Round 10 Train DatasetThis is the training data used to create and evaluate trojan detection software solutions. This data, generated at NIST, consists of object detection AIs trained on the COCO dataset. 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 144 AI models using a small set of model architectures.

This repository contains scripts and supporting information for a set of demonstrations related to the 2022 AM-Bench challenge series.In particular, the data used in this demonstration comes from the 2018 AM-Bench challenge documented at https://www.nist.gov/ambench/amb2018-02-description.The script queries the AM-Bench 2018 repository located at https://ambench.nist.gov/, then processes the returned XML-based results, does some image processing, and generates plots of melt pool dep

Abstract (from the manuscript): High cycle fatigue life of laser-powder bed fusion (L-PBF) parts depends on several factors; as-built surfaces, when present, are a particular concern. This work measures as-built L-PBF surfaces with X-ray computed tomography, and uses rotating beam fatigue (RBF) testing to measure high cycle fatigue life. Surfaces with different, but consistent, characteristics are achieved by build in vertical specimens and changing only the number of contour passes.

The development of large residual elastic strains and stresses during laser powder-bed fusion (LPBF) additive manufacturing is one of the most significant barriers to widespread adoption. Accurate modeling of these strains and stresses is broadly recognized as an effective tool for mitigating these challenges, but rigorous validation data are needed. This data publication includes measurement data from diffraction-based characterizations of residual elastic strains in as-built (not heat treated) artifacts manufactured as part the the 2018 Additive Manufacturing Benchmark Series (AM Bench).

Round rl-lavaworld-jul2023-train Train DatasetThis is the training data used to create and evaluate trojan detection software solutions. This data, generated at NIST, consists of Reinforcement Learning agents trained to navigate the Lavaworld Minigrid environment. 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.

nlp-question-answering-aug2023-trainThis is the train data used to 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.

Round rl-randomized-lavaworld-aug2023-train Train DatasetThis is the training data used to create and evaluate trojan detection software solutions. This data, generated at NIST, consists of Reinforcement Learning agents trained to navigate the Lavaworld Minigrid environment. 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 data repository provides a central location for a body of work using one build of nickel-based alloy 718 (IN718) material and resulted in three different studies. The IN718 parts were manufactured by laser powder bed fusion using a range of laser energy densities (manipulation of processing variables) and orientations with respect to the build direction. The influence of processing variables on resulting grain structures, pore structures, and mechanical properties were studied in the as-built (not heat treated) condition.

Round 5 Test DatasetThis is the test 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 text sentiment classification 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 5 Holdout DatasetThis is the holdout 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 text sentiment classification 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.