unit8co/darts

mc_dropout with predict_likelihood_parameters

Open

#2,105 创建于 2023年12月4日

在 GitHub 查看
 (3 评论) (0 反应) (0 负责人)Python (6,832 star) (762 fork)batch import
documentationgood first issue

描述

Hi, thanks for the prompt response regarding issue #2097 . I'm now able to utilize mc_dropout in historical_forecasts.

My current objective is to incorporate mc_dropout for obtaining epistemic uncertainty and predict_likelihood_parameters for acquiring aleatoric uncertainty. However, when setting predict_likelihood_parameters to True, I encounter an issue where I'm unable to use the sample method. Consequently, I'm limited to obtaining only one result with each call to historical_forecasts with predict_likelihood_parameters=True and mc_dropout=True.

I suspect that the solution might involve distinguishing the sample generation process by using mc_dropout and sample generation process based on the fitted distribution. Now they are using the same num_samples parameter.

贡献者指南

技术栈
python
领域
machine learningdata
议题类型
feature
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
3
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
1-2 days
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
fresh
清晰度议题是否清楚说明期望改动、验收标准和下一步。
mostly clear
前置要求
Familiarity with darts libraryUnderstanding of Monte Carlo DropoutUnderstanding of aleatoric vs epistemic uncertainty
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
35
研究方向
Investigate the implementation of `historical forecasts` in darts, specifically how `num samples` is used for both `mc dropout` and `predict likelihood parameters`. Examine the sampling logic to separate the two processes. Check the related issue #2097 for context. Consider modifying the function to allow different sample counts for each uncertainty source.