Text Simplification On Turkcorpus

評価指標

BLEU

FKGL

SARI (EASSEu003e=0.2.1)

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名	BLEU	FKGL	SARI (EASSEu003e=0.2.1)	Paper Title	Repository
GPT-175B (6 SARI-selected examples, high/low)	79.83	9.33	43.46	Metric-Based In-context Learning: A Case Study in Text Simplification
SEMoses	74.49	-	36.70	Simple and Effective Text Simplification Using Semantic and Neural Methods	-
Pointer + Multi-task Entailment and Paraphrase Generation	81.49	-	37.45	Dynamic Multi-Level Multi-Task Learning for Sentence Simplification	-
UNMT (Unsupervised)	74.02	-	37.20	Unsupervised Neural Text Simplification
DMASS-DCSS	-	-	40.45	Integrating Transformer and Paraphrase Rules for Sentence Simplification
NTS-SARI	80.69	-	37.25	Exploring Neural Text Simplification Models	-
EditNTS	86.69	-	38.22	EditNTS: An Neural Programmer-Interpreter Model for Sentence Simplification through Explicit Editing
BART	-	-	-	The GEM Benchmark: Natural Language Generation, its Evaluation and Metrics	-
Dress	77.18	-	37.08	Sentence Simplification with Deep Reinforcement Learning
UNTS-10k (Weakly supervised)	-	-	37.15	Unsupervised Neural Text Simplification
UNTS (Unsupervised)	-	-	36.29	Unsupervised Neural Text Simplification
PBMT-R	-	-	38.04	-	-
MUSS (BART+ACCESS Unsupervised)	-	8.79	40.85	MUSS: Multilingual Unsupervised Sentence Simplification by Mining Paraphrases
NSELSTM-B	92.02	-	33.43	Sentence Simplification with Memory-Augmented Neural Networks	-
SeqLabel	-	-	37.08*	Learning How to Simplify From Explicit Labeling of Complex-Simplified Text Pairs
NSELSTM-S	80.43	-	36.88	Sentence Simplification with Memory-Augmented Neural Networks	-
ACCESS	72.53	-	41.38	Controllable Sentence Simplification
Hybrid	48.97*	-	31.40*	-	-
SBMT-SARI	73.08*	-	39.56	Optimizing Statistical Machine Translation for Text Simplification
Control Prefixes (BART)	-	7.74	42.32	Control Prefixes for Parameter-Efficient Text Generation

0 of 25 row(s) selected.