Summary of Infer Human’s Intentions Before Following Natural Language Instructions, by Yanming Wan et al.
Infer Human’s Intentions Before Following Natural Language Instructions
by Yanming Wan, Yue Wu, Yiping Wang, Jiayuan Mao, Natasha Jaques
First submitted to arxiv on: 26 Sep 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL); 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 paper proposes a new framework, Follow Instructions with Social and Embodied Reasoning (FISER), to improve AI agents’ ability to follow natural language instructions in everyday cooperative tasks. FISER addresses ambiguities inherent in human language by explicitly inferring human goals and intentions as intermediate reasoning steps. The framework is implemented using Transformer-based models and evaluated on the HandMeThat benchmark, outperforming purely end-to-end approaches. By incorporating social reasoning, FISER achieves better performance compared to strong baselines, including Chain of Thought prompting on pre-trained language models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine having a helpful AI assistant that can understand what you want it to do. This paper tries to make that happen by figuring out how humans think and what they really mean when they give instructions. Right now, AI assistants are good at following simple commands, but they struggle with more complicated tasks that require understanding human intentions. The new framework proposed in this paper gets better results by thinking about what the human wants to achieve before deciding what to do next. It’s like having a conversation with someone and understanding their goals and plans. |
Keywords
» Artificial intelligence » Prompting » Transformer
