Challenge: Existing learned metrics perform unsatisfactory across text generation tasks or require human annotations for training on specific tasks.
Approach: They propose a self-supervised approach to train a model-based metric for text generation evaluation using sentences retrieved from a corpus.
Outcome: The proposed model outperforms all prior unsupervised metrics on four text generation evaluation benchmarks, with an average Kendall improvement of 0.158.

Similar Papers

Not All Errors are Equal: Learning Text Generation Metrics using Stratified Error Synthesis (2022.findings-emnlp)

Copied to clipboard

Challenge: Existing learning metrics are limited to tasks where large human ratings are available.
Approach: They propose a model-based natural language generation (NLG) evaluation metric that is highly correlated with human judgements without requiring human annotation.
Outcome: The proposed metric outperforms all prior unsupervised metrics on multiple NLG tasks including translation, image captioning, and WebNLG text generation.
On the Blind Spots of Model-Based Evaluation Metrics for Text Generation (2023.acl-long)

Copied to clipboard

Challenge: Existing methods for text generation evaluation metrics are lacking in robustness analysis.
Approach: They propose to use stress tests to test for errors in text generation evaluation metrics . they find that BERTScore is confused by truncation errors in summarization .
Outcome: The proposed stress tests show that they are insensitive to errors in open-ended generation, translation, and summarization.
Toward Human-Like Evaluation for Natural Language Generation with Error Analysis (2023.acl-long)

Copied to clipboard

Challenge: Pre-trained language models (PLMs) have been used to evaluate language generation tasks . pretrained error analysis can be used to refine the generated sentence toward higher confidence .
Approach: They propose to combine pretrained language model based metrics with human-like error analysis to improve sentence confidence.
Outcome: The proposed method outperforms top-scoring metrics in 19/25 settings.
INSTRUCTSCORE: Towards Explainable Text Generation Evaluation with Automatic Feedback (2023.emnlp-main)

Copied to clipboard

Challenge: Existing methods to evaluate the quality of language generation do not provide explicit explanation of their verdicts.
Approach: They propose a fine-grained explainable evaluation metric for text generation that harnesses human instruction and implicit knowledge of GPT-4 to fine-tune it.
Outcome: The proposed model outperforms all other unsupervised metrics on translation, captioning, data-to-text, and commonsense generation tasks.
DecompEval: Evaluating Generated Texts as Unsupervised Decomposed Question Answering (2023.acl-long)

Copied to clipboard

Challenge: Existing evaluation metrics for natural language generation (NLG) tasks face the challenges on generalization ability and interpretability.
Approach: They propose a metric that evaluates natural language generation tasks as an instruction-style question answering task and utilizes instruction-tuned pre-trained language models without training on evaluation datasets.
Outcome: The proposed metric achieves state-of-the-art performance in untrained metrics for evaluating text summarization and dialogue generation, which exhibits strong dimension-level / task-level generalization ability and interpretability.
Curious Case of Language Generation Evaluation Metrics: A Cautionary Tale (2020.coling-main)

Copied to clipboard

Challenge: a few popular metrics are still used to evaluate language generation systems despite their known limitations.
Approach: They propose to use automatic metrics to evaluate language generation systems . they show that they prefer system outputs to human-authored texts .
Outcome: The proposed metrics are insensitive to correct translations of rare words and can yield high scores when given a single sentence as system output for the entire test set.
Correction of Errors in Preference Ratings from Automated Metrics for Text Generation (2023.findings-acl)

Copied to clipboard

Challenge: Existing evaluation methods are over-confident in assigning significant differences between systems . Currently, the most reliable evaluation methods for text generation are human-based evaluations.
Approach: They propose to combine human ratings with automated ratings to reduce the amount of human ratings needed to arrive at robust results.
Outcome: The proposed evaluation protocol reduces the amount of human ratings by 50% while yielding the same evaluation outcome as the pure human evaluation in 95% of cases.
BLEURT: Learning Robust Metrics for Text Generation (2020.acl-main)

Copied to clipboard

Challenge: Text generation has made significant advances, but evaluation metrics have lagged behind.
Approach: They propose a learning evaluation metric for English based on BERT . BLEURT can model human judgment with a few thousand possibly biased training examples .
Outcome: The proposed model can model human judgment with a few thousand potentially biased training examples.
CTRLEval: An Unsupervised Reference-Free Metric for Evaluating Controlled Text Generation (2022.acl-long)

Copied to clipboard

Challenge: Existing reference-free metrics have obvious limitations for evaluating controlled text generation models.
Approach: They propose an unsupervised reference-free metric which evaluates controlled text generation from different aspects by formulating each aspect into multiple text infilling tasks.
Outcome: The proposed metric has higher correlations with human judgments while obtaining better generalization of evaluating generated texts from different models and with different qualities.
How Do Seq2Seq Models Perform on End-to-End Data-to-Text Generation? (2022.acl-long)

Copied to clipboard

Challenge: Existing models for data-to-text generation are based on pipelines and end-to end architectures.
Approach: They use multidimensional quality metrics to evaluate models on end-to-end data-totext generation and compare their performance against pipeline models.
Outcome: The proposed model improves in Omission and Inaccuracy Extrinsic errors but increases errors such as Addition.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations