Xiaozhong Lyu

ReViV: Reconstructing the Viewer and the View in 4D from Monocular Egocentric Video

Xiaozhong Lyu*, Gen Li*, Zhiyin Qian, Xucong Zhang, Marc Pollefeys, Siyu Tang

European Conference on Computer Vision (ECCV), 2026  (*equal contribution)

Paper / Project Page / Code

ReViV reconstructs viewer-centric human motion (body, hand, and gaze) and view-centric scene geometry (camera and depth) from a single egocentric RGB video in a unified feed-forward model.

LaMP: Learning Robust Latent Motion Prior for Optimization-Based Human Motion Generation

Xiaozhong Lyu, Korrawe Karunratanakul, Kaifeng Zhao, Siyu Tang

International Conference on 3D Vision (3DV), 2026

Paper / Project Page / Code

LaMP learns a robust latent motion prior with a part-based masked autoencoder, enabling optimization-based motion editing, blending, refinement, and collision avoidance.

EgoGen: An Egocentric Synthetic Data Generator

Gen Li, Kaifeng Zhao, Siwei Zhang, Xiaozhong Lyu, Mihai Dusmanu, Yan Zhang, Marc Pollefeys, Siyu Tang

Conference on Computer Vision and Pattern Recognition (CVPR), 2024  Oral

Paper / Project Page / Code

EgoGen is a new synthetic data generator that produces accurate and rich ground-truth training data for egocentric perception tasks.

RedPajama: an Open Dataset for Training Large Language Models

Maurice Weber, Daniel Fu, Quentin Anthony, Yonatan Oren, Shane Adams, Anton Alexandrov, Xiaozhong Lyu, Huu Nguyen, Xiaozhe Yao, Virginia Adams, Ben Athiwaratkun, Rahul Chalamala, Kezhen Chen, Max Ryabinin, Tri Dao, Percy Liang, Christopher Ré, Irina Rish, Ce Zhang

Conference on Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track, 2024

Paper / Dataset / Code

RedPajama is a fully open dataset family for LLM pretraining: an open reproduction of the LLaMA training corpus (V1) and a 100+ trillion token web dataset with quality signals and metadata (V2) for transparent, principled data curation.

Improving Retrieval-Augmented Large Language Models via Data Importance Learning

Xiaozhong Lyu, Stefan Grafberger, Samantha Biegel, Shaopeng Wei, Meng Cao, Sebastian Schelter, Ce Zhang

Paper / Code

A multilinear-extension algorithm that evaluates the data importance of retrieved data points in polynomial time, improving retrieval-augmented LLMs by pruning or reweighting the retrieval corpus without any further training.