人工智能培训

搜索

论文代码开源:噪声标签存在下的高光谱图像分类(Hyperspectral Image Classification in the Presence of Noi

[复制链接]
admin 发表于 2018-9-15 09:48:52 | 显示全部楼层 |阅读模式
admin 2018-9-15 09:48:52 1612 0 显示全部楼层
人工智能论文代码开源:噪声标签存在下的高光谱图像分类(Hyperspectral Image Classification in the Presence of Noisy Labels)请注意该人工智能论文代码开源在github,大部分是python写的,框架可能是tensorflow或者pytorch。医疗应用通过要求高精度和易于解释来挑战当今的文本分类技术。尽管深度学习在准确性方面提供了一个飞跃,但这种飞跃是在牺牲可解释性的基础上实现的。为了解决这种准确性 - 可解释性挑战,我们首次引入了一种文本分类方法,该方法利用了最近引入的Tsetlin机器。简而言之,我们将文本的术语表示为命题变量。从这些,我们使用简单的命题公式捕获类别,例如:如果“rash”和“反应”和“青霉素”然后过敏。 Tsetlin机器从标签文本中学习这些公式,利用连接子句来表示每个类别的特定方面。实际上,即使没有术语(否定的特征)也可用于分类目的。我们的实证结果非常有说服力.Tsetlin机器在20个新闻组和IMDb数据集以及非公共临床数据集上的表现均优于或优于所有评估方法。平均而言,Tsetlin机器可在整个数据集中提供最佳的回忆和精度分数。 Tsetlin Machine的GPU实现速度比神秘网络的GPU实现速度快8倍。因此,我们相信我们的新方法可以对广泛的文本分析应用产生重大影响,形成一个有希望的开始点,使用TsetlinMachine更深入地理解自然语言。
Medical applications challenge today's text categorization techniques bydemanding both high accuracy and ease-of-interpretation.Although deep learninghas provided a leap ahead in accuracy, this leap comes at the sacrifice ofinterpretability.To address this accuracy-interpretability challenge, we hereintroduce, for the first time, a text categorization approach that leveragesthe recently introduced Tsetlin Machine.In all brevity, we represent the termsof a text as propositional variables.From these, we capture categories usingsimple propositional formulae, such as: if "rash" and "reaction" and"penicillin" then Allergy.The Tsetlin Machine learns these formulae from alabelled text, utilizing conjunctive clauses to represent the particular facetsof each category.Indeed, even the absence of terms (negated features) can beused for categorization purposes.Our empirical results are quite conclusive.The Tsetlin Machine either performs on par with or outperforms all of theevaluated methods on both the 20 Newsgroups and IMDb datasets, as well as on anon-public clinical dataset.On average, the Tsetlin Machine delivers the bestrecall and precision scores across the datasets.The GPU implementation of theTsetlin Machine is further 8 times faster than the GPU implementation of theneural network.We thus believe that our novel approach can have a significantimpact on a wide range of text analysis applications, forming a promisingstarting point for deeper natural language understanding with the TsetlinMachine.论文代码开源:噪声标签存在下的高光谱图像分类(Hyperspectral Image Classification in the Presence of Noisy Labels) e0Oq1UH01Wz1zQee.jpg
URL地址:https://arxiv.org/abs/1809.04547v1     ----pdf下载地址:https://arxiv.org/pdf/1809.04547v1    ----         ----github下载地址:https://github.com/cair/TextUnderstandingTsetlinMachine    ----    论文代码开源:噪声标签存在下的高光谱图像分类(Hyperspectral Image Classification in the Presence of Noisy Labels)请注意该人工智能论文代码开源在github,大部分是python写的,框架可能是tensorflow或者pytorch,keras,至于具体是哪一个没有完全测试。
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

admin当前离线
管理员

查看:1612 | 回复:0

快速回复 返回顶部 返回列表