Sung-Ju Lee, Seungjoo Lee, Thanh-Long Le, Jaemin Shin
We lifted 18 functions out of this paper's own repositories and ran 9 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| seungjoo-ai/FLFL-NeurIPS24 | — | 8 of 16 |
| diaoenmao/semifl-semi-supervised-federated-learning-for-unlabeled-clients-with-alternate-training | reimplementation | 1 of 2 |
| Function | Status | Where it lives |
|---|---|---|
| ASAM | Ran | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("dafb8299d9942e92") |
| SAM | Ran | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("5937ee54bec0a47e") |
| ce_loss | Ran | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("ac50b71552df6ea1") |
| compute_probs | Ran | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("3b8a3109e8983ec9") |
| gen_ulb_targets | Ran | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("4cbb9c5dc8b55431") |
| kldiv_with_mask | Ran | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("6a242dc4c1f7039f") |
| make_controls | Ran | diaoenmao/semifl-semi-supervised-federated-learning-for-unlabeled-clients-with-alternate-training/src/process.py code served (permissive licence) · get_code("d817b1bf3eb37bf5") |
| smooth_targets | Ran | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("90f8ce204153de70") |
| to_device | Ran | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("7cebb43e567583f6") |
| AlgorithmBase | Not yet run | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("b9b91fbb32f08903") |
| FLFL | Not yet run | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("967d7bdbfda20315") |
| consistency_loss | Not yet run | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("129da2b3defd51b4") |
| get_model | Not yet run | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("33d367de0331ee93") |
| get_sam_optimizer | Not yet run | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("2b9135b26d4bfac9") |
| make_control_list | Not yet run | diaoenmao/semifl-semi-supervised-federated-learning-for-unlabeled-clients-with-alternate-training/src/process.py code served (permissive licence) · get_code("89204826fa1414cd") |
| make_data_loader | Not yet run | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("3e5e0e1a3245fac2") |
| recur | Not yet run | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("882ae83b71de025e") |
| seed_worker | Not yet run | seungjoo-ai/FLFL-NeurIPS24/src/algorithm/flfl.py code served (permissive licence) · get_code("5a53903788b0993f") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Federated Learning (FL) is a distributed machine learning framework that trains accurate global models while preserving clients' privacy-sensitive data. However, most FL approaches assume that clients possess labeled data, which is often not the case in practice. Federated Semi-Supervised Learning (FSSL) addresses this label deficiency problem, targeting situations where only the server has a small amount of labeled data while clients do not. However, a significant performance gap exists between Centralized Semi-Supervised Learning (SSL) and FSSL. This gap arises from confirmation bias, which is more pronounced in FSSL due to multiple local training epochs and the separation of labeled and unlabeled data. We propose (FL) 2 , a robust training method for unlabeled clients using sharpness-aware consistency regularization. We show that regularizing the original pseudo-labeling loss is suboptimal, and hence we carefully select unlabeled samples for regularization. We further introduce client-specific adaptive thresholding and learning statusaware aggregation to adjust the training process based on the learning progress of each client. Our experiments on three benchmark datasets demonstrate that our approach significantly improves performance and bridges the gap with SSL, particularly in scenarios with scarce labeled data. The source code is available at https://github.com/seungjoo-ai/FLFL-NeurIPS24
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2410.23227")
get_code_for_paper("2410.23227")
have("2410.23227")
Connect an agent — have() is free.