Finding Replicable Human Evaluations via Stable Ranking Probability

• URL: http://arxiv.org/abs/2404.01474v1
• Date: Mon, 1 Apr 2024 20:50:13 GMT
• Title: Finding Replicable Human Evaluations via Stable Ranking Probability
• Authors: Parker Riley, Daniel Deutsch, George Foster, Viresh Ratnakar, Ali Dabirmoghaddam, Markus Freitag,
• Abstract summary: We use machine translation and its state-of-the-art human evaluation framework, MQM, as a case study to understand how to set up reliable human evaluations. Our study on two language pairs provides concrete recommendations for designing replicable human evaluation studies.
• Score: 28.87806354986128
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Reliable human evaluation is critical to the development of successful natural language generation models, but achieving it is notoriously difficult. Stability is a crucial requirement when ranking systems by quality: consistent ranking of systems across repeated evaluations is not just desirable, but essential. Without it, there is no reliable foundation for hill-climbing or product launch decisions. In this paper, we use machine translation and its state-of-the-art human evaluation framework, MQM, as a case study to understand how to set up reliable human evaluations that yield stable conclusions. We investigate the optimal configurations for item allocation to raters, number of ratings per item, and score normalization. Our study on two language pairs provides concrete recommendations for designing replicable human evaluation studies. We also collect and release the largest publicly available dataset of multi-segment translations rated by multiple professional translators, consisting of nearly 140,000 segment annotations across two language pairs.

Related papers

• FLASK: Fine-grained Language Model Evaluation based on Alignment Skill Sets [69.91340332545094]
  We introduce FLASK, a fine-grained evaluation protocol for both human-based and model-based evaluation. We experimentally observe that the fine-graininess of evaluation is crucial for attaining a holistic view of model performance.
  arXiv  Detail & Related papers  (2023-07-20T14:56:35Z)
• Learning and Evaluating Human Preferences for Conversational Head Generation [101.89332968344102]
  We propose a novel learning-based evaluation metric named Preference Score (PS) for fitting human preference according to the quantitative evaluations across different dimensions. PS can serve as a quantitative evaluation without the need for human annotation.
  arXiv  Detail & Related papers  (2023-07-20T07:04:16Z)
• Disco-Bench: A Discourse-Aware Evaluation Benchmark for Language Modelling [70.23876429382969]
  We propose a benchmark that can evaluate intra-sentence discourse properties across a diverse set of NLP tasks. Disco-Bench consists of 9 document-level testsets in the literature domain, which contain rich discourse phenomena. For linguistic analysis, we also design a diagnostic test suite that can examine whether the target models learn discourse knowledge.
  arXiv  Detail & Related papers  (2023-07-16T15:18:25Z)
• SEAHORSE: A Multilingual, Multifaceted Dataset for Summarization Evaluation [52.186343500576214]
  We introduce SEAHORSE, a dataset for multilingual, multifaceted summarization evaluation. SEAHORSE consists of 96K summaries with human ratings along 6 dimensions of text quality. We show that metrics trained with SEAHORSE achieve strong performance on the out-of-domain meta-evaluation benchmarks TRUE and mFACE.
  arXiv  Detail & Related papers  (2023-05-22T16:25:07Z)
• Large Language Models are Diverse Role-Players for Summarization Evaluation [82.31575622685902]
  A document summary's quality can be assessed by human annotators on various criteria, both objective ones like grammar and correctness, and subjective ones like informativeness, succinctness, and appeal. Most of the automatic evaluation methods like BLUE/ROUGE may be not able to adequately capture the above dimensions. We propose a new evaluation framework based on LLMs, which provides a comprehensive evaluation framework by comparing generated text and reference text from both objective and subjective aspects.
  arXiv  Detail & Related papers  (2023-03-27T10:40:59Z)
• Revisiting the Gold Standard: Grounding Summarization Evaluation with Robust Human Evaluation [136.16507050034755]
  Existing human evaluation studies for summarization either exhibit a low inter-annotator agreement or have insufficient scale. We propose a modified summarization salience protocol, Atomic Content Units (ACUs), which is based on fine-grained semantic units. We curate the Robust Summarization Evaluation (RoSE) benchmark, a large human evaluation dataset consisting of 22,000 summary-level annotations over 28 top-performing systems.
  arXiv  Detail & Related papers  (2022-12-15T17:26:05Z)
• Consistent Human Evaluation of Machine Translation across Language Pairs [21.81895199744468]
  We propose a new metric called XSTS that is more focused on semantic equivalence and a cross-lingual calibration method. We demonstrate the effectiveness of these novel contributions in large scale evaluation studies across up to 14 language pairs.
  arXiv  Detail & Related papers  (2022-05-17T17:57:06Z)
• Does Summary Evaluation Survive Translation to Other Languages? [0.0]
  We translate an existing English summarization dataset, SummEval dataset, to four different languages. We analyze the scores from the automatic evaluation metrics in translated languages, as well as their correlation with human annotations in the source language.
  arXiv  Detail & Related papers  (2021-09-16T17:35:01Z)
• Experts, Errors, and Context: A Large-Scale Study of Human Evaluation for Machine Translation [19.116396693370422]
  We propose an evaluation methodology grounded in explicit error analysis, based on the Multidimensional Quality Metrics framework. We carry out the largest MQM research study to date, scoring the outputs of top systems from the WMT 2020 shared task in two language pairs. We analyze the resulting data extensively, finding among other results a substantially different ranking of evaluated systems from the one established by the WMT crowd workers.
  arXiv  Detail & Related papers  (2021-04-29T16:42:09Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.