Comments (10)
As long as the legal team is alright with it - I think that including MNIST in the title is actually pretty great and not confusing, because even though the dataset is totally different than MNIST, it is related since it's a drop-in replacement for MNIST. It sets out to fix many issues in the original MNIST handwritten digits dataset; and serves the exact same purpose as the original.
from fashion-mnist.
The original NIST digits database is now called "NIST Special Database 19" at the time Yann Lecun created MNIST it was SD 3 and 1. Someone could have made a modified version of the NIST fingerprint or mugshot datasets and named them MNIST prints and MNIST mugs, with just as much accuracy. But nobody did and so the shorthand is globally unique in the world that it is meaningful in (AI research dataset names) But not specific if you want to be pedantic about it you have to tack on the "handwritten digits" part. NIST acknowledges as much by having an official NIST version of MNIST called EMNIST https://www.nist.gov/itl/iad/image-group/emnist-dataset
In the horse and buggy days "dashboard" meant one thing and then something else in the age of the auto and yet something else in the information age so too MNIST in this case is (IMHO) an equally understandable moniker that does not mean what it technically is supposed to mean but captures a gestalt that everyone immediately grasps.
from fashion-mnist.
Good idea, @hanxiao. Otherwise, great job with this benchmark and thank you for sharing it! I'm already running TPOT on your benchmark to see how it works.
@robbiebarrat, technically MNIST was called the "MNIST database of handwritten digits". The handwritten digits database is not called MNIST, though I think that fact was lost in time. A more appropriate shorthand name would be "digits" or something similar.
Some cheeky commenters have suggested "fashioNISTa" as a possible name for this new benchmark.
from fashion-mnist.
@rhiever Oooh i didn't know that the handwritten dataset wasn't actually referred to as MNIST... Oops...
I think fashioNISTa would be a great name, though!
from fashion-mnist.
Some updates in this thread:
- Zalando's legal team said they are fine with the name "Fashion-MNIST".
- "MNIST" is not a registered trademark according to http://tmsearch.uspto.gov/bin/gate.exe?f=searchss&state=4810:neet52.1.1
- We found datasets with "*NIST" in the name e.g. http://www.cs.toronto.edu/~tijmen/affNIST/, http://www.garrickorchard.com/datasets/n-mnist
- Yann LeCun himself posted Fashion-MNIST on his Facebook page, so it seems that he is not angry with the name. https://www.facebook.com/yann.lecun/posts/10154714939492143
I will close this issue soon.
from fashion-mnist.
thanks for reporting. Will consult the company's legal team on Monday and check.
from fashion-mnist.
Even if it's legal, using the name MNIST for something that is not related to the NIST database seems confusing.
from fashion-mnist.
or just call it fashion-digits...
from fashion-mnist.
To be clear: I raised this issue not because of legal concerns, but because (as discussed above) naming it with MNIST in the name doesn't actually make sense. MNIST refers to the institution that originally provided the digits dataset, not the dataset itself.
from fashion-mnist.
OK. I still believe that it's a poor decision to continue misusing the MNIST name in this manner, but ultimately it's not my repo so I'll close the issue.
from fashion-mnist.
Related Issues (20)
- why running the benchmark/runner.py very slow? very slow !!!
- Benchmark - Free Pascal Implementation - 94% with ~174K parameters
- I NEED HELP WITH THE GRAPHICS
- Fashion-MNIST converted to png
- Benchmark : CNN-Fashion-MNIST
- Cant find the csv file of the data set HOT 2
- How to create clustering visulization? label:"help wanted"
- optional native extra labels.
- What's the mean and std of FMNIST? HOT 1
- Benchmark: E2E-3M Accuracy Result of 95.92% for Fashion-MNIST
- Interactive Visualization of the dataset
- Mislabeled Instances Found HOT 5
- Loading the dataset from the local path using tensorflow 2.0 HOT 5
- Benchmark: Ensemble 4 CNNs on censored images
- Link to Pytorch Dataset broken HOT 2
- convolution network mean acc achieve 0.9765
- Incremental learning HOT 1
- Direct download links not working HOT 3
- Invalid Tensor size in TF.js
- Image of full body in trainingset HOT 1
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 fashion-mnist.