Tian Liu

Hi there! I am a 2nd year PhD student in the Computer Vision Lab at CSE department of Texas A&M University (TAMU), working with Prof. Shu Kong and Prof. James Caverlee. Previously, I obtained my M.S. degree of Computer Science, at TAMU in 2023, and a M.S. degree of Petroleum Engineering, at TAMU in 2019. I worked as a field engineer for Schlumberger (2019-2020) and interned as a System Software Engineer (2021, 2022) at HPE.

Email  /  Google Scholar  /  CV  /  LinkedIn  /  Github

profile photo

News

  • Nov 7, 2024: our paper ERIC has won ACM BuildSys'24 Best paper award !!!
  • Oct 2024: 1 paper on adapting foundation model for video understanding was accepted to WACV'25.
  • Sep 2024: 1 paper on building efficient computer vision system was accepted to ACM BuildSys'24 as Best paper candidate!
  • June 2024: Coordinated the 4th Open World Vision Workshop at CVPR'24.
  • June 2024: I will present our paper in CVPR'24, Seattle.
  • June 2024: Our paper "The Neglected Tails in Vision-Language Models" was accepted to ICML 2024 DMLR Oral.
  • Mar 2024: I was awarded TAMU CSE department travel grant.
  • Mar 2024: I received TAMU CSE Graduate Teaching Assistant Excellence Award (1 each year).
  • Mar 2024: I passed Ph.D. qualify exam with 99% percentile.
  • Feb 2024: 1 paper on improving Vision Language Models for zero-shot recognition was accepted to CVPR'24.

Research

My research focuses on computer vision with Vision Language Models, addressing a central question: "how to adapt pretrained foundation models to better serve specific downstream tasks, with none or limited labeled data?". This question drives my efforts to identify and analyze the limitations of foundation models, examining their pretraining data to understand the origins of these limitations. My prior work includes developing advanced prompting and retrieval-augmented learning techniques for zero-shot recognition, as well as stage-wise augmented finetuning methods for few-shot recognition. I am also exploring the application of foundation models for video understanding, with specific interests in detecting unusual activities, such as failure actions or behaviors indicative of autism in children.

In addition, I have broad interest in cyber-physical systems, including developing efficient machine learning systems on resource-constrained edge devices. Applications include efficient vision system for precision residential irrigation and voice assistance system for emergency medical services.

Beyond these areas, my research extends to machine learning for healthcare and geoscience. This includes developing safe reinforcement learning algorithms for personalized medicine; enhancing the fairness of cardiovascular disease (CVD) risk prediction models for underrepresented populations; advancing geoscientific methods for characterizing subsurface fracture distribution and integrating 4D seismic data for more accurate reservoir model calibration.



Few-Shot Recognition via Stage-Wise Retrieval-Augmented Finetuning

Tian Liu, Huixin Zhang, Shubham Parashar, Shu Kong

arxiv / project / code

We explore retrieval-augmented learning for solving few-shot recognition, and propose stage-wise retrieval-Augmented fineTuning (SWAT) method to mitigate the imbalanced distribution and domain gap issues, outperforming SOTA methods by >6% accuracy.

ERIC: Estimating Rainfall with Commodity Doorbell Camera for Precision Residential Irrigation

Tian Liu, Liuyi Jin, Radu Stoleru, Amran Haroon, Charles Swanson, Kexin Feng

[BuildSys 2024]   Best paper award (1 out of 89 submissions)

paper / arxiv / video presentation / slides / code

We develop efficient vision system to estimate hyperlocal rainfall from doorbell camera for precision residential irrigation, saving > 9,000 gallons of water per month.

The Neglected Tails in Vision-Language Models

Shubham Parashar*, Zhiqiu Lin*, Tian Liu* (*co-first authors), Xiangjue Dong, Yanan Li, Deva Ramanan, James Caverlee, Shu Kong

[CVPR 2024, ICML 2024 DMLR Oral]   paper / DMLR / arxiv / project / code

We expose the long-tailed concept distributions in VLMs' pretraining data and reveal failues of SOTA multimodal systems (e.g. GPT-4V, DALL-E 3). We propose retrieval-augmented learning, achieving SOTA zero-shot recognition performance.

EMSAssist: An End-to-End Mobile Voice Assistant at the Edge for Emergency Medical Services

Liuyi Jin, Tian Liu, Amran Haroon, Radu Stoleru, Michael Middleton, Ziwei Zhu, Theodora Chaspari

[MobiSys 2023]   paper / demo / video / code

We build the first end-to-end mobile voice assistant system to assist Emergency Medical Technicians in selecting proper protocols for critical medical intervention.

Safe Reinforcement Learning with Contextual Information: Theory and Applications

Junyu Cao, Esmaeil Keyvanshokooh, Tian Liu

ssrn / code

We develop a safe reinforcement learning algorithm for personalized medical prescription considering patient's contextual information (e.g. age, gender, race etc.), achieveing sub-linear regret with zero safety violation.

A Machine Learning-based Hybrid Model for Fracture Parameterization and Distribution Prediction in Unconventional Reservoirs

Tian Liu, Ruxin Zhang

[Computers and Geotechnics 2024]   paper

We develop Variational Autoencoder (VAE) model for fracture parameterization and distribution prediction using reservoir production data.

Integration of Time-lapse Seismic Data using the Onset Time Approach: the Impact of Seismic Survey Frequency

Tian Liu, Hongquan Chen, Gill Hetz, Akhil Datta-Gupta

[Journal of Petroleum Science and Engineering 2020]  

1st Place of TAMU Student Paper Contest, 3rd Place of International Championship

JPSE journal paper / ATCE paper

We develop onset-time approach for efficient and robust integration of 4D seismic data for reservoir model calibration, achieveing 2x error reduction and 6x speedup compared to traditional amplitude-matching methods.

Teaching Assistance

  • CSCE606: Software Engineering, 2023 Fall
  • CSCE313: Introduction to Computer Systems, 2023 Summer
  • CSCE110: Programming I, 2023 Summer

Professional Services

  • Coordinator of 4th Open World Vision Workshop at CVPR'24.
  • Reviewer for Pattern Recognition, WACV'25
  • Reviewer for IEEE Internet of Things Journal
  • Reviewer for Applied Thermal Engineering, Geoenergy Science and Enginerring, SPE Journal

Selected Awards

  • BuildSys, Best Paper Award, 2024
  • TAMU CSE Department Travel Grant, 2024
  • TAMU CSE Department Graduate Teaching Assistant Excellence Award (1 each year), 2024
  • 1st place of SPE Student Paper Contest in TAMU, 1st place of Gulf Coast Region, 3rd place of International Championship, 2018
  • 2nd place of SPE Petrobowl Knowledge Contest in North American Region, 2017
  • 1st place of SPE Petrobowl Knowledge Contest in Asia-Pacific Region, 2015
  • Dean's Award (4 out of 296), China University of Petroleum Beijing, 2014
  • National Scholarship (highest honor in China), Ministry of Education of China, 2012

Miscellaneous

  • I have two big (20 lbs :) orange Maine Coon brothers, named Thor and Loki.

Template modified from Jon Barron