Round 8 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 extractive question answering (QA 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.

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.

Round 8 Holdout 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 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 dataset contains mechanical testing data at two different loading rates for two different aging conditions (wet and dry at 70 °C) for five different extraction time points for ultra high molar mass polyethylene (UHMMPE) unidirectional (UD) laminate. In each folder there is a spreadsheet (.csv) with the test data collected during loading, and a (.avi) video file for each of the tests done under those conditions.

The DVU dataset consists of 10 creative common movies for training and 4 creative common movies for testing used in the Deep Video Understanding grand challenges in 2020 and 2021. Total duration is about 17 hr.

A compilation of single fiber tensile tests for 6 different materials at 8 different gauge lengths. Fibers included are 3 different aramids, a liquid crystal polyester, regenerated cellulose, and nylon 6,6.

Equal Employment Opportunity Tabulation (5-year ACS data)

This dataset contains the raw data for the paper: Garey, W.; Henderson, T.; Sun, Y.; Rouil, R. and Gamboa, S. (2021). Modeling MCPTT and User Behavior in ns-3. In Proceedings of the 11th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - SIMULTECH, ISBN 978-989-758-528-9, pages 30-41. DOI: 10.5220/0010513200300041. This includes: Figure 4 - The talk spurt and talk session duration CDF. Figure 5 - Orchestrator pusher model example. Figure 7 - MCPTT Access Time for immediately granted requests with queuing disabled.

Flow cytometry measurements of FISH-labeled RNA at timepoints following rifampicin addition. These data folders contain raw data (.fcs files), analysis files, and summary statistics ("summary.csv") for each experiment. The "v1" data represent biological replicates. The "v2" data represent technical replicates of the original sample preparations for each experiment. Both "v2" datasets were collected on July 16, 2021.

This dataset presents statistics on: the number of establishments; sales, value of shipments, or revenue; annual payroll; and number of employees whose NAICS classification has changed between the current and the previous economic censuses. Data are shown for 6-digit current economic census NAICS industries and their 8-digit previous economic census NAICS components for the U.S. Includes only establishments of firms with paid employees.