Probing Neural Network Comprehension of Natural Language Arguments

"We are surprised to find that BERT's peak performance of 77% on the Argument Reasoning Comprehension Task reaches just three points below the average untrained human baseline. However, we show that this result is entirely accounted for by exploitation of spurious statistical cues in the dataset. We analyze the nature of these cues and demonstrate that a range of models all exploit them."

Timothy Niven and Hung-Yu Kao: https://arxiv.org/abs/1907.07355

#naturallanguage #neuralnetwork #reasoning #unsupervisedlearning

Mathematical Reasoning in Latent Space
Lee et al.: https://arxiv.org/pdf/1909.11851v1.pdf
#Mathematics #Reasoning #LatentSpace

SocialIQA: Commonsense Reasoning about Social Interactions
Sap et al.: https://arxiv.org/abs/1904.09728
#Commonsense #MachineLearning #Reasoning

Reasoning about Time and Knowledge in Neural Symbolic Learning Systems
Artur Garcez and Luis C. Lamb : https://papers.nips.cc/paper/2490-reasoning-about-time-and-knowledge-in-neural-symbolic-learning-systems
#ArtificialIntelligence #Reasoning #SymbolicAI

Differentiable Reasoning on Large Knowledge Bases and Natural Language
Minervini et al.: https://arxiv.org/abs/1912.10824
Open-source neuro-symbolic reasoning framework, in TensorFlow: https://github.com/uclnlp/gntp
#AIDebate #NeuroSymbolic #Reasoning

Neural Module Networks for Reasoning over Text
Gupta et al.: https://arxiv.org/abs/1912.04971
Code: https://nitishgupta.github.io/nmn-drop
#NeuralNetworks #Reasoning #SymbolicAI