baidu / baidu/AnyQ

simnet语义匹配使用MLP与直接用cosine比较有什么不同吗

Open
#130 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
C++
Stars
2.6k
Forks
657
PR merge metrics
No merged PRs in 30d

Description

simnet语义匹配过程中Representation-based Match这部分,将待匹配的两端都转换到等长的语义表示向量里。然后在两端对应的两个语义表示向量基础上,进行匹配度计算,两种计算方法:一种是通过固定的度量函数计算,实际中最常用的就是 cosine 函数,这种方式简单高效,并且得分区间可控意义明确;还有就是将两个向量再过一个多层感知器网络(MLP),通过数据训练拟合出一个匹配度得分,这种方式更加灵活拟合能力更强,但对训练的要求也更高。这里发现采用MLP一般会降低匹配准确度,那么具体为什么要用MLP呢

Contributor guide

No contributing guide indexed for this repository

Research direction

No files, tests, or entry points are named. Start with the issue's Representation-based Match description and compare the cosine and MLP approaches, including why training may reduce accuracy. Done means providing a clear explanation of when MLP is useful and why it may underperform here.

Written by the indexing model from the issue text.

Assessment

Tech stack
machine-learning
Domain
machine-learning
Issue type
Documentation
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.