Chen Gong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Chen Gong has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
- Fan Wu
Shanghai Jiao Tong University, Shanghai, China
13 shared publications - Zhenzhe Zheng
Shanghai Jiao Tong University, Shanghai, China
13 shared publications - Guihai Chen
Shanghai Jiao Tong University
11 shared publications - Yan Zhuang
Shanghai Jiao Tong University, Shanghai, China
5 shared publications - Sheng Wang
ByteDance, Hangzhou, China
4 shared publications - Yiliu Chen
ByteDance, Hangzhou, China
4 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6291-6300
Well Begun is Half Done: Efficient Initialization for Online Continual Learning via Two-Stage Model Merging
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6291-6300
Open AccessWell Begun is Half Done: Efficient Initialization for Online Continual Learning via Two-Stage Model Merging
- Open Access
- 26 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 867-884
Sparsity-Aware Scheduling for Efficient MoE-Based LLM Serving on Edge Servers
- 26 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 867-884
Open AccessSparsity-Aware Scheduling for Efficient MoE-Based LLM Serving on Edge Servers
- Open Access
- 20 July 2026
- Published by Association for Computing Machinery (ACM) in ACM SIGMETRICS Performance Evaluation Review
- Vol. 54 (1) , 67-69
Optimizing Storage Overhead of User Behavior Log for ML-embedded Mobile Apps
Machine learning (ML) models are increasingly integrated into modern mobile apps to enable personalized and intelligent services. These models typically rely on massive input features derived from historical user behaviors to capture user intents. However, as ML-driven services become more prevalent, recording necessary user behavior data imposes high storage cost on mobile apps, leading to lower system responsiveness and more app uninstalls. In this work, we present AdaLog system to improve the storage efficiency of user behavior log in ML-embedded mobile apps, without compromising model inference accuracy or latency. We identify two key inefficiencies in current industrial practices of user behavior log: (1) redundant logging of overlapping behavior data across different features and models, and (2) sparse storage caused by storing behaviors with heterogeneous attribute descriptions in a single log file. To solve these issues, AdaLog first formulates the elimination of feature-level redundant data as a maximum weighted matching problem in hypergraphs, and proposes an efficient designs an incremental update mechanism to minimize the I/O operations needed for adapting outdated behavior log. We implement a prototype of AdaLog and evaluate it in popular mobile apps in collaboration with our industry partner. Evaluations on real-world user data show that AdaLog reduces behavior log size by 19% to 44% with minimal system overhead (only 2 s latency and 15 MB memory usage), providing a more efficient data foundation for broader adoption of on-device ML.
- 20 July 2026
- Published by Association for Computing Machinery (ACM) in ACM SIGMETRICS Performance Evaluation Review
- Vol. 54 (1) , 67-69
Open AccessOptimizing Storage Overhead of User Behavior Log for ML-embedded Mobile Apps
Machine learning (ML) models are increasingly integrated into modern mobile apps to enable personalized and intelligent services. These models typically rely on massive input features derived from historical user behaviors to capture user intents. However, as ML-driven services become more prevalent, recording necessary user behavior data imposes high storage cost on mobile apps, leading to lower system responsiveness and more app uninstalls. In this work, we present AdaLog system to improve the storage efficiency of user behavior log in ML-embedded mobile apps, without compromising model inference accuracy or latency. We identify two key inefficiencies in current industrial practices of user behavior log: (1) redundant logging of overlapping behavior data across different features and models, and (2) sparse storage caused by storing behaviors with heterogeneous attribute descriptions in a single log file. To solve these issues, AdaLog first formulates the elimination of feature-level redundant data as a maximum weighted matching problem in hypergraphs, and proposes an efficient designs an incremental update mechanism to minimize the I/O operations needed for adapting outdated behavior log. We implement a prototype of AdaLog and evaluate it in popular mobile apps in collaboration with our industry partner. Evaluations on real-world user data show that AdaLog reduces behavior log size by 19% to 44% with minimal system overhead (only 2 s latency and 15 MB memory usage), providing a more efficient data foundation for broader adoption of on-device ML.
- Open Access
- 15 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 2778-2782
Efficient Video Anomaly Detection for Edge Devices via Background Feature Caching
- 15 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 2778-2782
Open AccessEfficient Video Anomaly Detection for Edge Devices via Background Feature Caching
- Open Access
- 8 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 67-69
Optimizing Storage Overhead of User Behavior Log for ML-embedded Mobile Apps
- 8 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 67-69
Open AccessOptimizing Storage Overhead of User Behavior Log for ML-embedded Mobile Apps
- Open Access
- 10 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 407-420
Optimizing Feature Extraction for On-device Model Inference with User Behavior Sequences
- 10 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 407-420
Open AccessOptimizing Feature Extraction for On-device Model Inference with User Behavior Sequences
- 3 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 4570-4588
A Device-Cloud Collaborative Data Enrichment Framework for Continual Learning on Devices
In modern mobile and edge applications, users and devices frequently encounter dynamic contexts, necessitating ondevice continual learning (CL) to ensure consistent performance of machine learning models. While existing research predominantly focused on developing lightweight CL frameworks, we identify that data scarcity is a more severe bottleneck for ondevice CL. In this work, we explore the potential of leveraging abundant cloud-side data to enrich scarce on-device data, and propose Delta, a private, efficient and effective data enrichment framework. Specifically, Delta introduces a directory dataset to decompose the data enrichment problem into device-side and cloud-side sub-problems without sharing raw sensitive data, mitigating privacy concerns and reducing communication overhead. Next, Delta designs a soft data matching strategy to effectively solve device-side sub-problem with sparse user data and an optimal data sampling scheme for cloud server to efficiently retrieve the most suitable dataset for enrichment. Further, Delta refines the data sampling scheme by jointly considering the impact of enriched data on both new and past contexts, mitigating the catastrophic forgetting issue from a new aspect. Finally, we enhance Delta’s applicability to more private and rare user contexts by integrating differential privacy and domain adaptation techniques. Comprehensive experiments across four typical edge computing tasks with varied data modalities demonstrate that Delta significantly enhances overall model accuracy by an average of 15.1%, 12.4%, 1.1% and 5.6% for visual, IMU, audio and textual tasks compared with few-shot CL. Additionally, Delta reduces communication costs by over 90% compared to federated CL, making it suitable for resource-constrained edge networks.
- 3 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 4570-4588
A Device-Cloud Collaborative Data Enrichment Framework for Continual Learning on Devices
In modern mobile and edge applications, users and devices frequently encounter dynamic contexts, necessitating ondevice continual learning (CL) to ensure consistent performance of machine learning models. While existing research predominantly focused on developing lightweight CL frameworks, we identify that data scarcity is a more severe bottleneck for ondevice CL. In this work, we explore the potential of leveraging abundant cloud-side data to enrich scarce on-device data, and propose Delta, a private, efficient and effective data enrichment framework. Specifically, Delta introduces a directory dataset to decompose the data enrichment problem into device-side and cloud-side sub-problems without sharing raw sensitive data, mitigating privacy concerns and reducing communication overhead. Next, Delta designs a soft data matching strategy to effectively solve device-side sub-problem with sparse user data and an optimal data sampling scheme for cloud server to efficiently retrieve the most suitable dataset for enrichment. Further, Delta refines the data sampling scheme by jointly considering the impact of enriched data on both new and past contexts, mitigating the catastrophic forgetting issue from a new aspect. Finally, we enhance Delta’s applicability to more private and rare user contexts by integrating differential privacy and domain adaptation techniques. Comprehensive experiments across four typical edge computing tasks with varied data modalities demonstrate that Delta significantly enhances overall model accuracy by an average of 15.1%, 12.4%, 1.1% and 5.6% for visual, IMU, audio and textual tasks compared with few-shot CL. Additionally, Delta reduces communication costs by over 90% compared to federated CL, making it suitable for resource-constrained edge networks.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 3687-3703
Resource-Efficient LLM Customization on Mobile Devices Through Proxy Submodel Tuning
Considering limited on-device resources, current practices are attempting to deploy a system-level mixture-of-experts (MoE)-based foundation LLM on a mobile device to serve multiple apps and support mobile intelligence. However, mobile apps are hard to customize their services that require fine-tuning adapters associated with the LLM using private in-app data. The difficulty arises due to both the limited on-device resources and the restricted control that apps have over the foundation LLM. To address this issue, in this work, we propose LiteMoE, a novel proxy submodel tuning framework that supports mobile apps to efficiently fine-tune customized adapters on devices using proxy submodels. The key technique behind LiteMoE is a post-training submodel extraction method, whereby without additional retraining, we can identify and reserve critical experts, match and merge moderate experts, to extract a lightweight and effective proxy submodel from the foundation LLM for a specific app. To further enhance scalability and adaptability, LiteMoE incorporates adapter reuse and continuous tuning mechanisms to handle multi-task requirements and evolving user preferences. We implemented a prototype of LiteMoE and evaluated it over various MoE-based LLMs and mobile computing tasks. The results show that with LiteMoE, mobile apps are able to fine-tune customized adapters on resource-limited devices, achieving 12.7% accuracy improvement and 6.6× memory reduction compared with operating the original foundation LLM.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 3687-3703
Resource-Efficient LLM Customization on Mobile Devices Through Proxy Submodel Tuning
Considering limited on-device resources, current practices are attempting to deploy a system-level mixture-of-experts (MoE)-based foundation LLM on a mobile device to serve multiple apps and support mobile intelligence. However, mobile apps are hard to customize their services that require fine-tuning adapters associated with the LLM using private in-app data. The difficulty arises due to both the limited on-device resources and the restricted control that apps have over the foundation LLM. To address this issue, in this work, we propose LiteMoE, a novel proxy submodel tuning framework that supports mobile apps to efficiently fine-tune customized adapters on devices using proxy submodels. The key technique behind LiteMoE is a post-training submodel extraction method, whereby without additional retraining, we can identify and reserve critical experts, match and merge moderate experts, to extract a lightweight and effective proxy submodel from the foundation LLM for a specific app. To further enhance scalability and adaptability, LiteMoE incorporates adapter reuse and continuous tuning mechanisms to handle multi-task requirements and evolving user preferences. We implemented a prototype of LiteMoE and evaluated it over various MoE-based LLMs and mobile computing tasks. The results show that with LiteMoE, mobile apps are able to fine-tune customized adapters on resource-limited devices, achieving 12.7% accuracy improvement and 6.6× memory reduction compared with operating the original foundation LLM.
- Open Access
- 2 December 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Measurement and Analysis of Computing Systems
- Vol. 9 (3) , 1-28
Optimizing Storage Overhead of User Behavior Log for ML-embedded Mobile Apps
Machine learning (ML) models are increasingly integrated into modern mobile apps to enable personalized and intelligent services. These models typically rely on rich input features derived from historical user behaviors to capture user intents. However, as ML-driven services become more prevalent, recording necessary user behavior data imposes substantial storage cost on mobile apps, leading to lower system responsiveness and more app uninstalls. To address this storage bottleneck, we present AdaLog, a lightweight and adaptive system designed to improve the storage efficiency of user behavior log in ML-embedded mobile apps, without compromising model inference accuracy or latency. We identify two key inefficiencies in current industrial practices of user behavior log: (i) redundant logging of overlapping behavior data across different features and models, and (ii) sparse storage caused by storing behaviors with heterogeneous attribute descriptions in a single log file. To solve these issues, AdaLog first formulates the elimination of feature-level redundant data as a maximum weighted matching problem in hypergraphs, and proposes a hierarchical algorithm for efficient on-device deployment. Then, AdaLog employs a virtually hashed attribute design to distribute heterogeneous behaviors into a few log files with physically dense storage. Finally, to ensure scalability to dynamic user behavior patterns, AdaLog designs an incremental update mechanism to minimize the I/O operations needed for adapting outdated behavior log. We implement a prototype of AdaLog and deploy it into popular mobile apps in collaboration with our industry partner. Evaluations on real-world user data show that AdaLog reduces behavior log size by 19% to 44% with minimal system overhead (only 2 seconds latency and 15 MB memory usage), providing a more efficient data foundation for broader adoption of on-device ML.
- 2 December 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Measurement and Analysis of Computing Systems
- Vol. 9 (3) , 1-28
Open AccessOptimizing Storage Overhead of User Behavior Log for ML-embedded Mobile Apps
Machine learning (ML) models are increasingly integrated into modern mobile apps to enable personalized and intelligent services. These models typically rely on rich input features derived from historical user behaviors to capture user intents. However, as ML-driven services become more prevalent, recording necessary user behavior data imposes substantial storage cost on mobile apps, leading to lower system responsiveness and more app uninstalls. To address this storage bottleneck, we present AdaLog, a lightweight and adaptive system designed to improve the storage efficiency of user behavior log in ML-embedded mobile apps, without compromising model inference accuracy or latency. We identify two key inefficiencies in current industrial practices of user behavior log: (i) redundant logging of overlapping behavior data across different features and models, and (ii) sparse storage caused by storing behaviors with heterogeneous attribute descriptions in a single log file. To solve these issues, AdaLog first formulates the elimination of feature-level redundant data as a maximum weighted matching problem in hypergraphs, and proposes a hierarchical algorithm for efficient on-device deployment. Then, AdaLog employs a virtually hashed attribute design to distribute heterogeneous behaviors into a few log files with physically dense storage. Finally, to ensure scalability to dynamic user behavior patterns, AdaLog designs an incremental update mechanism to minimize the I/O operations needed for adapting outdated behavior log. We implement a prototype of AdaLog and deploy it into popular mobile apps in collaboration with our industry partner. Evaluations on real-world user data show that AdaLog reduces behavior log size by 19% to 44% with minimal system overhead (only 2 seconds latency and 15 MB memory usage), providing a more efficient data foundation for broader adoption of on-device ML.
- Open Access
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 673-684
A Two-Stage Data Selection Framework for Data-Efficient Model Training on Edge Devices
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 673-684
Open AccessA Two-Stage Data Selection Framework for Data-Efficient Model Training on Edge Devices