From d46df9c1a2df73fc56bcfcc942ab0f84e7871d14 Mon Sep 17 00:00:00 2001 From: yuweibao Date: Fri, 7 Jul 2023 14:47:36 -0400 Subject: [PATCH] adding readme --- README.md | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/README.md b/README.md index e38f04e..ec9bb5b 100644 --- a/README.md +++ b/README.md @@ -49,8 +49,7 @@ Comparative Learning is the process of finding the similarities and differences Highlights: - **Acquisition Process:** We define the word acquisition as two parts of learning: **Information Filteration** and **Representation-Word Mapping**. It is to learn a computation as well as a representation. All learned feature-word mapping will be stored in memory. -- **Continual Learning:** In this work, we compute the centroid of a SIM batch to extract their shared feature, and refine the scope of this feature with the DIF batch. With the help of memory storage, 1) New words can be continually added to the memory; 2) the already learned word-feature can be pulled out of the memory, updated and refined when more examples are availble. +- **Continual Learning:** In this work, we compute the centroid of a SIM batch to extract their shared feature, and refine the scope of this feature with the DIF batch. With the help of memory storage, 1) New words can be continually added to the memory; 2) the existing word-feature can be pulled out of the memory, updated and refined when more examples are availble. - ![alt text](https://github.com/sled-group/Comparative-Learning/blob/main/assets/pipeline.png)