bhavitvyamalik commited on
Commit
306db2e
1 Parent(s): 68fa3ca

add bias, new images

Browse files
misc/examples/female_biker_resized.jpg ADDED
misc/examples/female_dev_1.jpg ADDED
misc/examples/female_dev_2.jpg ADDED
misc/examples/female_dev_4.jpg ADDED
misc/examples/female_doctor.jpg ADDED
misc/examples/female_doctor_1.jpg ADDED
misc/examples/women_cricket.jpg ADDED
sections/bias.md ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ ## Bias Analysis
2
+ Due to the gender bias in data, gender identification by an image captioning model suffers. Also, the gender-activity bias, owing to the word-by-word prediction, influences other words in the caption prediction, resulting in the well-known problem of label bias.
3
+
4
+ One of the reasons why we chose Conceptual 12M over COCO captioning dataset for training our Multi-lingual Image Captioning model was that in former all named entities of type Person were substituted by a special token <PERSON>. Because of this, the gendered terms in our captions became quite infrequent. We'll present a few captions from our model to analyse how our model performed on different images on which different pre-trained image captioning model usually gives gender prediction biases