TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Visualization of a dense grid search over neural network hyperparameters

3 点作者 jchook超过 1 年前

2 条评论

westurner超过 1 年前
AutoML and [Partially] automated feature engineering have hyperparameters too. Some algorithms have no hyperparameters. And, OT did a complete grid search instead of a PSO or gradient descent, for which there are also adversarial cases.<p>Featuretools supports Dask EntitySets for larger-than-RAM feature matrices, or pandas on multiple cores: <a href="https:&#x2F;&#x2F;featuretools.alteryx.com&#x2F;en&#x2F;stable&#x2F;guides&#x2F;using_dask_entitysets.html" rel="nofollow">https:&#x2F;&#x2F;featuretools.alteryx.com&#x2F;en&#x2F;stable&#x2F;guides&#x2F;using_dask...</a><p>&quot;Hyperparameter optimization with Dask&quot;: <a href="https:&#x2F;&#x2F;examples.dask.org&#x2F;machine-learning&#x2F;hyperparam-opt.html" rel="nofollow">https:&#x2F;&#x2F;examples.dask.org&#x2F;machine-learning&#x2F;hyperparam-opt.ht...</a> :<p>&gt; <i>HyperbandSearchCV is Dask-ML’s meta-estimator to find the best hyperparameters. It can be used as an alternative to RandomizedSearchCV to find similar hyper-parameters in less time by not wasting time on hyper-parameters that are not promising. Specifically, it is almost guaranteed that it will find high performing models with minimal training.</i><p>Note that e.g. TabPFN is faster or converges more quickly than xgboost and other gradient boosting <i>with hyperparameter</i> methods: <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=37269376#37274671">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=37269376#37274671</a><p>&quot;Stochastic gradient descent written in SQL&quot; (2023) <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=35063522">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=35063522</a> :<p>&gt; <i>What are some</i> adversarial <i>cases for gradient descent, and&#x2F;or what sort of e.g. DVC.org or W3C PROV provenance information should be tracked for a production ML workflow?</i>
westurner超过 1 年前
<a href="https:&#x2F;&#x2F;x.com&#x2F;jaschasd&#x2F;status&#x2F;1756930247633825827" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;jaschasd&#x2F;status&#x2F;1756930247633825827</a> :<p>&gt; <i>So it shouldn&#x27;t (post-hoc) be a surprise that hyperparameter landscapes are fractal. This is a general phenomenon: in these panes we see fractal hyperparameter landscapes for every neural network configuration I tried, including deep linear networks.</i>