Neural-Symbolic VQA: Disentangling Reasoning from Vision and Language Understanding
Open paper
Benchmark
Model
Rank
Results
visual-question-answering-on-clevr
NS-VQA (1K programs)
#1
Accuracy: 99.8
visual-question-answering-on-clevr-humans
NS-VQA (1K programs)
#4
Accuracy: 67.8
Rank counts only results with a code link.