Plug-and-Blend: A Framework for Controllable Story Generation with Blended Control Codes
arXiv:2104.04039
Abstract
Large pre-trained neural language models (LM) have very powerful text generation capabilities. However, in practice, they are hard to control for creative purposes. We describe a Plug-and-Play controllable language generation framework, Plug-and-Blend, that allows a human user to input multiple control codes (topics). In the context of automated story generation, this allows a human user loose or fine-grained control of the topics and transitions between them that will appear in the generated story, and can even allow for overlapping, blended topics. Automated evaluations show our framework, working with different generative LMs, controls the generation towards given continuous-weighted control codes while keeping the generated sentences fluent, demonstrating strong blending capability. A human participant evaluation shows that the generated stories are observably transitioning between two topics.
arXiv comment8 pages, To appear at AIIDE 2021
References in corpus (14)
- CTRL: A Conditional Transformer Language Model for Controllable Generation
- Prefix-Tuning: Optimizing Continuous Prompts for Generation
- Toward Controlled Generation of Text
- Plug and Play Language Models: A Simple Approach to Controlled Text Generation
- A Corpus and Evaluation Framework for Deeper Understanding of Commonsense Stories
- PlotMachines: Outline-Conditioned Generation with Dynamic Plot State Tracking
- Transformer-based Conditional Variational Autoencoder for Controllable Story Generation
- A Distributional Approach to Controlled Text Generation
- Narrative Interpolation for Generating and Understanding Stories
- CoCon: A Self-Supervised Approach for Controlled Text Generation
- DExperts: Decoding-Time Controlled Text Generation with Experts and Anti-Experts
- Plug and Play Autoencoders for Conditional Text Generation
- UNION: An Unreferenced Metric for Evaluating Open-ended Story Generation
- Directed Beam Search: Plug-and-Play Lexically Constrained Language Generation