#generative models
32 resultsError Analysis of Neural-Network-Based Engression
Juntong Chen, Zijian Guo, Xinwei Shen
The paper analyzes the theoretical error of neural‑network‑based engression, a method for learning conditional distributions via an energy score, and derives convergence rates by d…
Forecasting Land Art Under Climate Scenarios
Alev Cinbarci, Sean Kalaycioglu
The paper builds a two‑stage pipeline to forecast visual complexity of the Spiral Jetty land artwork under future climate scenarios, using climate model outputs, statistical regres…
RIPPLE: Generating Multi-Channel Phase, Not Recovering It
Jaehyuk Lee, Yeajin Lee, Dayeon Shin +1
The paper introduces RIPPLE, a method that generates inter‑channel phase directly using a prior‑based Griffin–Lim approach and rectified flow, improving phase coherence for multi‑c…
Amortized Moment Matching for Visual Generation
Wenze Liu, Xintao Wang, Pengfei Wan +1
The paper introduces amortized moment matching, using neural networks to learn data moments as training signals, and proposes the Amortized Fréchet Distance loss to improve one-ste…
BayesAME: Bayesian Active Model Evaluation
Paula Cordero Encinar, Taylan Cemgil, Arnaud Doucet +2
BayesAME is a Bayesian sequential framework that automatically determines the size of a coreset for evaluating large generative models, using latent ability models and information‑…
Allure of Craquelure: A Variational-Generative Approach to Crack Detection in Paintings
Laura Paul, Holger Rauhut, Martin Burger +2
The paper presents a hybrid method that treats crack detection in digitized paintings as an inverse problem, using a deep generative model to represent the crack-free artwork and a…
Escaping Model Collapse via Synthetic Data Verification: Near-term Improvements and Long-term Convergence
Bingji Yi, Qiyuan Liu, Yuwei Cheng +1
The paper proposes using an external synthetic data verifier to prevent model collapse during iterative retraining of generative models on their own synthetic data, providing theor…
Conditional Generative Learning Enabled Wireless UAV Sensing and Tracking via Point Cloud Imaging
Xinhong Dai, Yuan Gao, Hao Jiang +2
The paper introduces a method that uses a base‑station antenna array to illuminate a UAV and reconstruct its 3‑D electromagnetic point cloud from reflected echoes, employing a cond…
RAD: Retrieval High-quality Demonstrations to Enhance Decision-making
Lu Guo, Yixiang Shan, Zhengbang Zhu +5
The paper proposes RAD, a method that improves offline reinforcement learning by retrieving high-return states from the dataset and generating sub-trajectories toward these targets…
TanGO: Training-Free 3D Editing via Tangent-Space Guidance and Optimization
Siwoo Lim, Sunjae Yoon, Gwanhyeong Koo +2
TanGO is a training-free framework that edits 3D content from flow‑matching generative models by applying adaptive per‑token control in the tangent space of the model's dynamics, r…
Zero-Shot Quantization for Object Detectors using Off-the-Shelf Generative Models
Hyunho Lee, Kyomin Hwang, Hyeonjin Kim +3
The paper proposes GoodQ, a method that uses off-the-shelf generative models to create synthetic training data for zero-shot quantization of object detectors, enabling low-bit quan…
BridgeFlow: Fast and Robust SE(2)-Equivariant Motion Planning with Flow Matching
Xinzhe Zhou, Xuyang Wang, Xiaoming Duan +1
BridgeFlow is a generative motion planning method that achieves exact SE(2) equivariance through a lightweight canonicalization step and uses a Brownian bridge prior with optimal t…
Fully AI-Generated Image Detection: Definition, Recent Advances and Challenges
Qijie Xu, Can Wang, Jiawei Chen +2
The paper surveys methods for detecting fully AI‑generated images, focusing on how datasets are built and how detectors extract artifacts left by generative models.
MultiAnimate: A Unified Framework for Controllable Multi-Character Animation
Zhongyi Zhang, Guangyuan Wang, Li Hu +6
The paper presents MultiAnimate, a framework that can animate several characters together in a shared scene while keeping each character's appearance and spatial relationships cons…
DNA: Dual-stage Native Attribution for Generated Image Source Tracing
Chao Wang, Kejiang Chen, Zijin Yang +4
The paper proposes DNA, a two‑stage framework that attributes generated images to their source models without additional training by first screening at the family level and then pi…
Beyond Chain-of-Thought: Rewrite as a Universal Interface for Generative Multimodal Embeddings
Peixi Wu, Ke Mei, Feipeng Ma +15
The paper introduces RIME, a rewrite-driven framework that improves multimodal embeddings by jointly optimizing generation and retrieval-friendly rewriting, aligning generative and…
The Steering Budget: Examples beat Knobs
Raj Kumar Rajendran
The paper shows that controllable generation in generative models is limited by a "steering budget" set by the training data, and that providing concrete examples can reach the ful…
Generative Synthetic Data for Causal Inference: Pitfalls, Remedies, and Opportunities
Yichen Xu
The paper examines how fully generative synthetic data models can preserve predictive performance but distort causal estimates, and proposes a hybrid synthetic-data approach that s…
Lyapunov Guidance: A Unified Framework for Stabilizing Generative Flows
Jingdong Zhang, Xinze Li, Yize Jiang +3
The paper introduces LyaGuide, a Lyapunov‑based framework that treats flow guidance as a control problem, providing explicit stability guarantees for generative flow models while r…
FORCE-Interior: A Poisson Flow Generative Prior for Interior Tomography Reconstruction
Kang Chen, Wenjun Xia, Jianxu Wang +2
The paper introduces FORCE-Interior, a Poisson‑flow generative reconstruction framework that incorporates measurement‑constrained initialization and per‑step data consistency to im…
SeamGen: Artist-Aligned UV Seam Generation via Graph Flow Matching
Hao Xu, Yuqing Zhang, Yiqian Wu +5
The paper introduces SeamGen, a generative model that learns artist-preferred UV seam layouts from existing data using flow‑matching and a mesh‑aware Transformer, enabling controll…
Contrastive-Augmented Flow Matching for Style-Content Disentanglement
Yusong Li, Pingchuan Ma, Ming Gui +2
The paper proposes Contrastive Augmented Flow Matching (CAtFM), a method that adds contrastive regularization to invertible flow matching to learn disentangled content and style re…
Sample Efficient Generative Optimization for Molecular Design
Sarina Kopf, Cristina Nevado, Philippe Schwaller
The paper proposes SEGO, a Bayesian optimization framework that steers a generative model to propose molecules, achieving strong molecular design performance with far fewer expensi…
WanToFight: Real-Time Generative Game Engine for Multi-Player Combat Interaction
Li Hu, Guangyuan Wang, Peng Zhang +1
WanToFight is a generative game engine that uses a video diffusion transformer to produce real-time, two-player fighting game visuals from keyboard inputs, handling multi-player co…