Summary of Can We Rely on Llm Agents to Draft Long-horizon Plans? Let’s Take Travelplanner As An Example, by Yanan Chen et al.
Can We Rely on LLM Agents to Draft Long-Horizon Plans? Let’s Take TravelPlanner as an Example
by Yanan Chen, Ali Pesaranghader, Tanmana Sadhu, Dong Hoon Yi
First submitted to arxiv on: 12 Aug 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: 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 This study examines the capabilities of large language models (LLMs) in autonomous agents, aiming to bridge the gap between promising generalization and emergent capabilities. The researchers leveraged the TravelPlanner benchmark, a realistic scenario for real-world planning tasks, to investigate four key research questions: robustness in lengthy and noisy contexts, few-shot prompting’s impact on performance, refinement’s potential, and fine-tuning with positive and negative feedback. The results show that LLMs often fail to attend to crucial parts of long context, struggle with analyzing long plans, but can be improved through Feedback-Aware Fine-Tuning (FAFT), outperforming Supervised Fine-Tuning (SFT). |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at how big language models work in machines that make decisions on their own. The researchers used a special test called TravelPlanner to see if these models could plan well in real-life situations. They asked four big questions: can the model handle long and messy information? Does it get better with a little help from examples? Can it fix mistakes and make better plans? And does getting feedback – both good and bad – help the model improve? The answers show that these models are not perfect, but they can be made better by using special training methods. |
Keywords
» Artificial intelligence » Few shot » Fine tuning » Generalization » Prompting » Supervised
