Guolong Wang

AAAI Conference 2026 Conference Paper

EAGLE: Episodic Appearance- and Geometry-aware Memory for Unified 2D-3D Visual Query Localization in Egocentric Vision

• Yifei Cao
• Yu Liu
• Guolong Wang
• Zhu Liu
• Kai Wang
• Xianjie Zhang
• Jizhe Yu
• Xun Tu

Egocentric visual query localization is vital for embodied AI and VR/AR, yet remains challenging due to camera motion, viewpoint changes, and appearance variations. We present EAGLE, a novel framework that leverages episodic appearance- and geometry-aware memory to achieve unified 2D-3D visual query localization in egocentric vision. Inspired by avian memory consolidation, EAGLE synergistically integrates segmentation guided by an appearance-aware meta-learning memory (AMM), with tracking driven by a geometry-aware localization memory (GLM). This memory consolidation mechanism, through structured appearance and geometry memory banks, stores high-confidence retrieval samples, effectively supporting both long- and short-term modeling of target appearance variations. This enables precise contour delineation with robust spatial discrimination, leading to significantly improved retrieval accuracy. Furthermore, by integrating the VQL-2D output with a visual geometry grounded Transformer (VGGT), we achieve a efficient unification of 2D and 3D tasks, enabling rapid and accurate back-projection into 3D space. Our method achieves state-of-the-art performance on the Ego4D-VQ benchmark.

EAAI Journal 2025 Journal Article

Corrigendum to “Intelligent evaluation of pavement friction at high speeds with artificial intelligence powered three-dimensional laser imaging technology” [Eng. Appl. Artific. Intellig. 150 (2025) 1–19 110580]

• Guolong Wang
• Kelvin C.P. Wang
• Guangwei Yang

EAAI Journal 2025 Journal Article

Intelligent evaluation of pavement friction at high speeds with artificial intelligence powered three-dimensional laser imaging technology

• Guolong Wang
• Kelvin C.P. Wang
• Guangwei Yang

Accurate non-contact pavement friction evaluation at high speeds faces many challenges when using low-resolution (LR) three-dimensional (3D) texture images or other deficient texture data. With artificial intelligence (AI) powered 0. 1-mm (0. 1-mm) 3D laser imaging technology, this paper proposed a two-step deep learning (DL) network, named Friction-8KNet, for accurate and intelligent non-contact pavement friction evaluation at high speeds. Particularly, a 3D laser imagining device was employed to collect LR texture images at a speed of 30 mph, which were processed via a DL-based super-resolution (SR) algorithm to obtain 0. 1 mm high-resolution (HR) images with a size of 8192 × 4096 (8K) pixels. The Friction-8KNet comprises a network backbone for texture feature extraction in Step 1 and a triple attention network for friction evaluation in Step 2. The network backbone is developed to precisely extract features of an HR image without requiring large graphics processing unit (GPU) memory. The triple attention net is designed with three function-specific attention modules to utterly mine the extracted features for accurate friction prediction. Experimental results show that Friction-8KNet can achieve 99. 19 % prediction accuracy and transcends models using other texture data, including small HR 3D images, LR 3D images, and two-dimensional (2D) texture profiles. This research promotes an accurate and efficient measurement of pavement friction for production level in a non-contact manner in the future.

NeurIPS Conference 2024 Conference Paper

Boosting Text-to-Video Generative Model with MLLMs Feedback

• Xun Wu
• Shaohan Huang
• Guolong Wang
• Jing Xiong
• Furu Wei

Recent advancements in text-to-video generative models, such as Sora, have showcased impressive capabilities. These models have attracted significant interest for their potential applications. However, they often rely on extensive datasets of variable quality, which can result in generated videos that lack aesthetic appeal and do not accurately reflect the input text prompts. A promising approach to mitigate these issues is to leverage Reinforcement Learning from Human Feedback (RLHF), which aims to align the outputs of text-to-video generative with human preferences. However, the considerable costs associated with manual annotation have led to a scarcity of comprehensive preference datasets. In response to this challenge, our study begins by investigating the efficacy of Multimodal Large Language Models (MLLMs) generated annotations in capturing video preferences, discovering a high degree of concordance with human judgments. Building upon this finding, we utilize MLLMs to perform fine-grained video preference annotations across two dimensions, resulting in the creation of VideoPrefer, which includes 135, 000 preference annotations. Utilizing this dataset, we introduce VideoRM, the first general-purpose reward model tailored for video preference in the text-to-video domain. Our comprehensive experiments confirm the effectiveness of both VideoPrefer and VideoRM, representing a significant step forward in the field.

NeurIPS Conference 2024 Conference Paper

Multimodal Large Language Models Make Text-to-Image Generative Models Align Better

• Xun Wu
• Shaohan Huang
• Guolong Wang
• Jing Xiong
• Furu Wei

Recent studies have demonstrated the exceptional potentials of leveraging human preference datasets to refine text-to-image generative models, enhancing the alignment between generated images and textual prompts. Despite these advances, current human preference datasets are either prohibitively expensive to construct or suffer from a lack of diversity in preference dimensions, resulting in limited applicability for instruction tuning in open-source text-to-image generative models and hinder further exploration. To address these challenges and promote the alignment of generative models through instruction tuning, we leverage multimodal large language models to create VisionPrefer, a high-quality and fine-grained preference dataset that captures multiple preference aspects. We aggregate feedback from AI annotators across four aspects: prompt-following, aesthetic, fidelity, and harmlessness to construct VisionPrefer. To validate the effectiveness of VisionPrefer, we train a reward model VP-Score over VisionPrefer to guide the training of text-to-image generative models and the preference prediction accuracy of VP-Score is comparable to human annotators. Furthermore, we use two reinforcement learning methods to supervised fine-tune generative models to evaluate the performance of VisionPrefer, and extensive experimental results demonstrate that VisionPrefer significantly improves text-image alignment in compositional image generation across diverse aspects, e. g. , aesthetic, and generalizes better than previous human-preference metrics across various image distributions. Moreover, VisionPrefer indicates that the integration of AI-generated synthetic data as a supervisory signal is a promising avenue for achieving improved alignment with human preferences in vision generative models.

IJCAI Conference 2018 Conference Paper

Collaborative and Attentive Learning for Personalized Image Aesthetic Assessment

• Guolong Wang
• Junchi Yan
• Zheng Qin

The ever-increasing volume of visual images has stimulated the demand for organizing such data by aesthetic quality. Automatic and especially learning based aesthetic assessment methods have shown potential by recent works. Existing image aesthetic prediction is often user-agnostic which may ignore the fact that the rating to an image can be inherently individual. We fill this gap by formulating the personalized image aesthetic assessment problem with a novel learning method. Specifically, we collect user-image textual reviews in addition with visual images from the public dataset to organize a review-augmented benchmark. Using this enriched dataset, we devise a deep neural network with a user/image relation encoding input for collaborative filtering. Meanwhile an attentive mechanism is designed to capture the user-specific taste for image semantic tags and regions of interest by fusing the image and user's review. Extensive and promising experimental results on the review-augmented benchmark corroborate the efficacy of our approach.