Author name cluster

Ji Li

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

12 papers

2 author rows

AAAI Conference 2026 Conference Paper

Image Restoration via Primal Dual Hybrid Gradient and Flow Generative Model

Ji Li
Chao Wang

Regularized optimization has been a classical approach to solving imaging inverse problems, where the regularization term enforces desirable properties of the unknown image. Recently, the integration of flow matching generative models into image restoration has garnered significant attention, owing to their powerful prior modeling capabilities. In this work, we incorporate such generative priors into a Plug-and-Play (PnP) framework based on proximal splitting, where the proximal operator associated with the regularizer is replaced by a time-dependent denoiser derived from the generative model. While existing PnP methods have achieved notable success in inverse problems with smooth squared ℓ2 data fidelity--typically associated with Gaussian noise--their applicability to more general data fidelity terms remains underexplored. To address this, we propose a general and efficient PnP algorithm inspired by the primal-dual hybrid gradient (PDHG) method. Our approach is computationally efficient, memory-friendly, and accommodates a wide range of fidelity terms. In particular, it supports both ℓ1 and ℓ2 norm-based losses, enabling robustness to non-Gaussian noise types such as Poisson and impulse noise. We validate our method on several image restoration tasks, including denoising, super-resolution, deblurring, and inpainting, and demonstrate that ℓ1 and ℓ2 fidelity terms outperform the conventional squared ℓ2 loss in the presence of non-Gaussian noise.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Integrating Reweighted Least Squares with Plug-and-Play Diffusion Priors for Noisy Image Restoration

Ji Li
Chao Wang

Existing plug-and-play image restoration methods typically employ off-the-shelf Gaussian denoisers as proximal operators within classical optimization frameworks based on variable splitting. Recently, denoisers induced by generative priors have been successfully integrated into regularized optimization methods for image restoration under Gaussian noise. However, their application to non-Gaussian noise--such as impulse noise--remains largely unexplored. In this paper, we propose a plug-and-play image restoration framework based on generative diffusion priors for robust removal of general noise types, including impulse noise. Within the maximum a posteriori (MAP) estimation framework, the data fidelity term is adapted to the specific noise model. Departing from the conventional least-squares loss used for Gaussian noise, we introduce a generalized Gaussian scale mixture-based loss, which approximates a wide range of noise distributions and leads to an ℓq-norm fidelity term. This optimization problem is addressed using an iteratively reweighted least squares (IRLS) approach, wherein the proximal step involving the generative prior is efficiently performed via a diffusion-based denoiser. Experimental results on benchmark datasets demonstrate that the proposed method effectively removes non-Gaussian impulse noise and achieves superior restoration performance.

PDF Details DOI

AAAI Conference 2025 Conference Paper

DesignEdit: Unify Spatial-Aware Image Editing via Training-free Inpainting with a Multi-Layered Latent Diffusion Framework

Yueru Jia
Aosong Cheng
Yuhui Yuan
Chuke Wang
Ji Li
Huizhu Jia
Shanghang Zhang

Spatial-aware image editing focuses on modifying the position and size of elements within a given image. However, previous works still struggle with maintaining background harmony in the original editing areas, as well as preserving the initial identity of the edited elements, making it difficult to achieve complex multi-object editing in a single pass. In this paper, we aim to perform flexible spatial editing in a simple yet straightforward manner. We propose to inpaint the background first and develop a two-stage multi-layered latent diffusion framework to edit each element independently. Specifically, we design a key-masking self-attention scheme alongside artifact suppression to achieve background inpainting within the denoising process, leveraging the powerful generative capabilities of the Latent Diffusion Model, Stable Diffusion XL-1.0. The latent decomposition and fusion framework is capable of unifying various spatial-aware operations, including removal, resizing, relocation, flipping, addition, camera panning, zooming out, occlusion-aware editing, and cross-image editing. Experiments demonstrate the superior inpainting quality for object removal, along with enhanced versatility and higher precision in spatial-aware editing achieved by our method.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

MMMG: A Massive, Multidisciplinary, Multi-Tier Generation Benchmark for Text-to-Image Reasoning

Yuxuan Luo
Ryan Yuan
Junwen Chen
Haonan Cai
Ziyi Yue
Yuwei Yang
Fatima Zohra Daha
Ji Li

In this paper, we introduce knowledge image generation as a new task, alongside the Massive Multi-Discipline Multi-Tier Knowledge-Image Generation Benchmark (MMMG) to probe the reasoning capability of image generation models. Knowledge images have been central to human civilization and to the mechanisms of human learning—a fact underscored by dual-coding theory and the picture-superiority effect. Generating such images is challenging, demanding multimodal reasoning that fuses world knowledge with pixel-level grounding into clear explanatory visuals. To enable comprehensive evaluation, MMMG offers $4, 456$ expert-validated (knowledge) image-prompt pairs spanning $10$ disciplines, $6$ educational levels, and diverse knowledge formats such as charts, diagrams, and mind maps. To eliminate confounding complexity during evaluation, we adopt a unified Knowledge Graph (KG) representation. Each KG explicitly delineates a target image’s core entities and their dependencies. We further introduce MMMG-Score to evaluate generated knowledge images. This metric combines factual fidelity, measured by graph-edit distance between KGs, with visual clarity assessment. Comprehensive evaluations of $21$ state-of-the-art text-to-image generation models expose serious reasoning deficits—low entity fidelity, weak relations, and clutter—with GPT-4o achieving an MMMG-Score of only $50. 20$, underscoring the benchmark’s difficulty. To spur further progress, we release FLUX-Reason (MMMG-Score of $34. 45$), an effective and open baseline that combines a reasoning LLM with diffusion models and is trained on $16, 000$ curated knowledge image–prompt pairs.

PDF Details

NeurIPS Conference 2024 Conference Paper

Aligning Vision Models with Human Aesthetics in Retrieval: Benchmarks and Algorithms

Miaosen Zhang
Yixuan Wei
Zhen Xing
Yifei Ma
Zuxuan Wu
Ji Li
Zheng Zhang
Qi Dai

Modern vision models are trained on very large noisy datasets. While these models acquire strong capabilities, they may not follow the user's intent to output the desired results in certain aspects, e. g. , visual aesthetic, preferred style, and responsibility. In this paper, we target the realm of visual aesthetics and aim to align vision models with human aesthetic standards in a retrieval system. Advanced retrieval systems usually adopt a cascade of aesthetic models as re-rankers or filters, which are limited to low-level features like saturation and perform poorly when stylistic, cultural or knowledge contexts are involved. We find that utilizing the reasoning ability of large language models (LLMs) to rephrase the search query and extend the aesthetic expectations can make up for this shortcoming. Based on the above findings, we propose a preference-based reinforcement learning method that fine-tunes the vision models to distill the knowledge from both LLMs reasoning and the aesthetic models to better align the vision models with human aesthetics. Meanwhile, with rare benchmarks designed for evaluating retrieval systems, we leverage large multi-modality model (LMM) to evaluate the aesthetic performance with their strong abilities. As aesthetic assessment is one of the most subjective tasks, to validate the robustness of LMM, we further propose a novel dataset named HPIR to benchmark the alignment with human aesthetics. Experiments demonstrate that our method significantly enhances the aesthetic behaviors of the vision models, under several metrics. We believe the proposed algorithm can be a general practice for aligning vision models with human values.

PDF Details DOI

ICML Conference 2024 Conference Paper

DiffFPR: Diffusion Prior for Oversampled Fourier Phase Retrieval

Ji Li
Chao Wang

This paper tackled the challenging Fourier phase retrieval problem, the absolute uniqueness of which does not hold. The existence of equivalent solution (a. k. a. trivial solution ambiguity) hinders the successful recovery, especially for multi-channel color image. The traditional iterative engine, such as the Relaxed Averaged Alternating Reflections (RAAR), can be applied to reconstruct the image channel-wisely. However, due to the relative uniqueness of the solution, the restoration is not automatically aligned with the accurate orientation for each channel, resulting in a reconstructed image that deviates significantly from the true solution manifold. To address this issue, by penalizing the mismatch of the image channels, a diffusion model as the strong prior of the color image is integrated into the iterative engine. The combination of the traditional iterative engine and the diffusion model provides an effective solution to the oversampled Fourier phase retrieval. The formed algorithm, DiffFPR, is validated by experiments. The code is available at https: //github. com/Chilie/DiffFPR.

Details

EAAI Journal 2024 Journal Article

Multi-objective modeling of price and pollution in large-scale energy hubs with load management

Shuangbin Ma
Tao Li
Jun Ni
Jinxin Chen
Ji Li

This study introduces a novel multi-objective model that addresses emission costs, large-scale user operational expenses, and the efficiency of the electric storage system within an integrated energy hub encompassing heating, water, power, and gas sources. A key objective of this model is to maximize the performance of the electrical storage system. The study employs the epsilon-constraint method to construct the Pareto front, aiming to balance profitability and emission reduction from virtual power plant units. The final decision-making process utilizes a fuzzy decision-making technique. Additionally, demand response program (DRP) is incorporated to optimize peak-hour demand and align the load profile with defined objectives. The proposed approach is evaluated across various operational scenarios within a sample system, demonstrating its effectiveness and potential benefits in terms of cost reduction and environmental impact mitigation. Focusing on environmental impact, the carbon dioxide (CO2) emissions are also lower under the DRP scenario, amounting to 10253. 92 kg without DRP versus 10127. 74 kg with DRP. This reduction in emissions aligns with sustainable energy management goals, showcasing DRP's capability to mitigate environmental impacts associated with energy generation and consumption. The percentage improvements in total costs and emissions further highlight the advantages of employing DRP. The reductions in total costs range from approximately 1. 2%–1. 4%, demonstrating cost savings across different cost categories. Similarly, the decrease in CO2 emissions by approximately 1. 2% underscores DRP's role in promoting environmentally friendly energy practices.

Details DOI

AAAI Conference 2024 Conference Paper

Revisiting Disentanglement in Downstream Tasks: A Study on Its Necessity for Abstract Visual Reasoning

Ruiqian Nai
Zixin Wen
Ji Li
Yuanzhi Li
Yang Gao

In representation learning, a disentangled representation is highly desirable as it encodes generative factors of data in a separable and compact pattern. Researchers have advocated leveraging disentangled representations to complete downstream tasks with encouraging empirical evidence. This paper further investigates the necessity of disentangled representation in downstream applications. Specifically, we show that dimension-wise disentangled representations are unnecessary on a fundamental downstream task, abstract visual reasoning. We provide extensive empirical evidence against the necessity of disentanglement, covering multiple datasets, representation learning methods, and downstream network architectures. Furthermore, our findings suggest that the informativeness of representations is a better indicator of downstream performance than disentanglement. Finally, the positive correlation between informativeness and disentanglement explains the claimed usefulness of disentangled representations in previous works. The source code is available at https://github.com/Richard-coder-Nai/disentanglement-lib-necessity.git

PDF Details DOI

EAAI Journal 2023 Journal Article

Robust key parameter identification of dedicated hybrid engine performance indicators via K-fold filter collaborated feature selection

Xu He
Ji Li
Quan Zhou
Guoxiang Lu
Hongming Xu

Dedicated hybrid engine technology using auxiliary electronic components has been proven as an energy-saving solution to public concerns about energy consumption and carbon emissions. This paper proposes a generic approach of K-fold filter-collaborated feature selection (KFFC-FS) to robustly identify the key parameters of three engine performance indicators, i. e. , volumetric efficiency, thermal efficiency, and fuel consumption. By using this approach, five filters are collaborated to provide a robust rank of feature importance and avoid the feature overestimation caused by the single filter. Meanwhile, the K-fold cross validation method is introduced to avoid random precision issues and overfitting, further enhancing the robustness of key parameter identification for the independent engine performance indicators. In this research, the modelling data is collected from an experimental test bench with a BYD 1. 5L gasoline engine. Under the basics of the studied three engine performance indicators by using a multiple-layer perceptron network, the proposed approach further reduces by at least 10. 3% root-mean-square error (RMSE) and at least 30% reduction of the model inputs.

Details DOI

IJCAI Conference 2023 Conference Paper

Towards Lossless Head Pruning through Automatic Peer Distillation for Language Models

Bingbing Li
Zigeng Wang
Shaoyi Huang
Mikhail Bragin
Ji Li
Caiwen Ding

Pruning has been extensively studied in Transformer-based language models to improve efficiency. Typically, we zero (prune) unimportant model weights and train a derived compact model to improve final accuracy. For pruned weights, we treat them as useless and discard them. This usually leads to significant model accuracy degradation. In this paper, we focus on attention head pruning as head attention is a key component of the transformer-based language models and provides interpretable knowledge meaning. We reveal the relationship between pruned attention heads and retained heads and provide a solution to recycle the discarded knowledge from the pruned heads, named peer distillation. We also develop an automatic framework to locate the to-be-pruned attention heads in each layer, freeing the time-consuming human labor in tuning hyperparameters. Experimental results on the General Language Understanding Evaluation (GLUE) benchmark are provided using BERT model. By recycling discarded knowledge from pruned heads, the proposed method maintains model performance across all nine tasks while reducing heads by over 58% on average and outperforms state-of-the-art techniques (e. g. , Random, HISP, L0 Norm, SMP).

PDF Details DOI

IJCAI Conference 2013 Conference Paper

Guarantees of Augmented Trace Norm Models in Tensor Recovery

Ziqiang Shi
Jiqing Han
Tieran Zheng
Ji Li

This paper studies the recovery guarantees of the models of minimizing X ∗ + 1 2α X 2 F where X is a tensor and X ∗ and X F are the trace and Frobenius norm of respectively. We show that they can efﬁciently recover low-rank tensors. In particular, they enjoy exact guarantees similar to those known for minimizing X ∗ under the conditions on the sensing operator such as its null-space property, restricted isometry property, or spherical section property. To recover a low-rank tensor X0, minimizing X ∗ + 1 2α X 2 F returns the same solution as minimizing X ∗ almost whenever α ≥ 10 max i X0 (i) 2.

PDF Details DOI

IROS Conference 1995 Conference Paper

Grasp stiffness matrix-fundamental properties in analysis of grasping and manipulation

Ji Li
Imin Kao

In this paper, we present fundamental properties of stiffness matrix as applied to analysis of grasping and dextrous manipulation. The investigation unveils insights of stiffness matrix which are important in grasping and manipulation for robotic hands and fingers in R/sup 3/ space. A general grasp stiffness matrix can be broken into two parts-symmetric and antisymmetric. The symmetric part is derived from a conservative quadratic potential function in the Hermitian form; while the antisymmetric part is a function of nonconservative curl vector field of the grasp. The conservative part stores and interchanges energy with the environment with which the fingers make contact. The nonconservative part dissipates or increases energy. The theory suggests that it is possible to introduce a nonsymmetric stiffness matrix in robotic control so as to have energy dissipation (damping) effects. This is useful when passive damping effects are desirable in grasping. Application of the theory to the analysis of stiffness matrix in 3D is presented for analysis of grasping and manipulation.

Details