Research map / Time series / Foundation models for time series
CALDA: Improving Multi-Source Time Series Domain Adaptation With Contrastive Adversarial Learning
Garrett Wilson, Janardhan Rao Doppa, Diane J. Cook
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add domain adversary (ablation) | CALDA-Any,R · Real-world average | accuracy: 73.2 -> 80.2 | Improved |
| add domain adversary (ablation) | CALDA-XS,H · Real-world average | accuracy: 73.1 -> 79.9 | Improved |
| add pseudo-labeled target data to contrastive loss (ablation) | CALDA-Any,R · Real-world average | accuracy: 80.2 -> 76.2 | Hurt |
| add pseudo-labeled target data to contrastive loss (ablation) | CALDA-XS,H · Real-world average | accuracy: 79.9 -> 75.4 | Hurt |
| use cross-source/any-source pairs instead of within-source pairs (ablation) | CALDA · Synthetic datasets with shift | accuracy: 71.5 -> 75.0 | Improved |
| hard sampling instead of random sampling (ablation) | CALDA-Any · Real-world average | accuracy: 80.2 -> 79.7 | Hurt |
| hard sampling instead of random sampling (ablation) | CALDA-Any · Synthetic datasets with shift | accuracy: 71.5 -> 75.0 | Improved |
| hard sampling instead of random sampling (ablation) | CALDA-XS · WISDM AR | accuracy: 80.0 -> 81.4 | Improved |
| add contrastive learning plus adversary over CoDATS | CALDA-Any,R · Real-world average | accuracy: 78.6 -> 80.2 | Improved |
| add weak supervision regularizer | CALDA-Any,R · Real-world average | accuracy: 80.2 -> 83.7 | Improved |
| add weak supervision regularizer | CALDA-XS,H · WISDM AT | accuracy: 71.0 -> 83.3 | Improved |
| add label-proportion noise budget 0.4 to weak supervision | CoDATS-WS · WISDM AT | accuracy: 72.1 -> 56.3 | Hurt |
Closest papers by meaning
- Contrastive Learning for Unsupervised Domain Adaptation of Time Series (ICLR 2023 · similarity 0.89)
- Enhancing Multivariate Time-Series Domain Adaptation via Contrastive Frequency Graph Discovery and Language-Guided Adversary Alignment (AAAI 2025 · similarity 0.88)
- CASUAL: Conditional Support Alignment for Domain Adaptation with Label Shift (AAAI 2025 · similarity 0.86)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · similarity 0.86)
- Understanding and Improving Source-free Domain Adaptation from a Theoretical Perspective (CVPR 2024 · similarity 0.85)
- Link-based Contrastive Learning for One-Shot Unsupervised Domain Adaptation (CVPR 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
