Mingtao Ji
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Mingtao Ji has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
- Zhuzhong Qian
Nanjing University
6 shared publications - Baoliu Ye
Nanjing University
3 shared publications - Lei Jiao
University of Oregon
3 shared publications Tuo Cao
3 shared publicationsBaoliu Ye
2 shared publications- Hehan Zhao
State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China
2 shared publications
Latest Publications
- 25 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-16
Feature Map Compression for Split Learning in Mobile Wireless Environment: Theory and Practice
Split Learning (SL) enables collaborative model training by distributing computation between resource constrained mobile edge devices and a central server. In this distributed framework, mobile edge devices compute the early layers of the model and transmit feature maps to the server. However, due to practical mobile wireless bandwidth limitations, these feature maps should be compressed before transmission. In this paper, we propose a lightweight Mask-Encoded Spar sification (MS) compressor that enhances Top-k sparsification with a narrow-bit mask, significantly reducing compression error while improving computational and communication efficiency. Our theoretical analyses reveal that feature map compression generally induces estimator bias in the gradients. Furthermore, we establish the relationships between compression error and convergence, as well as between the choice of cutlayer and output error. The first relationship demonstrates that reducing compression error facilitates convergence, confirming the advantage of MS, whereas the second provides theoretical guidance for determining the cutlayer, indicating that earlier layers are more sensitive to compression errors. Extensive experiments on vision and language models demonstrate that our method outperforms existing compression techniques in terms of compression error and model performance, achieving up to a 21.38× speedup in communication efficiency under wireless environments.
- 25 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-16
Feature Map Compression for Split Learning in Mobile Wireless Environment: Theory and Practice
Split Learning (SL) enables collaborative model training by distributing computation between resource constrained mobile edge devices and a central server. In this distributed framework, mobile edge devices compute the early layers of the model and transmit feature maps to the server. However, due to practical mobile wireless bandwidth limitations, these feature maps should be compressed before transmission. In this paper, we propose a lightweight Mask-Encoded Spar sification (MS) compressor that enhances Top-k sparsification with a narrow-bit mask, significantly reducing compression error while improving computational and communication efficiency. Our theoretical analyses reveal that feature map compression generally induces estimator bias in the gradients. Furthermore, we establish the relationships between compression error and convergence, as well as between the choice of cutlayer and output error. The first relationship demonstrates that reducing compression error facilitates convergence, confirming the advantage of MS, whereas the second provides theoretical guidance for determining the cutlayer, indicating that earlier layers are more sensitive to compression errors. Extensive experiments on vision and language models demonstrate that our method outperforms existing compression techniques in terms of compression error and model performance, achieving up to a 21.38× speedup in communication efficiency under wireless environments.
- 26 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 24 (9) , 7947-7964
Edge AI Inference as a Service via Dynamic Resources From Repeated Auctions
To enable edge AI providers to recruit edge devices and use them to deploy AI models and provision inference services, we conduct a comprehensive mathematical and algorithmic study on a novel incentive and optimization mechanism based on repeated auctions. We first model and formulate a time-cumulative social cost optimization problem to capture the challenges of the trade-off between cost and accuracy, the dependency between adjacent auctions, and the need of achieving desired economic properties. Then, to solve this intractable non-linear integer program in an online manner, we design a set of polynomial-time algorithms that work together. Our approach dynamically chooses and switches winning bids under careful control, incorporates online learning to overcome posterior inference accuracy and workload queue dynamics, and leverages randomization to strategically convert fractional decisions of model placement and query dispatch into integers. We also allocate payments to meet the necessary and sufficient conditions for the desired economic properties. Further, we rigorously prove the constant competitive ratio, the sub-linear regret and fit, and the truthfulness and individual rationality for our proposed approach. Finally, through extensive experiments using real devices, AI models, and data traces, we have validated the substantial advantages of our proposed approach compared to the baselines and the state-of-the-art methods.
- 26 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 24 (9) , 7947-7964
Edge AI Inference as a Service via Dynamic Resources From Repeated Auctions
To enable edge AI providers to recruit edge devices and use them to deploy AI models and provision inference services, we conduct a comprehensive mathematical and algorithmic study on a novel incentive and optimization mechanism based on repeated auctions. We first model and formulate a time-cumulative social cost optimization problem to capture the challenges of the trade-off between cost and accuracy, the dependency between adjacent auctions, and the need of achieving desired economic properties. Then, to solve this intractable non-linear integer program in an online manner, we design a set of polynomial-time algorithms that work together. Our approach dynamically chooses and switches winning bids under careful control, incorporates online learning to overcome posterior inference accuracy and workload queue dynamics, and leverages randomization to strategically convert fractional decisions of model placement and query dispatch into integers. We also allocate payments to meet the necessary and sufficient conditions for the desired economic properties. Further, we rigorously prove the constant competitive ratio, the sub-linear regret and fit, and the truthfulness and individual rationality for our proposed approach. Finally, through extensive experiments using real devices, AI models, and data traces, we have validated the substantial advantages of our proposed approach compared to the baselines and the state-of-the-art methods.
- 7 January 2025
- Published by Elsevier in Computer Networks
- Vol. 258
Towards strong continuous consistency in edge-assisted VR-SGs: Delay-differences sensitive online task redistribution
- 7 January 2025
- Published by Elsevier in Computer Networks
- Vol. 258
Towards strong continuous consistency in edge-assisted VR-SGs: Delay-differences sensitive online task redistribution
- 14 November 2024
- Published by Elsevier in Computer Networks
- Vol. 256
CPN meets learning: Online scheduling for inference service in Computing Power Network
- 14 November 2024
- Published by Elsevier in Computer Networks
- Vol. 256
CPN meets learning: Online scheduling for inference service in Computing Power Network
- 24 January 2024
- Published by Elsevier in Computer Networks
- Vol. 241
INTaaS: Provisioning In-band Network Telemetry as a service via online learning
- 24 January 2024
- Published by Elsevier in Computer Networks
- Vol. 241
INTaaS: Provisioning In-band Network Telemetry as a service via online learning
- 19 December 2023
- Published by Elsevier in Computer Networks
- Vol. 239
Walking on two legs: Joint service placement and computation configuration for provisioning containerized services at edges
- 19 December 2023
- Published by Elsevier in Computer Networks
- Vol. 239
Walking on two legs: Joint service placement and computation configuration for provisioning containerized services at edges
- Open Access
- 24 September 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Efficient Inference Task Deployment for Enhanced Accuracy at the Edge
With the development of the Internet of Things and Artificial Intelligence, deploying the inference tasks at the network edge is widely adopted by operators so as to provide intelligent services. However, due to limited edge resources and the relatively high resource consumption of the model, deploying the appropriate model to ensure good inference accuracy can be challenging. In this work, we model this problem whose aim is to maximize overall efficiency and benefit by selecting AI models pre-trained on cloud servers to be deployed to edge servers. However, the benefit of a pre-trained model is unable to predict due to the dynamic stochastic inputs of users, which can only be disclosed to us after the inference task is executed. Therefore, we improve on the classic algorithm and propose an algorithm for selecting pre-trained models for edge servers based on the multi-armed bandit. With rigorous proof, we prove that regret grows sub-linearly along with time. We also conduct experiments and show the superiority of our algorithm compared to the state-of-the-art algorithms.
- 24 September 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Open AccessEfficient Inference Task Deployment for Enhanced Accuracy at the Edge
With the development of the Internet of Things and Artificial Intelligence, deploying the inference tasks at the network edge is widely adopted by operators so as to provide intelligent services. However, due to limited edge resources and the relatively high resource consumption of the model, deploying the appropriate model to ensure good inference accuracy can be challenging. In this work, we model this problem whose aim is to maximize overall efficiency and benefit by selecting AI models pre-trained on cloud servers to be deployed to edge servers. However, the benefit of a pre-trained model is unable to predict due to the dynamic stochastic inputs of users, which can only be disclosed to us after the inference task is executed. Therefore, we improve on the classic algorithm and propose an algorithm for selecting pre-trained models for edge servers based on the multi-armed bandit. With rigorous proof, we prove that regret grows sub-linearly along with time. We also conduct experiments and show the superiority of our algorithm compared to the state-of-the-art algorithms.
- 14 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ACM Transactions on Networking
- Vol. 31 (5) , 2313-2328
Scheduling In-Band Network Telemetry With Convergence-Preserving Federated Learning
Conducting federated learning across distributed sites with In-Band Network Telemetry (INT) based data collection faces critical challenges, including control decisions of different frequencies, convergence of the models being trained, and resource provisioning coupled over time. To study this problem, we formulate a non-linear mixed-integer program to optimize the long-term INT overhead, resource cost, and federated learning cost. We then design polynomial-time online algorithms to solve this problem with only observable inputs on the fly, featuring laziness-aware resource adaption, online-learning-based INT flow selection and model aggregation control, as well as expectation-preserving randomized dependent rounding. We rigorously prove the parameterized-constant competitive ratio of our approach against the offline optimum, and the time-averaged constraint violation that vanishes in the long run. With extensive trace-driven evaluations, we confirm the superiority of our approach over other alternative approaches for reducing total cost and the efficacy of our trained models for solving real machine learning problems, reducing the real-time cost by $34\%$ on average.
- 14 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ACM Transactions on Networking
- Vol. 31 (5) , 2313-2328
Scheduling In-Band Network Telemetry With Convergence-Preserving Federated Learning
Conducting federated learning across distributed sites with In-Band Network Telemetry (INT) based data collection faces critical challenges, including control decisions of different frequencies, convergence of the models being trained, and resource provisioning coupled over time. To study this problem, we formulate a non-linear mixed-integer program to optimize the long-term INT overhead, resource cost, and federated learning cost. We then design polynomial-time online algorithms to solve this problem with only observable inputs on the fly, featuring laziness-aware resource adaption, online-learning-based INT flow selection and model aggregation control, as well as expectation-preserving randomized dependent rounding. We rigorously prove the parameterized-constant competitive ratio of our approach against the offline optimum, and the time-averaged constraint violation that vanishes in the long run. With extensive trace-driven evaluations, we confirm the superiority of our approach over other alternative approaches for reducing total cost and the efficacy of our trained models for solving real machine learning problems, reducing the real-time cost by $34\%$ on average.