Di Niu
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
Di Niu has a total of 152 co-authors. Here are the top co-authors by number of shared publications:
- Mingjun Zhao
University of Alberta
9 shared publications - Baochun Li
University of Toronto
7 shared publications - Chenglin Li
University of Alberta, Edmonton, AB, Canada
6 shared publications - Linglong Kong
University of Alberta
6 shared publications Weidong Guo
Platform and Content Group Tencent
6 shared publicationsYu Xu
Platform and Content Group Tencent
5 shared publications
Latest Publications
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 8683-8694
PAUSE: A User-Centric Benchmark for Personal AI Assistants in Unified Service Environments
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 8683-8694
Open AccessPAUSE: A User-Centric Benchmark for Personal AI Assistants in Unified Service Environments
- 13 February 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 339
CTRL: Continuous-time representation learning on temporal heterogeneous information network
- 13 February 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 339
CTRL: Continuous-time representation learning on temporal heterogeneous information network
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 56 (3) , 2132-2142
A Real-Time Fault-Tolerant Constrained Control Scheme for Elastic Cable-Driven Robots
This article proposes an integrated fault-tolerant control architecture for unidirectional constrained control of elastic cable-driven parallel robots (CDPRs). In CDPRs, the cable tension can only be positive, requiring unidirectional control effort. In the proposed structure, the positive tension distribution (PTD) of the elastic CDPR is ensured using a prevention function. Compared to existing methods for elastic CDPRs, this new technique has significantly lower computation time, which is essential for real-time applications. Moreover, two adaptive observers are utilized to tackle undesirable events, particularly partial failures. The stability analysis of the closed-loop system is provided using Lyapunov’s second method. The simulation study shows that the proposed method improves the tracking accuracy and is significantly faster than conventional methods. The effectiveness of the proposed method is evaluated in the presence of model uncertainties and various faulty conditions.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 56 (3) , 2132-2142
A Real-Time Fault-Tolerant Constrained Control Scheme for Elastic Cable-Driven Robots
This article proposes an integrated fault-tolerant control architecture for unidirectional constrained control of elastic cable-driven parallel robots (CDPRs). In CDPRs, the cable tension can only be positive, requiring unidirectional control effort. In the proposed structure, the positive tension distribution (PTD) of the elastic CDPR is ensured using a prevention function. Compared to existing methods for elastic CDPRs, this new technique has significantly lower computation time, which is essential for real-time applications. Moreover, two adaptive observers are utilized to tackle undesirable events, particularly partial failures. The stability analysis of the closed-loop system is provided using Lyapunov’s second method. The simulation study shows that the proposed method improves the tracking accuracy and is significantly faster than conventional methods. The effectiveness of the proposed method is evaluated in the presence of model uncertainties and various faulty conditions.
- Open Access
rSIM: Incentivizing Reasoning Capabilities of LLMs via Reinforced Strategy Injection
Large language models (LLMs) are post-trained through reinforcement learning (RL) to evolve into Reasoning Language Models (RLMs), where the hallmark of this advanced reasoning is ``aha'' moments when they start to perform strategies, such as self-reflection and deep thinking, within chain of thoughts (CoTs). Motivated by this, this paper proposes a novel reinforced strategy injection mechanism (rSIM), that enables any LLM to become an RLM by employing a small planner to guide the LLM's CoT through the adaptive injection of reasoning strategies. To achieve this, the planner (leader agent) is jointly trained with an LLM (follower agent) using multi-agent RL (MARL), based on a leader-follower framework and straightforward rule-based rewards. Experimental results show that rSIM enables Qwen2.5-0.5B to become an RLM and significantly outperform Qwen2.5-14B. Moreover, the planner is generalizable: it only needs to be trained once and can be applied as a plug-in to substantially improve the reasoning capabilities of existing LLMs. In addition, the planner supports continual learning across various tasks, allowing its planning abilities to gradually improve and generalize to a wider range of problems.
Open AccessrSIM: Incentivizing Reasoning Capabilities of LLMs via Reinforced Strategy Injection
Large language models (LLMs) are post-trained through reinforcement learning (RL) to evolve into Reasoning Language Models (RLMs), where the hallmark of this advanced reasoning is ``aha'' moments when they start to perform strategies, such as self-reflection and deep thinking, within chain of thoughts (CoTs). Motivated by this, this paper proposes a novel reinforced strategy injection mechanism (rSIM), that enables any LLM to become an RLM by employing a small planner to guide the LLM's CoT through the adaptive injection of reasoning strategies. To achieve this, the planner (leader agent) is jointly trained with an LLM (follower agent) using multi-agent RL (MARL), based on a leader-follower framework and straightforward rule-based rewards. Experimental results show that rSIM enables Qwen2.5-0.5B to become an RLM and significantly outperform Qwen2.5-14B. Moreover, the planner is generalizable: it only needs to be trained once and can be applied as a plug-in to substantially improve the reasoning capabilities of existing LLMs. In addition, the planner supports continual learning across various tasks, allowing its planning abilities to gradually improve and generalize to a wider range of problems.
- Open Access
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 260-269
HyperGenFL: Hypernetwork-Generated Model Aggregation in Federated Learning
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 260-269
Open AccessHyperGenFL: Hypernetwork-Generated Model Aggregation in Federated Learning
- Open Access
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 3773-3782
Enhancing and Assessing Instruction-Following with Fine-Grained Instruction Variants
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 3773-3782
Open AccessEnhancing and Assessing Instruction-Following with Fine-Grained Instruction Variants
- Open Access
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1491-1504
TransFusion: End-to-End Transformer Acceleration via Graph Fusion and Pipelining
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1491-1504
Open AccessTransFusion: End-to-End Transformer Acceleration via Graph Fusion and Pipelining
- Open Access
- 6 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 142007-142014
Machine Learning-Enabled Fast Prediction of GGNMOS Performance and Inverse Design for Electrostatic Discharge Applications
Electrostatic discharge (ESD) protection is generally required in integrated circuit (IC) chips. The grounded-gate n-channel metal-oxide-semiconductor (GGNMOS) is a popular device for ESD protection in circuit design. However, the design optimization of GGNMOS for every circuit is carried out using the trail-and-error method and requiring iterative device simulations. The device simulation is usually performed using technology computer-aided design (TCAD) software and is time-consuming. To address this issue, we developed machine learning models for fast prediction of GGNMOS performance and inverse design of its structure according to performance metrics. Here, we generated data for machine learning using Sentaurus TCAD. We applied AutoML and weight-sharing deep neural networks to predict current-voltage (I-V) characteristics of GGNMOS and extract performance metrics: triggering point (It1, Vt1), holding point (Ih, Vh), thermal breakdown point (It2, Vt2), and discharge rsistance Ron. Additionally, Bayesian optimization was employed for inverse design, allowing rapid identification of optimal structural parameters for desired performance metrics. This approach significantly accelerates the ESD design process, minimizing dependency on costly and time-consuming TCAD simulations. Our work represents an advancement in design of electronic devices and circuits using machine learning.
- 6 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 142007-142014
Open AccessMachine Learning-Enabled Fast Prediction of GGNMOS Performance and Inverse Design for Electrostatic Discharge Applications
Electrostatic discharge (ESD) protection is generally required in integrated circuit (IC) chips. The grounded-gate n-channel metal-oxide-semiconductor (GGNMOS) is a popular device for ESD protection in circuit design. However, the design optimization of GGNMOS for every circuit is carried out using the trail-and-error method and requiring iterative device simulations. The device simulation is usually performed using technology computer-aided design (TCAD) software and is time-consuming. To address this issue, we developed machine learning models for fast prediction of GGNMOS performance and inverse design of its structure according to performance metrics. Here, we generated data for machine learning using Sentaurus TCAD. We applied AutoML and weight-sharing deep neural networks to predict current-voltage (I-V) characteristics of GGNMOS and extract performance metrics: triggering point (It1, Vt1), holding point (Ih, Vh), thermal breakdown point (It2, Vt2), and discharge rsistance Ron. Additionally, Bayesian optimization was employed for inverse design, allowing rapid identification of optimal structural parameters for desired performance metrics. This approach significantly accelerates the ESD design process, minimizing dependency on costly and time-consuming TCAD simulations. Our work represents an advancement in design of electronic devices and circuits using machine learning.
- Open Access
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 5640-5651
MethaneS2CM: A Dataset for Multispectral Deep Methane Emission Detection
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 5640-5651
Open AccessMethaneS2CM: A Dataset for Multispectral Deep Methane Emission Detection
- Open Access
Grounding Degradations in Natural Language for All-In-One Video Restoration
In this work, we propose an all-in-one video restoration framework that grounds degradation-aware semantic context of video frames in natural language via foundation models, offering interpretable and flexible guidance. Unlike prior art, our method assumes no degradation knowledge in train or test time and learns an approximation to the grounded knowledge such that the foundation model can be safely disentangled during inference adding no extra cost. Further, we call for standardization of benchmarks in all-in-one video restoration, and propose two benchmarks in multi-degradation setting, three-task (3D) and four-task (4D), and two time-varying composite degradation benchmarks; one of the latter being our proposed dataset with varying snow intensity, simulating how weather degradations affect videos naturally. We compare our method with prior works and report state-of-the-art performance on all benchmarks.
Open AccessGrounding Degradations in Natural Language for All-In-One Video Restoration
In this work, we propose an all-in-one video restoration framework that grounds degradation-aware semantic context of video frames in natural language via foundation models, offering interpretable and flexible guidance. Unlike prior art, our method assumes no degradation knowledge in train or test time and learns an approximation to the grounded knowledge such that the foundation model can be safely disentangled during inference adding no extra cost. Further, we call for standardization of benchmarks in all-in-one video restoration, and propose two benchmarks in multi-degradation setting, three-task (3D) and four-task (4D), and two time-varying composite degradation benchmarks; one of the latter being our proposed dataset with varying snow intensity, simulating how weather degradations affect videos naturally. We compare our method with prior works and report state-of-the-art performance on all benchmarks.