Haoxin Liu (@hessianliu) 's Twitter Profile
Haoxin Liu

@hessianliu

PhD Student in Computational Science & Engineering @GTCSE @GeorgiaTech |
Working on ML/AI for Time Series Analysis and Applications

ID: 1782519396944232451

linkhttps://hliu763.github.io/ calendar_today22-04-2024 21:18:46

8 Tweet

32 Followers

75 Following

Harsha Kamarthi (@harsha_64) 's Twitter Profile Photo

Can we build foundational models for time series similar to text and vision that can be deployed on a wide range of domains with superior accuracy, data efficiency, and scalability? We present two recent works with B. Aditya Prakash that make significant progress to achieve this goal. 1/

Zhiyuan (Leo) Zhao (@leozhao_zhiyuan) 's Twitter Profile Photo

Catch me at my poster session at #ICLR2024 : What: PINNsFormer: A Transformer-Based Framework For Physics-Informed Neural Networks When: May 10:45 am — 12:45 pm CEST Where: Halle B #31, Messe Wien Paper: faculty.cc.gatech.edu/~badityap/pape… Code: github.com/AdityaLab/pinn…

Haoxin Liu (@hessianliu) 's Twitter Profile Photo

1/2 Thrilled to share that our work "Time-Series Forecasting for Out-of-Distribution Generalization Using Invariant Learning" will appear in #ICML2024. Thanks to my advisor B. Aditya Prakash and collaborators Chao Zhang Harsha Kamarthi Zhiyuan (Leo) Zhao. Let us meet in Vienna! GaTech CSE

Haoxin Liu (@hessianliu) 's Twitter Profile Photo

2/2 In this paper, we develop FOIL a new framework to alleviate the inherent out-of-distribution problem in time-series forecasting via invariant learning. Paper: arxiv.org/abs/2406.09130 Repo: github.com/AdityaLab/FOIL

2/2 In this paper, we develop FOIL a new framework to alleviate the inherent out-of-distribution problem in time-series forecasting via invariant learning.  
Paper: arxiv.org/abs/2406.09130
Repo:  github.com/AdityaLab/FOIL
Haoxin Liu (@hessianliu) 's Twitter Profile Photo

I am at #ICML24 all week. Catch me at my poster session: What: Time-Series Forecasting for Out-of-Distribution Generalization Using Invariant Learning When: Wed 24 Jul 11:30 a.m. — 1 p.m. CEST Where: Hall C 4-9 #2200 Paper: proceedings.mlr.press/v235/liu24ae.h… Code: github.com/AdityaLab/FOIL

Zhiyuan (Leo) Zhao (@leozhao_zhiyuan) 's Twitter Profile Photo

We're presenting our poster at #ACL2024: What: LSTPrompt: Large Language Models as Zero-Shot Time Series Forecasters by Long-Short-Term Prompting When: Aug 14, 12:15-13:15 ICT Where: Convention Center A1 Paper: arxiv.org/abs/2402.16132 Code: github.com/AdityaLab/lstp…

B. Aditya Prakash (@badityap) 's Twitter Profile Photo

Our new paper ACL 2025 on LLMs for time series forecasting (TSF)! We propose LSTPrompt, a new way for prompting for zero-shot TSF tasks. Main idea: split TSF into short+long-term sub-tasks, tailoring prompts to each & guide LLMs to adapt GaTech CSE Machine Learning at Georgia Tech github.com/AdityaLab/lstp…