MVP: Deep Incomplete Multi-view Learning via Cyclic Permutation of VAEs

Xin Gao, Jian Pu

[Project page]

• MVP introduces a robust generative approach to incomplete multi-view representation learning by leveraging latent space correspondences in Variational Auto-Encoders. This enables the inference of missing modalities and enhances multi-view consistency even with irregularly missing information.

MacTok: Robust Continuous Tokenization for Image Generation

Hengyu Zeng#, Xin Gao#, Guanghao Li, Yuxiang Yan, Jiaoyang Ruan, Junpeng Ma, Haoyu Albert Wang, Jian Pu

• MacTok addresses the severe “posterior collapse” failure in continuous image tokenizers. By introducing DINO-guided semantic masking and multi-level representation alignment, it forces the model to infer robust semantics from incomplete visual evidence, yielding state-of-the-art generation fidelity.

Reasoning on the Manifold: Bidirectional Consistency for Self-Verification in Diffusion Language Models

Jiaoyang Ruan#, Xin Gao#, Yinda Chen, Hengyu Zeng, Liang Du, Guanghao Li, Jie Fu, Jian Pu

• BMC introduces a geometric perspective to reasoning in Diffusion Language Models. By framing verifiable trajectories as stable paths on a high-density manifold, it provides an unsupervised mechanism for logical self-verification, empowering models to dynamically verify generation steps and self-evolve.

GOOD: Training-Free Guided Diffusion Sampling for Out-of-Distribution Detection

Xin Gao, Jiyao Liu, Guanghao Li, Yueming Lyu, Jianxiong Gao, Weichen Yu, Ningsheng Xu, Liang Wang, Caifeng Shan, Ziwei Liu, Chenyang Si

• GOOD is a training-free diffusion guidance framework that shapes a robust OOD/ID classification boundary. It steers sampling with image-level and feature-level gradients to systematically generate diverse, controllable OOD examples for robust decision-making.

Learning Spatial-Aware Manipulation Ordering

Yuxiang Yan, Zhiyuan Zhou, Xin Gao, Guanghao Li, Shenglin Li, Jiaqi Chen, Qunyan Pu, Jian Pu

• OrderMind is a spatial-aware manipulation ordering framework that learns object priorities from local geometry via a kNN spatial graph and a lightweight module, supervised by VLM-distilled situational knowledge for robotic embodied tasks.

Escaping Low-Rank Traps: Interpretable Visual Concept Learning via Implicit Vector Quantization

Shujian Gao, Yuan Wang, Chenglong Ma, Xin Gao, Jiangtao Yan, Junzhi Ning, Cheng Tang, Changkai Ji, Huihui Xu, Wei Li, Ziyan Huang et al.

• IVQ acts as a structural regularizer to address representational collapse in Concept Bottleneck Models (CBMs). It anchors patch features to learned semantic prototypes continuously, preserving a high-rank, explainable representation space for visual concept learning.

Dark-ISP: Enhancing RAW Image Processing for Low-Light Object Detection

Jiasheng Guo#, Xin Gao#, Yuxiang Yan, Guanghao Li, Jian Pu

• Dark-ISP is a lightweight, self-adaptive ISP plugin that enhances low-light object detection by processing Bayer RAW images. It breaks down traditional pipelines into optimized sub-modules, using physics-informed priors for robust sensory perception.

ARTDECO: Toward High-Fidelity On-the-Fly Reconstruction with Hierarchical Gaussian Structure and Feed-Forward Guidance

Guanghao Li, Kerui Ren, Linning Xu, Zhewen Zheng, Changjian Jiang, Xin Gao, Bo Dai, Jian Pu, Mulin Yu, Jiangmiao Pang

[Project page]

• ARTDECO unifies feed-forward 3D foundation priors with SLAM-style global optimization. It utilizes a hierarchical 3D Gaussian scene representation to achieve efficient, robust, and high-fidelity on-the-fly monocular 3D reconstruction.

Loong: Synthesize Long Chain-of-Thoughts at Scale through Verifiers

Xingyue Huang, Rishabh, Gregor Franke, Ziyi Yang, Jiamu Bai, Weijie Bai, Jinhe Bi, Zifeng Ding, Yiqun Duan, Chengyu Fan, Wendong Fan, Xin Gao et al.

[Project page]

• Loong is an open-source framework for scalable, verifiable Chain-of-Thought (CoT) reasoning data generation. It pairs a large benchmark with a modular generator forming an agent–environment loop for broad-domain reasoning training and verification.

This document provides an overview of key concepts related to diffusion models, particularly focusing on the theoretical foundations, development timeline, and recent advancements in the field. The content includes detailed discussions on VAE, DDPM, DDIM, SDE, and ODE, as well as conditional guidance. It also covers the evolution of stable diffusion, including topics like Latent Diffusion, VQ-VAE, and DiT. Lastly, the document highlights the latest methodology, IC-Light, set to be presented at ICLR 2025.

This document introduces the t3-Variational Autoencoder (ICLR 2024), which uses Student’s t-distributions to model heavy-tailed data distributions and improve latent variable representations. It also explores the framework of Information Geometry, focusing on how generative models can be understood through statistical manifolds, divergences, and Riemannian metrics, providing a deeper understanding of probability distributions and their applications in machine learning, signal processing, and neuroscience.

This document introduces the X-metric framework (PAMI 2023), an N-dimensional information-theoretic approach designed for groupwise registration and deep combined computing, with applications in advanced machine learning tasks. It also covers the theoretical foundations, including entropy, mutual information, and the MLE algorithm, alongside the framework’s modifications for deep computing and network training.