About

Xiao Liu is an Assistant Professor in the Department of Educational Psychology at The University of Texas at Austin, specializing in Quantitative Methods. She holds a PhD in Quantitative Psychology and an MS in Applied Statistics from the University of Notre Dame, as well as a BS in Statistics from Renmin University of China. Her research focuses on developing and applying statistical methods for causal inference related to the effects of interventions in areas concerning children, families, communities, schools, and health. Her methodological expertise includes mediation and moderation analyses, longitudinal data analyses, sensitivity analysis, and methods for clustered data. She has previous experience as a research assistant in labs at the University of Notre Dame, contributing to the advancement of statistical methodologies for developmental and health research.

Research topics

• Computer Science
• Algorithm
• Artificial Intelligence
• Applied mathematics
• Mathematical analysis
• Geometry
• Mathematics
• Mathematical optimization

Selected publications

• Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow

  arXiv (Cornell University) · 2022 · 75 citations

  1st authorCorresponding
  • Computer Science
  • Algorithm
  • Computer Science

  We present rectified flow, a surprisingly simple approach to learning (neural) ordinary differential equation (ODE) models to transport between two empirically observed distributions π_0 and π_1, hence providing a unified solution to generative modeling and domain transfer, among various other tasks involving distribution transport. The idea of rectified flow is to learn the ODE to follow the straight paths connecting the points drawn from π_0 and π_1 as much as possible. This is achieved by solving a straightforward nonlinear least squares optimization problem, which can be easily scaled to large models without introducing extra parameters beyond standard supervised learning. The straight paths are special and preferred because they are the shortest paths between two points, and can be simulated exactly without time discretization and hence yield computationally efficient models. We show that the procedure of learning a rectified flow from data, called rectification, turns an arbitrary coupling of π_0 and π_1 to a new deterministic coupling with provably non-increasing convex transport costs. In addition, recursively applying rectification allows us to obtain a sequence of flows with increasingly straight paths, which can be simulated accurately with coarse time discretization in the inference phase. In empirical studies, we show that rectified flow performs superbly on image generation, image-to-image translation, and domain adaptation. In particular, on image generation and translation, our method yields nearly straight flows that give high quality results even with a single Euler discretization step.

  DOI

• FuseDream: Training-Free Text-to-Image Generation with Improved CLIP+GAN Space Optimization

  arXiv (Cornell University) · 2021 · 36 citations

  1st authorCorresponding
  • Computer Science
  • Computer Science
  • Artificial Intelligence

  Generating images from natural language instructions is an intriguing yet highly challenging task. We approach text-to-image generation by combining the power of the retrained CLIP representation with an off-the-shelf image generator (GANs), optimizing in the latent space of GAN to find images that achieve maximum CLIP score with the given input text. Compared to traditional methods that train generative models from text to image starting from scratch, the CLIP+GAN approach is training-free, zero shot and can be easily customized with different generators. However, optimizing CLIP score in the GAN space casts a highly challenging optimization problem and off-the-shelf optimizers such as Adam fail to yield satisfying results. In this work, we propose a FuseDream pipeline, which improves the CLIP+GAN approach with three key techniques: 1) an AugCLIP score which robustifies the CLIP objective by introducing random augmentation on image. 2) a novel initialization and over-parameterization strategy for optimization which allows us to efficiently navigate the non-convex landscape in GAN space. 3) a composed generation technique which, by leveraging a novel bi-level optimization formulation, can compose multiple images to extend the GAN space and overcome the data-bias. When promoted by different input text, FuseDream can generate high-quality images with varying objects, backgrounds, artistic styles, even novel counterfactual concepts that do not appear in the training data of the GAN we use. Quantitatively, the images generated by FuseDream yield top-level Inception score and FID score on MS COCO dataset, without additional architecture design or training. Our code is publicly available at \url{https://github.com/gnobitab/FuseDream}.

  DOI

Frequent coauthors

• Chengyue Gong

  16 shared

• Lemeng Wu

  12 shared

• Shujian Zhang

  10 shared

• Haiqiang Jiang

  Northeast Agricultural University

  10 shared

• Enliang Wang

  University of Warwick

  9 shared

• Ping Bie

  9 shared

• Yunfei Chen

  Southeast University

  9 shared

• Hailian Wang

  University of Electronic Science and Technology of China

  8 shared

Labs

• The Families, Schools, and Communities Translational Research LabPI

Awards & honors

• Outstanding Quantitative Dissertation Award, American Educat…

Similar researchers at University of Texas at Austin

• Carlos Caldash-218
• Thomas J. Hughesh-154
• George Georgiouh-107
• Keith Johnstonh-106
• Jeffrey Andrewsh-104