential issues such as posterior collapse, over-regularization, and poor disentanglement of latent variables. Overall, the best performance of the model is achieved with hierarchical priors and incrementally increasing 𝜷 from 0 to a threshold value of 0.25 over 75 epochs. The generative VAE model can be readily coupled with Quantitative Structure–Property Relationship (QSPR) analysis to develop an integrated end-to-end framework for fuel-property prediction and molecular design of novel promising fuels.

fuel design↗

Burst decoding of binary block codes on Q-ary output channels.

The burst-b distance between two binary vectors is defined and shown to be a metric. This definition is applied to a binary-input, Q-ary output channel where errors occur in bursts. A decoding algorithm is presented for such a channel that is an extension of Weldon's (1971) weighted erasure decoding. Examples are presented illustrating the techniques.

Wainberg, S.↗

Reed-Solomon decoder

A Reed-Solomon decoder with dedicated hardware for five sequential algorithms was designed with overall pipelining by memory swapping between input, processing and output memories, and internal pipelining through the five algorithms. The code definition used in decoding is specified by a keyword received with each block of data so that a number of different code formats may be decoded by the same hardware.

Lahmeyer, Charles R.↗

Robust Mapping of Incoherent Fiber-Optic Bundles

A method and apparatus for mapping between the positions of fibers at opposite ends of incoherent fiber-optic bundles have been invented to enable the use of such bundles to transmit images in visible or infrared light. The method is robust in the sense that it provides useful mapping even for a bundle that contains thousands of narrow, irregularly packed fibers, some of which may be defective. In a coherent fiber-optic bundle, the input and output ends of each fiber lie at identical positions in the input and output planes; therefore, the bundle can be used to transmit images without further modification. Unfortunately, the fabrication of coherent fiber-optic bundles is too labor-intensive and expensive for many applications. An incoherent fiber-optic bundle can be fabricated more easily and at lower cost, but it produces a scrambled image because the position of the end of each fiber in the input plane is generally different from the end of the same fiber in the output plane. However, the image transmitted by an incoherent fiber-optic bundle can be unscrambled (or, from a different perspective, decoded) by digital processing of the output image if the mapping between the input and output fiber-end positions is known. Thus, the present invention enables the use of relatively inexpensive fiber-optic bundles to transmit images.

Roberts, Harry E.↗

Fourier-DeepONet: Fourier-enhanced deep operator networks for full waveform inversion with improved accuracy, generalizability, and robustness

In this article, full waveform inversion (FWI) infers the subsurface structure information from seismic waveform data by solving a non-convex optimization problem. Data-driven FWI has been increasingly studied with various neural network architectures to improve accuracy and computational efficiency. Nevertheless, the applicability of pre-trained neural networks is severely restricted by potential discrepancies between the source function used in the field survey and the one utilized during training. Here, we develop a Fourier-enhanced deep operator network (Fourier-DeepONet) for FWI with the generalization of seismic sources, including the frequencies and locations of sources. Specifically, we employ the Fourier neural operator as the decoder of DeepONet, and we utilize source parameters as one input of Fourier-DeepONet, facilitating the resolution of FWI with variable sources. To test Fourier-DeepONet, we develop three new and realistic FWI benchmark datasets (FWI-F, FWI-L, and FWI-FL) with varying source frequencies, locations, or both. Our experiments demonstrate that compared with existing data-driven FWI methods, Fourier-DeepONet obtains more accurate predictions of subsurface structures in a wide range of source parameters. Moreover, the proposed Fourier-DeepONet exhibits superior robustness when handling data with Gaussian noise or missing traces and sources with Gaussian noise, paving the way for more reliable and accurate subsurface imaging across diverse real conditions.

42 ENGINEERING↗

LUNA Condition-Based Monitoring Update: Auto-encoders on Actuator Data [Slides]

An auto-encoder is a neural network trained to output a reconstruction of its input. There is a ‘bottleneck’ layer whose dimension is typically lower than that of the input. Because the input passes through the bottleneck, the auto-encoder must learn to compress (encode) the input into this low-dimensional form, and to decompress (decode) the compressed data into a close approximation of its original form.

97 MATHEMATICS AND COMPUTING↗

Weight distributions for turbo codes using random and nonrandom permutations

This article takes a preliminary look at the weight distributions achievable for turbo codes using random, nonrandom, and semirandom permutations. Due to the recursiveness of the encoders, it is important to distinguish between self-terminating and non-self-terminating input sequences. The non-self-terminating sequences have little effect on decoder performance, because they accumulate high encoded weight until they are artificially terminated at the end of the block. From probabilistic arguments based on selecting the permutations randomly, it is concluded that the self-terminating weight-2 data sequences are the most important consideration in the design of constituent codes; higher-weight self-terminating sequences have successively decreasing importance. Also, increasing the number of codes and, correspondingly, the number of permutations makes it more and more likely that the bad input sequences will be broken up by one or more of the permuters. It is possible to design nonrandom permutations that ensure that the minimum distance due to weight-2 input sequences grows roughly as the square root of (2N), where N is the block length. However, these nonrandom permutations amplify the bad effects of higher-weight inputs, and as a result they are inferior in performance to randomly selected permutations. But there are 'semirandom' permutations that perform nearly as well as the designed nonrandom permutations with respect to weight-2 input sequences and are not as susceptible to being foiled by higher-weight inputs.

Dolinar, S.↗

Multi-task Parallelism for Robust Pre-training of Graph Foundation Models on Multi-source, Multi-fidelity Atomistic Modeling Data

Graph foundation models using graph neural networks promise sustainable, efficient atomistic modeling. To tackle challenges of processing multi-source, multi-fidelity data during pre-training, recent studies employ multi-task learning, in which shared message passing layers initially process input atomistic structures regardless of source, then route them to multiple decoding heads that predict data-specific outputs. This approach stabilizes pre-training and enhances a model’s transferability to unexplored chemical regions. Preliminary results on approximately four million structures are encouraging, yet questions remain about generalizability to larger, more diverse datasets and scalability on supercomputers. We propose a multi-task parallelism method that distributes each head across computing resources with GPU acceleration. Implemented in the open-source HydraGNN architecture, our method was trained on over 24 million structures from five datasets and tested on the Perlmutter, Aurora, and Frontier supercomputers, demonstrating efficient scaling on all three highly heterogeneous super-computing architectures.

Lupo Pasini, Massimiliano [ORNL] (ORCID:0000000249↗

NuGraph2 with context-aware inputs: physics-inspired improvements in semantic segmentation

Graph neural networks have recently shown strong promise for event reconstruction tasks in Liquid Argon Time Projection Chambers, yet their performance remains limited for underrepresented classes of particles, such as Michel electrons. In this work, we investigate physics-informed strategies to improve semantic segmentation within the NuGraph2 architecture. We explore three complementary approaches: (i) enriching the input representation with context-aware features derived from detector geometry and track continuity, (ii) introducing auxiliary decoders to capture class-level correlations, and (iii) incorporating energy-based regularization terms motivated by Michel electron energy distributions. Experiments on MicroBooNE public datasets show that physics-inspired feature augmentation yields the largest gains, particularly boosting Michel electron precision and recall by disentangling overlapping latent space regions. In contrast, auxiliary decoders and energy-regularization terms provided limited improvements, partly due to the hit-level nature of NuGraph2, which lacks explicit particle- or event-level representations. Our findings highlight that embedding physics context directly into node-level inputs is more effective than imposing task-specific auxiliary losses, and suggest that future hierarchical architectures such as NuGraph3, with explicit particle- and event-level reasoning, will provide a more natural setting for advanced decoders and physics-based regularization. The code for this work is publicly available on Github at https://github.com/vitorgrizzi/nugraph_phys/tree/main_phys.

Other Experiments↗

A system for the input and storage of data in the Besm-6 digital computer

Computer programs used for the decoding and storage of large volumes of data on the the BESM-6 computer are described. The following factors are discussed: the programming control language allows the programs to be run as part of a modular programming system used in data processing; data control is executed in a hierarchically built file on magnetic tape with sequential index storage; and the programs are not dependent on the structure of the data.

Schmidt, K.↗

Robust deep learning framework for constitutive relations modeling

Modeling the full-range deformation behaviors of materials under complex loading and materials conditions is a significant challenge for constitutive relations (CRs) modeling. Here, we propose a general encoder-decoder deep learning framework that can model high-dimensional stress-strain data and complex loading histories with robustness and universal capability. The framework employs an encoder to project high-dimensional input information (e.g., loading history, loading conditions, and materials information) to a lower-dimensional hidden space and a decoder to map the hidden representation to the stress of interest. We evaluated various encoder architectures, including gated recurrent unit (GRU), GRU with attention, temporal convolutional network (TCN), and the Transformer encoder, on two complex stress-strain datasets that were designed to include a wide range of complex loading histories and loading conditions. All architectures achieved excellent test results with an root-mean-square error (RMSE) below 1 MPa. Additionally, we analyzed the capability of the different architectures to make predictions on out-of-domain applications, with an uncertainty estimation based on deep ensembles. The proposed approach provides a robust alternative to empirical/semi-empirical models for CRs modeling, offering the potential for more accurate and efficient materials design and optimization.

36 MATERIALS SCIENCE↗

Three-phase dc motor decoder

Circuit minimizes components required to provide six properly timed drive signals from three equal-interval sensor inputs.

Studer, P. A.↗

Error-Rate Bounds for Coded PPM on a Poisson Channel

Equations for computing tight bounds on error rates for coded pulse-position modulation (PPM) on a Poisson channel at high signal-to-noise ratio have been derived. These equations and elements of the underlying theory are expected to be especially useful in designing codes for PPM optical communication systems. The equations and the underlying theory apply, more specifically, to a case in which a) At the transmitter, a linear outer code is concatenated with an inner code that includes an accumulator and a bit-to-PPM-symbol mapping (see figure) [this concatenation is known in the art as "accumulate-PPM" (abbreviated "APPM")]; b) The transmitted signal propagates on a memoryless binary-input Poisson channel; and c) At the receiver, near-maximum-likelihood (ML) decoding is effected through an iterative process. Such a coding/modulation/decoding scheme is a variation on the concept of turbo codes, which have complex structures, such that an exact analytical expression for the performance of a particular code is intractable. However, techniques for accurately estimating the performances of turbo codes have been developed. The performance of a typical turbo code includes (1) a "waterfall" region consisting of a steep decrease of error rate with increasing signal-to-noise ratio (SNR) at low to moderate SNR, and (2) an "error floor" region with a less steep decrease of error rate with increasing SNR at moderate to high SNR. The techniques used heretofore for estimating performance in the waterfall region have differed from those used for estimating performance in the error-floor region. For coded PPM, prior to the present derivations, equations for accurate prediction of the performance of coded PPM at high SNR did not exist, so that it was necessary to resort to time-consuming simulations in order to make such predictions. The present derivation makes it unnecessary to perform such time-consuming simulations.

Moision, Bruce↗

Single phase dynamic CMOS PLA using charge sharing technique

A single phase dynamic CMOS NOR-NOR programmable logic array (PLA) using triggered decoders and charge sharing techniques for high speed and low power is presented. By using the triggered decoder technique, the ground switches are eliminated, thereby, making this new design much faster and lower power dissipation than conventional PLA's. By using the charge-sharing technique in a dynamic CMOS NOR structure, a cascading AND gate can be implemented. The proposed PLA's are presented with a delay-time of 15.95 and 18.05 nsec, respectively, which compare with a conventional single phase PLA with 35.5 nsec delay-time. For a typical example of PLA like the Signetics 82S100 with 16 inputs, 48 input minterms (m) and 8 output minterms (n), the 2-SOP PLA using the triggered 2-bit decoder is 2.23 times faster and has 2.1 times less power dissipation than the conventional PLA. These results are simulated using maximum drain current of 600 micro-A, gate length of 2.0 micron, V sub DD of 5 V, the capacitance of an input miniterm of 1600 fF, and the capacitance of an output minterm of 1500 fF.

Dhong, Y. B.↗