Natural Language Understanding On Pdp60

Metrics

Accuracy

Results

Performance results of various models on this benchmark

		Paper Title
HNN	90	A Hybrid Neural Network Model for Commonsense Reasoning
UDSSM-II (ensemble)	78.3	Unsupervised Deep Structured Semantic Models for Commonsense Reasoning
BERT-large 340M	78.3	BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
UDSSM-I (ensemble)	76.7	Unsupervised Deep Structured Semantic Models for Commonsense Reasoning
UDSSM-II	75	Unsupervised Deep Structured Semantic Models for Commonsense Reasoning
DSSM	75.0	Unsupervised Deep Structured Semantic Models for Commonsense Reasoning
BERT-base 110M + MAS	68.3	Attention Is (not) All You Need for Commonsense Reasoning
USSM + Supervised Deepnet + 3 Knowledge Bases	66.7	Attention Is (not) All You Need for Commonsense Reasoning
Word-level CNN+LSTM (full scoring)	60.0	A Simple Method for Commonsense Reasoning
Subword-level Transformer LM	58.3	Attention Is All You Need
USSM + Cause-Effect Knowledge Base	55.0	Probabilistic Reasoning via Deep Learning: Neural Association Models
Word-level CNN+LSTM (partial scoring)	53.3	A Simple Method for Commonsense Reasoning
USSM + Supervised Deepnet	53.3	Attention Is (not) All You Need for Commonsense Reasoning

0 of 13 row(s) selected.

Natural Language Understanding On Pdp60 | SOTA | HyperAI