Summary of Genie: Achieving Human Parity in Content-grounded Datasets Generation, by Asaf Yehudai et al.
Genie: Achieving Human Parity in Content-Grounded Datasets Generation
by Asaf Yehudai, Boaz Carmeli, Yosi Mass, Ofir Arviv, Nathaniel Mills, Assaf Toledo, Eyal Shnarch, Leshem Choshen
First submitted to arxiv on: 25 Jan 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed Genie method addresses the lack of high-quality data for content-grounded generation tasks by automatically generating synthetic data through three stages: Content Preparation, Generation, and Filtering. The method is showcased by creating large-scale datasets for Long-Form Question-Answering (LFQA), summarization, and information extraction. Human evaluation confirms that the generated data is natural and of high quality. Comparisons with models trained on human-written data show Genie-trained models to be competitive or superior in terms of accuracy and faithfulness. The method is applied to create LFQA data within the medical domain, demonstrating its potential for domain-specific applications. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Genie is a new way to make high-quality data for tasks like answering long questions, summarizing texts, and finding key information. This method helps by automatically creating lots of examples that are natural and accurate. The team tested this approach and found that it worked really well. They compared their results with models trained on human-written data and found that Genie-trained models were just as good or better. They even applied the method to a medical domain, showing how it can be used for specific areas of interest. |
Keywords
* Artificial intelligence * Question answering * Summarization * Synthetic data
