Summary of Hargpt: Are Llms Zero-shot Human Activity Recognizers?, by Sijie Ji et al.
HARGPT: Are LLMs Zero-Shot Human Activity Recognizers?
by Sijie Ji, Xinzhe Zheng, Chenshu Wu
First submitted to arxiv on: 5 Mar 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
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 paper explores the potential of Large Language Models (LLMs) as foundational models for interpreting the physical world. The study, HARGPT, investigates whether LLMs can recognize human activities from raw IMU data without prior training or fine-tuning, a task known as zero-shot human activity recognition (HAR). The authors demonstrate that LLMs can comprehend raw IMU data and perform HAR tasks in a zero-shot manner by utilizing role-play and think step-by-step strategies for prompting. The paper benchmarks HARGPT on GPT4 using two public datasets with different inter-class similarities, comparing various baselines based on traditional machine learning and state-of-the-art deep classification models. Surprisingly, LLMs successfully recognize human activities from raw IMU data and consistently outperform all the baselines on both datasets. This study’s findings indicate that effective prompting can enable LLMs to interpret raw sensor data of the physical world effectively. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at how well Large Language Models (LLMs) can understand human activities from simple sensors like accelerometers and gyroscopes. The researchers tested if these models could learn to recognize different human actions, like walking or running, without being trained on any specific data beforehand. They found that by using clever prompts, the LLMs were able to successfully identify human activities from the raw sensor data. This is important because it shows that these models can be used to analyze data from all sorts of sensors, not just language-based ones. |
Keywords
» Artificial intelligence » Activity recognition » Classification » Fine tuning » Machine learning » Prompting » Zero shot
