davanstrien HF staff commited on
Commit
b79c084
1 Parent(s): 9de292c
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -103,7 +103,7 @@ To illustrate we can look at some example Iconclass classifications.
103
 
104
  [source](https://iconclass.org/41A12)
105
 
106
- The construction of Iconclass of parts makes it particulalry interesting (and challenging) to tackle via Machine Learning. Whilst one could tackle this dataset as a (multi) label image classification problem, this is only one way of tackling it. For example in the above label `castle` giving the model the 'freedom' to predict only a partial label could result in the prediction `41A` i.e. housing. Whilst a very particular form of housing this prediction is not 'wrong' so much as it is not as precise as a human cataloguer may provide.
107
 
108
  ### Supported Tasks and Leaderboards
109
 
@@ -193,7 +193,7 @@ The annotations are derived from the source dataset see above. It is likely that
193
 
194
  ### Discussion of Biases
195
 
196
- Iconclass as a metadata standard absorbs biases from the time and place of it's creation (1940's Netherlands). In particular, '32B human races, peoples; nationalities' has been subject to criticism. '32B36 'primitive', 'pre-modern' peoples' is one example of a category which we may not wish to adopt. In general there are components of the subdivsions of `32B` which reflect a belief that race is a scientific category rather than socialy constructed.
197
 
198
  These limitations are actively being explored by the Iconclass community, for example, see [Revising Iconclass section 32B human races, peoples; nationalities](https://web.archive.org/web/20210425131753/https://iconclass.org/Updating32B.pdf).
199
 
 
103
 
104
  [source](https://iconclass.org/41A12)
105
 
106
+ The construction of Iconclass of parts makes it particularly interesting (and challenging) to tackle via Machine Learning. Whilst one could tackle this dataset as a (multi) label image classification problem, this is only one way of tackling it. For example in the above label `castle` giving the model the 'freedom' to predict only a partial label could result in the prediction `41A` i.e. housing. Whilst a very particular form of housing this prediction for 'castle' is not 'wrong' so much as it is not as precise as a human cataloguer may provide.
107
 
108
  ### Supported Tasks and Leaderboards
109
 
 
193
 
194
  ### Discussion of Biases
195
 
196
+ Iconclass as a metadata standard absorbs biases from the time and place of it's creation (1940's Netherlands). In particular, '32B human races, peoples; nationalities' has been subject to criticism. '32B36 'primitive', 'pre-modern' peoples' is one example of a category which we may not wish to adopt. In general there are components of the subdivsions of `32B` which reflect a belief that race is a scientific category rather than socially constructed.
197
 
198
  These limitations are actively being explored by the Iconclass community, for example, see [Revising Iconclass section 32B human races, peoples; nationalities](https://web.archive.org/web/20210425131753/https://iconclass.org/Updating32B.pdf).
199