Sitemap
A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.
Pages
Posts
Future Blog Post
Published:
This post will show up by default. To disable scheduling of future posts, edit config.yml
and set future: false
.
Blog Post number 4
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 3
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 2
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 1
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
portfolio
Portfolio 1
State of health (SOH) is a crucial indicator to evaluate the level of degradation of batteries that cannot be measured directly but requires estimation. Accurate SOH estimation enhances detection, control, and feedback for Li-ion batteries, allowing for safe and efficient energy management and guiding the development of new-generation batteries. Despite significant progress in data-driven SOH estimation methods, the time- and resource-consuming degradation experiments to generate lifelong training data pose a barrier to timely safety monitoring and the development of new battery technologies. To tackle this problem, We propose GPT4Battery, the first foundation model for battery to utilize the strong generalization ability of the large language model (LLM) for cross-battery SOH estimation. We design a translator to reprogram the voltage-time battery charging data into text prototype representations to align data from different modalities, adapting the LLM to battery tasks. We also design a novel physical guided test-time prompt tuning (PGTPT) method to learn adaptive prompts on the fly with a single test sample, enhancing the model’s generalization ability and fitting the real-world scenarios. PGTPT optimizes the prompt by guiding the LLM to generate a complete battery charging curve in accordance with the physics equations of the 1-RC ECM model of a LIB cell. The validation results demonstrate that the proposed model achieves state-of-the-art accuracy that is even on par with domain adaptation or fine-tuning methods that require additional training data on five widely recognized datasets collected from 65 batteries.
Portfolio item number 2
Short description of portfolio item number 2
publications
GPT4Battery: Cross-battery State of Health Estimation via Physical-Guided Test-time Prompt Learning with LLM
Published in Arxiv, Under Review, 2024
State of health (SOH) is a crucial indicator to evaluate the level of degradation of batteries that cannot be measured directly but requires estimation. Accurate SOH estimation enhances detection, control, and feedback for Li-ion batteries, allowing for safe and efficient energy management and guiding the development of new-generation batteries. Despite significant progress in data-driven SOH estimation methods, the time- and resource-consuming degradation experiments to generate lifelong training data pose a barrier to timely safety monitoring and the development of new battery technologies. To tackle this problem, We propose GPT4Battery, the first foundation model for battery to utilize the strong generalization ability of the large language model (LLM) for cross-battery SOH estimation. We design a translator to reprogram the voltage-time battery charging data into text prototype representations to align data from different modalities, adapting the LLM to battery tasks. We also design a novel physical guided test-time prompt tuning (PGTPT) method to learn adaptive prompts on the fly with a single test sample, enhancing the model generalization ability and fitting the real-world scenarios. PGTPT optimizes the prompt by guiding the LLM to generate a complete battery charging curve in accordance with the physics equations of the 1-RC ECM model of a LIB cell. The validation results demonstrate that the proposed model achieves state-of-the-art accuracy that is even on par with domain adaptation or fine-tuning methods that require additional training data on five widely recognized datasets collected from 65 batteries.
Recommended citation: To be announced. https://github.com/yuyuan6/yuyuan6.github.io/raw/master/files/gpt4battery.pdf
talks
Recent Advances of LLM for Time Series
Published:
This tutorial gives recent advances in the combination of large language models and time series analysis, involving two top conference papers and code reading. Download the slides here.
Explainable Time Series based on Feature Importance
Published:
This talk introduces explainable AI and collects recent publications about interpretable time series based on feature importance. Download the slides here.
A large File Downloading for Lab Assignments of Computer Network
Published:
This is a lab assignment to use Wireshark and capture packets while downloading an large file. Download the Video here.
teaching
Teaching experience 1
Undergraduate course, University 1, Department, 2014
This is a description of a teaching experience. You can use markdown like any other post.
Teaching experience 2
Workshop, University 1, Department, 2015
This is a description of a teaching experience. You can use markdown like any other post.