tilldream 发表于 2019-1-13 19:41:48

2018年最实用机器学习项目Top 6(附开源链接)

<p style="max-width: 100%; min-height: 1em; letter-spacing: 0.544px; line-height: 3em; text-align: center; box-sizing: border-box !important; overflow-wrap: break-word !important;"></p><p style="max-width: 100%; min-height: 1em; letter-spacing: 0.544px; line-height: 3em; text-align: center; box-sizing: border-box !important; overflow-wrap: break-word !important;"><span style="font-size: 15px; letter-spacing: 0.5px;"><br></span></p><p style="max-width: 100%; min-height: 1em; letter-spacing: 0.544px; line-height: 3em; text-align: center; box-sizing: border-box !important; overflow-wrap: break-word !important;"><span style="font-size: 15px; letter-spacing: 0.5px;">过去一年,是人工智能和机器学习蓬勃发展的一年。许多高影响力的机器学习应用被开发出来,特别是在医疗保健、金融、语音识别、增强现实以及更复杂的3D和视频应用中。</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">我们已经看到了更多的应用驱动研究,而不是理论研究。虽然这些研究有着一些不足,但当前的确产生了巨大的积极影响,也促成了很多可以迅速商业化的新研发。这一趋势也在机器学习的大部分开源项目中得到了强烈反映。</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">让我们来看看过去一年中前6大最实用的机器学习项目。这些项目发布了代码和数据集,允许个别开发人员和小型团队学习并能实现立即创造价值。它们可能不是理论上最具开创性的作品,但却非常适用且实用。</span></p><p><br></p><section class="" style="max-width: 100%; box-sizing: border-box; letter-spacing: 0.544px; border-width: 0px; border-style: initial; border-color: initial; clear: both; overflow-wrap: break-word !important;"><section class="" style="padding: 8px;max-width: 100%;box-sizing: border-box;border-left: 6px solid rgb(255, 202, 0);border-top-color: rgb(255, 202, 0);border-right-color: rgb(255, 202, 0);border-bottom-color: rgb(255, 202, 0);font-size: 18px;line-height: 1.4;font-family: inherit;font-weight: bold;text-decoration: inherit;color: rgb(10, 10, 10);overflow-wrap: break-word !important;"><section style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;letter-spacing: 1px;box-sizing: border-box !important;overflow-wrap: break-word !important;">Fast.ai——易于使用,流程简便</span></section></section></section><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="text-align: center;"></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">Github链接:</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">https://github.com/fastai/fastai?utm_source=mybridge&amp;utm_medium=blog&amp;utm_campaign=read_more</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">项目链接:https://docs.fast.ai/</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">Fast.ai库的编写是为了简化训练快速准确的神经网络。它去掉了在实践中实施深度神经网络可能带来的所有细节工作。</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">Fast.ai非常易于使用,并且设计成符合开发者的应用程序来构建思维模式。最初这个开源项目是为Fast.ai课程的学生创建的,该库以简洁易懂的方式编写在易于使用的Pytorch库上。</span></p><p><br></p><section class="" style="max-width: 100%; box-sizing: border-box; letter-spacing: 0.544px; border-width: 0px; border-style: initial; border-color: initial; clear: both; overflow-wrap: break-word !important;"><section class="" style="padding: 8px;max-width: 100%;box-sizing: border-box;border-left: 6px solid rgb(255, 202, 0);border-top-color: rgb(255, 202, 0);border-right-color: rgb(255, 202, 0);border-bottom-color: rgb(255, 202, 0);font-size: 18px;line-height: 1.4;font-family: inherit;font-weight: bold;text-decoration: inherit;color: rgb(10, 10, 10);overflow-wrap: break-word !important;"><section style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;letter-spacing: 1px;box-sizing: border-box !important;overflow-wrap: break-word !important;">Detectron——Facebook AI出品</span></section></section></section><p><br></p><p style="margin-right: 8px; margin-left: 8px;"></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">github链接:</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">https://github.com/facebookresearch/Detectron?utm_source=mybridge&amp;utm_medium=blog&amp;utm_campaign=read_more</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">Detectron是Facebook AI用于物体检测和实例分割研究所创建的研究平台,用Caffe2进行编写。它包含各种目标检测算法的实现,包括:</span></p><p><br></p><ul class=" list-paddingleft-2"><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">Mask R-CNN::使用更快的R-CNN结构的目标检测和实例分割;</span></p></li><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">RetinaNet:一个基于特征金字塔的网络,具有独特的Focal Loss来处理难题;</span></p></li><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">更快的R-CNN:目标检测网络最常见的结构</span></p></li></ul><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">所有网络都可以使用以下几种可选的分类主干之一:</span></p><p><br></p><ul class=" list-paddingleft-2"><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">ResNeXt {50101152}(https://arxiv.org/abs/1611.05431)</span></p></li><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">RESNET {50101152}(https://arxiv.org/abs/1512.03385)</span></p></li><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">特征金字塔网络(使用ResNet / ResNeXt)(https://arxiv.org/abs/1612.03144)</span></p></li><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">VGG16(https://arxiv.org/abs/1409.1556)</span></p></li></ul><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">更重要的是,以上都带有COCO数据集上的预训练模型,因此开发者可以立即使用它们。</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><section class="" style="max-width: 100%; box-sizing: border-box; letter-spacing: 0.544px; border-width: 0px; border-style: initial; border-color: initial; clear: both; overflow-wrap: break-word !important;"><section class="" style="padding: 8px;max-width: 100%;box-sizing: border-box;border-left: 6px solid rgb(255, 202, 0);border-top-color: rgb(255, 202, 0);border-right-color: rgb(255, 202, 0);border-bottom-color: rgb(255, 202, 0);font-size: 18px;line-height: 1.4;font-family: inherit;font-weight: bold;text-decoration: inherit;color: rgb(10, 10, 10);overflow-wrap: break-word !important;"><section style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;letter-spacing: 1px;box-sizing: border-box !important;overflow-wrap: break-word !important;">FastText——Facebook 的又一研究</span></section></section></section><p style="margin-right: 8px; margin-left: 8px; max-width: 100%; min-height: 1em; letter-spacing: 0.544px; box-sizing: border-box !important; overflow-wrap: break-word !important;"><br></p><p style="text-align: center;"></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">github链接:https://github.com/facebookresearch/fastText</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">这是来自Facebook的另一个研究,fastText库专为文本表述和分类而设计。它配备了预先训练的150多种语言的词向量模型。这些词向量可用于多种任务,包括文本分类、摘要和翻译。</span></p><p><br></p><section class="" style="max-width: 100%; box-sizing: border-box; border-width: 0px; border-style: initial; border-color: initial; clear: both; overflow-wrap: break-word !important;"><section class="" style="padding: 8px;max-width: 100%;box-sizing: border-box;border-left: 6px solid rgb(255, 202, 0);border-top-color: rgb(255, 202, 0);border-right-color: rgb(255, 202, 0);border-bottom-color: rgb(255, 202, 0);font-size: 18px;line-height: 1.4;font-family: inherit;font-weight: bold;text-decoration: inherit;color: rgb(10, 10, 10);overflow-wrap: break-word !important;"><section style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="letter-spacing: 0.544px;">Auto-Keras——</span><span style="letter-spacing: 1px;">降低机器学习门槛</span></section></section></section><p><br></p><p style="margin-right: 8px; margin-left: 8px; text-align: center;"></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">GitHub链接:</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">https://github.com/jhfjhfj1/autokeras?utm_source=mybridge&amp;utm_medium=blog&amp;utm_campaign=read_more</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">Auto-Keras是一个用于自动机器学习(AutoML)的开源软件库。它由德克萨斯州农工(Texas A&M)大学的DATA实验室和社区贡献者开发。</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">AutoML的最终目标是为只有有限数据科学或机器学习背景的领域专家提供易于访问的深度学习工具。Auto-Keras提供自动搜索深度学习模型的最佳架构和超参数的功能。</span></p><p><br></p><section class="" style="max-width: 100%; box-sizing: border-box; letter-spacing: 0.544px; border-width: 0px; border-style: initial; border-color: initial; clear: both; overflow-wrap: break-word !important;"><section class="" style="padding: 8px;max-width: 100%;box-sizing: border-box;border-left: 6px solid rgb(255, 202, 0);border-top-color: rgb(255, 202, 0);border-right-color: rgb(255, 202, 0);border-bottom-color: rgb(255, 202, 0);font-size: 18px;line-height: 1.4;font-family: inherit;font-weight: bold;text-decoration: inherit;color: rgb(10, 10, 10);overflow-wrap: break-word !important;"><section style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;letter-spacing: 1px;box-sizing: border-box !important;overflow-wrap: break-word !important;">Dopamine——灵活易使用</span></section></section></section><p style="margin-right: 8px; margin-left: 8px; max-width: 100%; min-height: 1em; letter-spacing: 0.544px; box-sizing: border-box !important; overflow-wrap: break-word !important;"><br></p><p style="margin-right: 8px; margin-left: 8px; text-align: center;"></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">GitHub链接:</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">https://github.com/google/dopamine?utm_source=mybridge&amp;utm_medium=blog&amp;utm_campaign=read_more</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">Dopamine由谷歌创建,是快速原型的强化学习算法的研究框架。它旨在灵活且易于使用,实现标准的RL算法、指标和基准。</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">根据Dopamine的文档,他们的设计原则是:</span></p><p><br></p><ul class=" list-paddingleft-2"><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">简单的实验:帮助新用户运行基准实验;</span></p></li><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">灵活的开发:为新用户提供新的创新想法;</span></p></li><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">紧凑且可靠:为一些较旧且更流行的算法提供实现的可能性;</span></p></li><li><p style="margin-right: 8px;margin-left: 8px;line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">可重复性:确保结果的可重复性。</span></p></li></ul><p><br></p><section class="" style="max-width: 100%; box-sizing: border-box; letter-spacing: 0.544px; border-width: 0px; border-style: initial; border-color: initial; clear: both; overflow-wrap: break-word !important;"><section class="" style="padding: 8px;max-width: 100%;box-sizing: border-box;border-left: 6px solid rgb(255, 202, 0);border-top-color: rgb(255, 202, 0);border-right-color: rgb(255, 202, 0);border-bottom-color: rgb(255, 202, 0);font-size: 18px;line-height: 1.4;font-family: inherit;font-weight: bold;text-decoration: inherit;color: rgb(10, 10, 10);overflow-wrap: break-word !important;"><section style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;letter-spacing: 1px;box-sizing: border-box !important;overflow-wrap: break-word !important;">vid2vid——英伟达的“黑科技”</span></section></section></section><p><br></p><p style="margin-right: 8px; margin-left: 8px; text-align: center;"></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">GitHub链接:</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;color: rgb(136, 136, 136);">https://github.com/NVIDIA/vid2vid?utm_source=mybridge&amp;utm_medium=blog&amp;utm_campaign=read_more</span></p><p><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">vid2vid项目是英伟达(Nvidia)最先进的视频到视频的合成算法。Pytorch实现了高分辨率(例如2048x1024)逼真的视频到视频转换方法。这一项目的目标是学习从输入源视频到精确描绘源视频内容的输出拟真视频的变换功能。</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; text-align: center;"></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="font-size: 15px;letter-spacing: 0.5px;">这个库的好处在于其选择多样性:它提供了几种不同的vid2vid应用程序,包括自动驾驶/城市场景,人脸和人体姿势。它还附带了丰富的指令和功能,包括数据集加载、任务评估、训练功能和多块GPU。</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><br></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="letter-spacing: 0.5px; color: rgb(136, 136, 136);">参考链接:</span></p><p style="margin-right: 8px; margin-left: 8px; line-height: 1.75em;"><span style="letter-spacing: 0.5px; color: rgb(136, 136, 136);">https://towardsdatascience.com/the-10-most-useful-machine-learning-projects-of-the-past-year-2018-5378bbd4919f</span></p><hr style="max-width: 100%; border-style: solid; border-right-width: 0px; border-bottom-width: 0px; border-left-width: 0px; border-color: rgba(0, 0, 0, 0.1); transform-origin: 0px 0px 0px; transform: scale(1, 0.5); box-sizing: border-box !important; overflow-wrap: break-word !important;"><p><br></p><p></p>
页: [1]
查看完整版本: 2018年最实用机器学习项目Top 6(附开源链接)