This research software package contains Python code to execute experiments on deep generative modeling of classical random process models for noise time series. Specifically, it includes Pytorch implementations of two generative adversarial network (GAN) models for time series based on convolutational neural networks (CNNs): WaveGAN, a 1-D CNN model, and STFT-GAN, a 2-D CNN model. In addition, there are methods for generating and evaluating noise time series defined several by classical random process models.

SolDet is an object-oriented package for solitonic feature detection in absorption images of Bose-Einstein condensate. with wider use for cold atom image analysis. Featured with classifier, object detector, and Mexican hat metric methods. Technical details are explained in https://arxiv.org/abs/2111.04881.

A code repository and accompanying data for incorporating imperfect theory into machine learning for improved prediction and explainability. Specifically, it focuses on the case study of the dimensions of a polymer chain in different solvent qualities. Jupyter Notebooks for quickly testing concepts and reproducing figures, as well as source code that computes the mean squared error as a function of dataset size for various machine learning models are included.For additional details on the data, please refer to the README.md associated with the data.

Atomic Bose-Einstein condensates (BECs) are widely investigated systems that exhibit quantum phenomena on a macroscopic scale. For example, they can be manipulated to contain solitonic excitations including conventional solitons, vortices, and many more. Broadly speaking, solitonic excitations are solitary waves that retain their size and shape and often propagate at a constant speed. They are present in many systems, at scales ranging from microscopic, to terrestrial and even astronomical.

This project produces synthetic datasets of spectrum sharing simulation results (I/Q data, metadata, and KPIs).

Using a modified Thomas-Fermi approximation, we model a reference semiconductor system comprising a quasi-1D nanowire with a series of five depletion gates whose voltages determine the number of quantum dots (QDs), the charges on each of the QDs, as well as the conductance through the wire. The original dataset, QFlow lite, consists of 1 001 idealized simulated measurements with gate configurations sampling over different realizations of the same type of device.

Round 9 Train DatasetThis is the training data used to construct and evaluate trojan detection software solutions. This data, generated at NIST, consists of natural language processing (NLP) AIs trained to perform one of three tasks, sentiment classification, named entity recognition, or extractive question answering on English text. A known percentage of these trained AI models have been poisoned with a known trigger which induces incorrect behavior. This data will be used to develop software solutions for detecting which trained AI models have been poisoned via embedded triggers.

This repository contains results for experiments on generative modeling of synthetic Orthogonal-Frequency Division Multiplexing (OFDM) communication signals. (This record supersedes Software and Data for Modeling OFDM Communication Signals with Generative Adversarial Networks, formerly at https://doi.org/10.18434/mds2-2428)

ANDiE the Autonomous Neutron Diffraction Explorer is a tool for autonomously discovering the magnetic transition temperature and transition dynamics of a material from neutron diffraction experiments. The Jupyter notebooks used to implement ANDiE can be found here: https://github.com/usnistgov/ANDiE-v1_0 The Jupyter notebooks contained therein are of ANDiE as implemented at the WAND2 instrument at the HB-2C beamline at the High-Flux Isotope Reactor (HFIR) at Oak Ridge National Laboratory (ORNL).

Round 8 Test DatasetThis is the training data used to construct and evaluate trojan detection software solutions. This data, generated at NIST, consists of natural language processing (NLP) AIs trained to perform extractive question answering. A known percentage of these trained AI models have been poisoned with a known trigger which induces incorrect behavior. This data will be used to develop software solutions for detecting which trained AI models have been poisoned via embedded triggers. This dataset consists of 360 QA AI models using a small set of model architectures.