A5109781363- Privacy-Preserving Technologies in Data
- Tunneling and Rock Mechanics
- Topic Modeling
- Innovative Human-Technology Interaction
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Robotics and Sensor-Based Localization
- Software Engineering Research
- Data Visualization and Analytics
- Infrastructure Maintenance and Monitoring
- Complexity and Algorithms in Graphs
- Image and Object Detection Techniques
- Cryptography and Data Security
Shenyang Institute of Automation
2025
Chinese Academy of Sciences
2025
Shanghai University
2024
Multi-temporal remote sensing image matching is crucial for tasks such as drone positioning under satellite-denial conditions, natural disaster monitoring, and land-cover-change detection. However, the significant differences between multi-temporal images often lead to reduced accuracy or even failure of most methods in these scenarios. To address this challenge, paper introduces a Discrete Cosine Transform (DCT) frequency analysis tailored characteristics images, proposes network that...
The state-of-the-art for training on-device language models mobile keyboard applications combines federated learning (FL) with differential privacy (DP) via the DP-Follow-the-Regularized-Leader (DP-FTRL) algorithm. Two variants of DP-FTRL are used in practice, tree aggregation and matrix factorization. However, suffers from significantly suboptimal privacy/utility tradeoffs, while mechanisms require expensive optimization parameterized by hard-to-estimate-in-advance constants, high runtime...
Visual programming provides beginner-level programmers with a coding-free experience to build their customized pipelines. Existing systems require users pipeline entirely from scratch, implying that novice need set up and link appropriate nodes all by themselves, starting blank workspace. We present InstructPipe, an AI assistant enables start prototyping machine learning (ML) pipelines text instructions. designed two LLM modules code interpreter execute our solution. generate pseudocode of...
We train language models (LMs) with federated learning (FL) and differential privacy (DP) in the Google Keyboard (Gboard). apply DP-Follow-the-Regularized-Leader (DP-FTRL)~\citep{kairouz21b} algorithm to achieve meaningfully formal DP guarantees without requiring uniform sampling of client devices. To provide favorable privacy-utility trade-offs, we introduce a new participation criterion discuss implication its configuration large scale systems. show how quantile-based clip...