From 7afea449e9c978cf357bfa1a071e7312287e0bfd Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Mart=C3=ADn=20Santill=C3=A1n=20Cooper?= Date: Tue, 5 Dec 2023 10:40:50 -0300 Subject: [PATCH] Active learning docs update --- docs/docs/dev/active_learning.md | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/docs/docs/dev/active_learning.md b/docs/docs/dev/active_learning.md index 62274f3..fe204f6 100644 --- a/docs/docs/dev/active_learning.md +++ b/docs/docs/dev/active_learning.md @@ -6,13 +6,29 @@ Label Sleuth guides the users through the labeling process by suggesting a ranke Label Sleuth currently supports the following active learning strategies: +::::{tab-set} +:::{tab-item} Binary mode + +The employed active learning strategy can be configured by setting the value of the `binary_flow.active_learning_strategy` parameter in the system's [configuration file](configuration.md). + | Active learning strategy | Description | |---|---| | `RANDOM` | Randomly sample from unlabeled data (active learning baseline) | | `HARD_MINING`
default | Uncertainty sampling/least confidence ([Lewis and Gale 1994](https://arxiv.org/abs/cmp-lg/9407020)) | | `RETROSPECTIVE` | Select top scored instances by the model ([Ein-Dor et al. 2019](https://arxiv.org/abs/1911.10763)) | +| `COMBINED_RETROSPECTIVE_HM` | a combination between the previous two strategies | + +::: +:::{tab-item} Multiclass mode + +The employed active learning strategy can be configured by setting the value of the `multiclass_flow.active_learning_strategy` parameter in the system's [configuration file](configuration.md). + +| Active learning strategy | Description | +|---|---| +| `MULTICLASS_ENTROPY`
default | Uncertainty sampling/least confidence ([Lewis and Gale 1994](https://arxiv.org/abs/cmp-lg/9407020)) | -The employed active learning strategy can be configured by setting the value of the `active_learning_strategy` parameter in the system's [configuration file](configuration.md). +::: +:::: ## Implement a new active learning strategy