Large language models (LLMs) have been recently leveraged as training data generators for various natural language processing (NLP) tasks. While previous research has explored different approaches to training models using generated data, they generally rely on simple class-conditional prompts, which may limit the diversity of the generated data and inherit systematic biases of LLM. Thus, we investigate training data generation with diversely attributed prompts (e.g.,

Large language models (LLMs) have been recently leveraged as training data generators for various natural language processing (NLP) tasks. While previous research has explored different approaches to training models using generated data, they generally rely on simple class-conditional prompts, which may limit the diversity of the generated data and inherit systematic biases of LLM. Thus, we investigate training data generation with diversely attributed prompts (e.g.,

Carnegie Mellon University students who designed and built a small, boxy robot, called Iris, have achieved a major milestone: their robot passed its critical design review by NASA and is on track to land on the moon in the fall of 2021. "We are moving forward … we're going to the moon," a triumphant project manager, Raewyn Duvall, told Iris team members during a Zoom meeting following the review. Officials at NASA and Astrobotic Inc., whose Peregrine lander will deliver the robot to the lunar surface, performed the review. Duvall, a Ph.D. student in the Electrical and Computer Engineering Department, said the process resulted in a few small design revisions, which the team is now incorporating. The team will replace prototype parts with flight components this summer, as they test the robot to prove that it can withstand the trip to the moon without causing problems for Peregrine or other payloads aboard the lunar lander.