Comments (1)
I remember that we also tried the setting in which we do not mask x_i, and it performs poorly on syntactic tasks.
We suspect the reason is that if you do not mask x_i, x_i+1 will only have a very small impact on x_i's representation, and make parsing difficult (since the impact values are quite close to each other).
But yeh, it's also possible only masking once can also do well on sentence segmentation task.
Would be great if you would like to share some of your results, e.g., performance of masking x_i and x_i+1 vs masking x_i+1
from perturbed-masking.
Related Issues (10)
- A Question about Figure 1 in Paper HOT 2
- Why the WSJ data is empty and lack WSJ25 test data in your code HOT 1
- parsing.py: error: argument HOT 3
- Could you share the visualization code to reproduce Figure 1? HOT 1
- Help needed to achieve the distance in the ABSC task HOT 5
- Why does the "CLS" row of the matrix have very high values on PUD? HOT 2
- Could you share the visualization code or package to reproduce Figure 2? HOT 2
- The results about the left/right in ABSC experiments HOT 2
- "Dist" and "Prob" versions HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from perturbed-masking.