Summary of Generating Visual Stories with Grounded and Coreferent Characters, by Danyang Liu et al.
Generating Visual Stories with Grounded and Coreferent Characters
by Danyang Liu, Mirella Lapata, Frank Keller
First submitted to arxiv on: 20 Sep 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
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 paper introduces a novel task, character-centric story generation, which focuses on predicting visual stories with grounded and coreferent character mentions. The authors present the first model capable of achieving this goal, finetuned on a new dataset built on top of the VIST benchmark. The dataset is enriched with visual and textual character coreference chains, allowing for the evaluation of character richness and coreference in generated stories. Experimental results demonstrate that the proposed model generates stories with recurring characters that are consistent and coreferent to a greater extent than baselines and state-of-the-art systems. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper creates new ways to tell stories that have real people in them. Right now, computers can make up stories based on events, but they don’t create characters that feel like they’re part of the story. The authors want to change this by making a computer model that can predict stories with specific characters. They created a special dataset for training the model and came up with new ways to measure how well it does. The results show that their model makes better stories than others do, with characters that appear again and are connected in meaningful ways. |
Keywords
» Artificial intelligence » Coreference
