DSSM #141
Merged
DSSM #141
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
lcy-seso
requested changes
Jul 5, 2017
dssm/README.md
Outdated
| @@ -0,0 +1,397 @@ | |||
| # Deep Structured Semantic Models (DSSM) | |||
| 两个单位间语义距离的权衡是一种非常常规的需求，本文将会演示如何使用 DSSM 模型建模两个字符串间的关系， | |||
There was a problem hiding this comment.
- “两个单位间语义距离的权衡是一种非常常规的需求” 这句话的含义不是特别精确，我没有从这个描述中精确地理解到这篇例子的模型要解决什么问题。
- “使用 DSSM 模型建模两个字符串间的关系” --> 什么关系？语义关系？
- “单位”这里具体都可以指什么呢？有些意义不明确。
- ”本文将会演示“ --> "本例演示"，和其它篇章保持一致。
dssm/README.md
Outdated
| 具体模型实现支持通用的数据格式，用户只需要替换数据就可以在自己的任务上训练和预测。 | ||
|
|
||
| ## 背景介绍 | ||
| DSSM \[[1](##参考文档)\]是微软研究院13年提出来的比较经典的语义模型，用于学习两个单位之间的语义距离，适用于如下场景： |
There was a problem hiding this comment.
- 比较经典 --> 这样口语化的描述尽量避免。技术文章里，经典就是经典，不经典就是不经打，不存在“比较经典”。
- “单位” 有没有其他说法？或者其它词汇，表意不具体，不明确。
dssm/README.md
Outdated
| @@ -0,0 +1,397 @@ | |||
| # Deep Structured Semantic Models (DSSM) | |||
There was a problem hiding this comment.
- 深度结构化语义模型 （Deep Structured Semantic Models, DSSM）
- 在存在中文的情况下，尽量给一个中文的解释
dssm/README.md
Outdated
|
|
||
| 1. CTR预估模型，衡量用户搜索词（Query）与候选网页集合（Documents）之间的相关联程度 | ||
| 2. 文本相关性，衡量两个字符串间的语义相关程度 | ||
| 3. 自动推荐，衡量User与被推荐的Item之间的关联程度 |
dssm/README.md
Outdated
| 2. 文本相关性，衡量两个字符串间的语义相关程度 | ||
| 3. 自动推荐，衡量User与被推荐的Item之间的关联程度 | ||
|
|
||
| DSSM 如今已经发展成了一个框架，可以很自然地建模两个单位之间的距离关系，比如对于相关性，可以用余弦相似度（COS)来刻画，模型具体训练时，可以用分类或者pairwise rank的方式训练。 |
There was a problem hiding this comment.
如今，DSSM 已经发展成一个框架，用于很建模两个单位之间的距离关系。例如：对于相关性问题，可以用余弦相似度（cosine）来刻画，[还有没有其它例子，一个例子用来说明 DSSM 发展为一个框架略显单薄行]。模型的训练可以使用分类或者pairwise rank的方式。
dssm/data/rank/test.txt
Outdated
| @@ -0,0 +1,20 @@ | |||
| 新手 汽车 驾驶 驾校 培训 苹果 6s 1 | |||
| 苹果 六 袋 苹果 6s 新手 汽车 驾驶 1 | |||
There was a problem hiding this comment.
- 示例数据，也将训练和测试数据放置为不同数据。
- 提供不同的数据，删掉重复。
dssm/data/rank/train.txt
Outdated
| @@ -0,0 +1,20 @@ | |||
| 新手 汽车 驾驶 驾校 培训 苹果 6s 1 | |||
| 苹果 六 袋 苹果 6s 新手 汽车 驾驶 1 | |||
There was a problem hiding this comment.
- 示例数据，也将训练和测试数据放置为不同数据。
- 提供不同的数据，删掉重复。
| logger.info("create fc layer [%s] which dimention is %d" % | ||
| (name, dim)) | ||
| fc = paddle.layer.fc( | ||
| name=name, |
| ModelType.CLASSIFICATION) | ||
|
|
||
| for rcd in dataset.train(): | ||
| print rcd |
dssm/utils.py
Outdated
|
|
||
| UNK = 0 | ||
|
|
||
| logger = logging.getLogger("logger") |
There was a problem hiding this comment.
直接 get “paddle” 这个logger，可以不需要再重新定义一个。
lcy-seso
approved these changes
Jul 13, 2017
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
DSSM 模型实现
支持 CLASSIFICATION/REGRESSION/RANK 3种损失函数 +
FC/RNN/CNN 三种模型结构的组合