Simin Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Simin Li has a total of 53 co-authors. Here are the top co-authors by number of shared publications:
- Xianglong Liu
Beihang University, Beijing, China
5 shared publications - Jiakai Wang
Zhongguancun Laboratory, Beijing, China
3 shared publications - Aishan Liu
Beihang University, Beijing, China
2 shared publications - Bo An
Nanyang Technological University
2 shared publications - Hewu Li
Tsinghua University, Beijing, China
2 shared publications - Pu Feng
State Key Laboratory of Software Development Environment, Beihang University, Beijing, China
2 shared publications
Latest Publications
- Open Access
- 15 June 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies
- Vol. 10 (2) , 1-38
The Pervasive Blind Spot: Benchmarking VLM Inference Risks on Everyday Personal Videos
Applying Vision-Language Models (VLMs) to pervasive personal videos introduces profound privacy risks. This paper addresses the critical yet unexplored inferential privacy threat, specifically the risk of inferring sensitive personal attributes from seemingly benign data. To address this gap, we crowdsourced a dataset of 508 everyday personal videos from 58 individuals, and benchmarked VLM inference capabilities against human performance. Our findings reveal three key insights: (1) VLMs surpass recruited human evaluators in inferential accuracy, analyzing temporal behavioral patterns rather than relying solely on object recognition. (2) Inferential risk is strongly correlated with specific video characteristics and prompting strategies. (3) VLM-driven explanation towards the inference is unreliable, as we observe a disconnect between the model's reasoning and evidential impact, where ubiquitous objects often serve as misleading confounders.
- 15 June 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies
- Vol. 10 (2) , 1-38
Open AccessThe Pervasive Blind Spot: Benchmarking VLM Inference Risks on Everyday Personal Videos
Applying Vision-Language Models (VLMs) to pervasive personal videos introduces profound privacy risks. This paper addresses the critical yet unexplored inferential privacy threat, specifically the risk of inferring sensitive personal attributes from seemingly benign data. To address this gap, we crowdsourced a dataset of 508 everyday personal videos from 58 individuals, and benchmarked VLM inference capabilities against human performance. Our findings reveal three key insights: (1) VLMs surpass recruited human evaluators in inferential accuracy, analyzing temporal behavioral patterns rather than relying solely on object recognition. (2) Inferential risk is strongly correlated with specific video characteristics and prompting strategies. (3) VLM-driven explanation towards the inference is unreliable, as we observe a disconnect between the model's reasoning and evidential impact, where ubiquitous objects often serve as misleading confounders.
- Open Access
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1614-1625
Bayesian Robust Financial Trading with Adversarial Synthetic Market Data
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1614-1625
Open AccessBayesian Robust Financial Trading with Adversarial Synthetic Market Data
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 3984-3995
CASE: Conflict-assessed Knowledge-sensitive Neuron Tuning for Lifelong Model Editing
Large Language Models (LLMs) inevitably encounter factual hallucinations and knowledge obsolescence, necessitating lifelong knowledge editing to sustain reliability and factual advancement. While mainstream lifelong editing paradigms aim to alleviate knowledge forgetting through allocating and updating isolated parameter subspaces, they often overlook conflict assessment among distinct editing processes, leading to unjustified subspace allocation and indiscriminate neuron tuning. To address these issues, we propose the Conflict-Assessed Sensitive Editing (CASE) framework, which integrates a Conflict-Assessed Editing Allocation (CAA) module and a Knowledge-sensitive Neuron Tuning (KNT) strategy. The CAA module quantitatively assesses editing conflicts to enable justified subspace allocation, thereby reducing globally significant conflicts and routing errors. The KNT strategy adaptively identifies and tunes knowledge-sensitive neurons through a calibrated sensitivity threshold, effectively eliminating local conflicts and enhancing subspace stability. Extensive experiments on standard lifelong editing benchmarks demonstrate that CASE achieves state-of-the-art performance, improving average editing accuracy by nearly 10% after 1,000 sequential edits. Overall, CASE substantially mitigates editing conflicts and enhances knowledge retention, offering a scalable and conflict-resilient solution for lifelong model editing.
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 3984-3995
Open AccessCASE: Conflict-assessed Knowledge-sensitive Neuron Tuning for Lifelong Model Editing
Large Language Models (LLMs) inevitably encounter factual hallucinations and knowledge obsolescence, necessitating lifelong knowledge editing to sustain reliability and factual advancement. While mainstream lifelong editing paradigms aim to alleviate knowledge forgetting through allocating and updating isolated parameter subspaces, they often overlook conflict assessment among distinct editing processes, leading to unjustified subspace allocation and indiscriminate neuron tuning. To address these issues, we propose the Conflict-Assessed Sensitive Editing (CASE) framework, which integrates a Conflict-Assessed Editing Allocation (CAA) module and a Knowledge-sensitive Neuron Tuning (KNT) strategy. The CAA module quantitatively assesses editing conflicts to enable justified subspace allocation, thereby reducing globally significant conflicts and routing errors. The KNT strategy adaptively identifies and tunes knowledge-sensitive neurons through a calibrated sensitivity threshold, effectively eliminating local conflicts and enhancing subspace stability. Extensive experiments on standard lifelong editing benchmarks demonstrate that CASE achieves state-of-the-art performance, improving average editing accuracy by nearly 10% after 1,000 sequential edits. Overall, CASE substantially mitigates editing conflicts and enhances knowledge retention, offering a scalable and conflict-resilient solution for lifelong model editing.
- Open Access
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 33-42
Towards Aligning Personalized AI Agents with Users' Privacy Preference
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 33-42
Open AccessTowards Aligning Personalized AI Agents with Users' Privacy Preference
- 1 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 36 (10) , 18118-18132
Robust Multi-Agent Reinforcement Learning by Mutual Information Regularization
In cooperative multi-agent reinforcement learning (MARL), ensuring robustness against cooperative agents making unpredictable or worst-case adversarial actions is crucial for real-world deployment. In multi-agent settings, each agent may be perturbed or unperturbed, leading to an exponential increase in potential threat scenarios as the number of agents grows. Existing robust MARL methods either enumerate, or approximate all possible threat scenarios, leading to intense computation and insufficient robustness. In contrast, humans develop robust behaviors by maintaining a general level of caution rather than preparing for every possible threat. Inspired by human decision making, we frame robust MARL as a control-as-inference problem, and optimize worst-case robustness across all threat scenarios implicitly optimized through off-policy evaluation. Specifically, we introduce mutual information regularization as robust regularization (MIR3), which maximizes a lower bound on robustness during routine training, serving as a kind of caution for MARL without adversarial inputs. Further insights show that MIR3 acts as an information bottleneck, preventing agents from over-reacting to others and aligning policies with robust action priors. In the presence of worst-case adversaries, our MIR3 significantly surpasses baseline methods in robustness and training efficiency, and maintaining cooperative performance in StarCraft II, quadrotor swarm control, and robot swarm control. When deploying the robot swarm control algorithm in the real world, our method also outperforms the best baseline by 14.29% in reward. See code and demo videos at https://github.com/DIG-Beihang/MIR3.
- 1 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 36 (10) , 18118-18132
Robust Multi-Agent Reinforcement Learning by Mutual Information Regularization
In cooperative multi-agent reinforcement learning (MARL), ensuring robustness against cooperative agents making unpredictable or worst-case adversarial actions is crucial for real-world deployment. In multi-agent settings, each agent may be perturbed or unperturbed, leading to an exponential increase in potential threat scenarios as the number of agents grows. Existing robust MARL methods either enumerate, or approximate all possible threat scenarios, leading to intense computation and insufficient robustness. In contrast, humans develop robust behaviors by maintaining a general level of caution rather than preparing for every possible threat. Inspired by human decision making, we frame robust MARL as a control-as-inference problem, and optimize worst-case robustness across all threat scenarios implicitly optimized through off-policy evaluation. Specifically, we introduce mutual information regularization as robust regularization (MIR3), which maximizes a lower bound on robustness during routine training, serving as a kind of caution for MARL without adversarial inputs. Further insights show that MIR3 acts as an information bottleneck, preventing agents from over-reacting to others and aligning policies with robust action priors. In the presence of worst-case adversaries, our MIR3 significantly surpasses baseline methods in robustness and training efficiency, and maintaining cooperative performance in StarCraft II, quadrotor swarm control, and robot swarm control. When deploying the robot swarm control algorithm in the real world, our method also outperforms the best baseline by 14.29% in reward. See code and demo videos at https://github.com/DIG-Beihang/MIR3.
- 21 June 2025
- Published by Elsevier in Neural Networks
- Vol. 191, 107747
Attacking cooperative multi-agent reinforcement learning by adversarial minority influence
- 21 June 2025
- Published by Elsevier in Neural Networks
- Vol. 191, 107747
Attacking cooperative multi-agent reinforcement learning by adversarial minority influence
- Open Access
- 18 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 71-80
A Comprehensive Evaluation Framework for Multi-Agent Reinforcement Learning
- 18 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 71-80
Open AccessA Comprehensive Evaluation Framework for Multi-Agent Reinforcement Learning
- 29 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 9 (12) , 11138-11145
Safe and Efficient Multi-Agent Collision Avoidance With Physics-Informed Reinforcement Learning
Reinforcement learning (RL) has shown great promise in addressing multi-agent collision avoidance challenges. However, existing RL-based methods often suffer from low training efficiency and poor action safety. To tackle these issues, we introduce a physics-informed reinforcement learning framework equipped with two modules: a Potential Field (PF) module and a Multi-Agent Multi-Level Safety (MAMLS) module. The PF module uses the Artificial Potential Field method to compute a regularization loss, adaptively integrating it into the critic's loss to enhance training efficiency. The MAMLS module formulates action safety as a constrained optimization problem, deriving safe actions by solving this optimization. Furthermore, to better address the characteristics of multi-agent collision avoidance tasks, multi-agent multi-level constraints are introduced. The results of simulations and real-world experiments showed that our physics-informed framework offers a significant improvement in terms of both the efficiency of training and safety-related metrics over advanced baseline methods.
- 29 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 9 (12) , 11138-11145
Safe and Efficient Multi-Agent Collision Avoidance With Physics-Informed Reinforcement Learning
Reinforcement learning (RL) has shown great promise in addressing multi-agent collision avoidance challenges. However, existing RL-based methods often suffer from low training efficiency and poor action safety. To tackle these issues, we introduce a physics-informed reinforcement learning framework equipped with two modules: a Potential Field (PF) module and a Multi-Agent Multi-Level Safety (MAMLS) module. The PF module uses the Artificial Potential Field method to compute a regularization loss, adaptively integrating it into the critic's loss to enhance training efficiency. The MAMLS module formulates action safety as a constrained optimization problem, deriving safe actions by solving this optimization. Furthermore, to better address the characteristics of multi-agent collision avoidance tasks, multi-agent multi-level constraints are introduced. The results of simulations and real-world experiments showed that our physics-informed framework offers a significant improvement in terms of both the efficiency of training and safety-related metrics over advanced baseline methods.
- 1 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 33 (10577149) , 2714-2729
Hierarchical Perceptual Noise Injection for Social Media Fingerprint Privacy Protection
Billions of people share images from their daily lives on social media every day. However, their biometric information ( e.g ., fingerprints) could be easily stolen from these images. The threat of fingerprint leakage from social media has created a strong desire to anonymize shared images while maintaining image quality, since fingerprints act as a lifelong individual biometric password. To guard the fingerprint leakage, adversarial attack that involves adding imperceptible perturbations to fingerprint images have emerged as a feasible solution. However, existing works of this kind are either weak in black-box transferability or cause the images to have an unnatural appearance. Motivated by the visual perception hierarchy ( i.e ., high-level perception exploits model-shared semantics that transfer well across models while low-level perception extracts primitive stimuli that result in high visual sensitivity when a suspicious stimulus is provided), we propose FingerSafe, a hierarchical perceptual protective noise injection framework to address the above mentioned problems. For black-box transferability, we inject protective noises into the fingerprint orientation field to perturb the model-shared high-level semantics ( i.e ., fingerprint ridges). Considering visual naturalness, we suppress the low-level local contrast stimulus by regularizing the response of the Lateral Geniculate Nucleus. Our proposed FingerSafe is the first to provide feasible fingerprint protection in both digital (up to 94.12%) and realistic scenarios (Twitter and Facebook, up to 68.75%). Our code can be found at https://github.com/nlsde-safety-team/FingerSafe.
- 1 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 33 (10577149) , 2714-2729
Hierarchical Perceptual Noise Injection for Social Media Fingerprint Privacy Protection
Billions of people share images from their daily lives on social media every day. However, their biometric information ( e.g ., fingerprints) could be easily stolen from these images. The threat of fingerprint leakage from social media has created a strong desire to anonymize shared images while maintaining image quality, since fingerprints act as a lifelong individual biometric password. To guard the fingerprint leakage, adversarial attack that involves adding imperceptible perturbations to fingerprint images have emerged as a feasible solution. However, existing works of this kind are either weak in black-box transferability or cause the images to have an unnatural appearance. Motivated by the visual perception hierarchy ( i.e ., high-level perception exploits model-shared semantics that transfer well across models while low-level perception extracts primitive stimuli that result in high visual sensitivity when a suspicious stimulus is provided), we propose FingerSafe, a hierarchical perceptual protective noise injection framework to address the above mentioned problems. For black-box transferability, we inject protective noises into the fingerprint orientation field to perturb the model-shared high-level semantics ( i.e ., fingerprint ridges). Considering visual naturalness, we suppress the low-level local contrast stimulus by regularizing the response of the Lateral Geniculate Nucleus. Our proposed FingerSafe is the first to provide feasible fingerprint protection in both digital (up to 94.12%) and realistic scenarios (Twitter and Facebook, up to 68.75%). Our code can be found at https://github.com/nlsde-safety-team/FingerSafe.
- Open Access
- 20 November 2023
- Published by Darcy & Roy Press Co. Ltd. in Frontiers in Computing and Intelligent Systems
- Vol. 6 (1) , 60-65
Image Emotion Analysis Combining Attention Mechanism and Multi-level Correlation
The development of social network has brought a large amount of image information, and the research on image emotion has gradually attracted wide attention. The current image emotion analysis methods based on multi-level features simply splice the features at each level and then classify the emotions, which not only ignores the correlation between features at different levels, but also ignores the synergistic effect between global features and local features. Therefore, this paper proposes an emotion model (MAML) based on mixed attention and multi-level dependence of images, which uses spatial and channel attention mechanisms to extract local emotion region features of images. Bi-directional Long Short Term Memory network (BiLSTM) is used to establish correlation between multi-level image global features. The experimental results of MAML model on artphoto and abstract data sets prove the validity of MAML model.
- 20 November 2023
- Published by Darcy & Roy Press Co. Ltd. in Frontiers in Computing and Intelligent Systems
- Vol. 6 (1) , 60-65
Open AccessImage Emotion Analysis Combining Attention Mechanism and Multi-level Correlation
The development of social network has brought a large amount of image information, and the research on image emotion has gradually attracted wide attention. The current image emotion analysis methods based on multi-level features simply splice the features at each level and then classify the emotions, which not only ignores the correlation between features at different levels, but also ignores the synergistic effect between global features and local features. Therefore, this paper proposes an emotion model (MAML) based on mixed attention and multi-level dependence of images, which uses spatial and channel attention mechanisms to extract local emotion region features of images. Bi-directional Long Short Term Memory network (BiLSTM) is used to establish correlation between multi-level image global features. The experimental results of MAML model on artphoto and abstract data sets prove the validity of MAML model.