Is the Best Better? Bayesian Statistical Model Comparison for Natural Language Processing (2020.emnlp-main)
Copied to clipboard
| Challenge: | a recent study raises concerns about the use of standard splits to compare models . we compare the performance of six English part-of-speech taggers to those of other models based on standard split analysis . |
| Approach: | They propose a Bayesian statistical model comparison technique using k-fold cross-validation . they rank six English part-of-speech taggers across two data sets and three evaluation metrics . |
| Outcome: | The proposed method ranks English part-of-speech taggers on two data sets and three evaluation metrics. |
Similar Papers
We Need to Talk about Standard Splits (P19-1)
Copied to clipboard
| Challenge: | Existing methods to evaluate systems with a held-out test set are insufficient for system comparison. |
| Approach: | They propose to use multiple random splits to compare performance of systems . they replicate results on standard split but fail to reproduce some rankings . |
| Outcome: | The proposed method is based on multiple random splits to replicate results with a set of part-of-speech taggers. |
Not All Claims are Created Equal: Choosing the Right Statistical Approach to Assess Hypotheses (2020.acl-main)
Copied to clipboard
| Challenge: | Empirical research in natural language processing has adopted a narrow set of principles for assessing hypotheses . alternative approaches to assess hypothese rely on p-value computation, which suffers from several known issues. |
| Approach: | They propose to compare different methods for assessing hypotheses . they argue that practitioners should first decide their target hypothesis before choosing a method . |
| Outcome: | The proposed method differs from other methods, but is not widely used in NLP . the proposed method is based on a p-value computation, but has a small gap in accuracy . |
Bayes Test of Precision, Recall, and F1 Measure for Comparison of Two Natural Language Processing Models (P19-1)
Copied to clipboard
| Challenge: | Existing t-tests for cross-validation (CV) are inappropriate for model comparison . existing t tests for cross validation (CV), such as 52 CV t test and F ttest, are inadequate . |
| Approach: | They propose to use a block-regularized 32 CV to compare two NLP models . they calibrate the posterior distributions of P, R, and F1 and derive an accurate interval estimation of P and R . |
| Outcome: | The proposed model could regularize the difference in certain frequency distributions over linguistic units and yield stable estimators of P, R, and F1. |
Understanding Cross-Lingual Alignment—A Survey (2024.findings-acl)
Copied to clipboard
| Challenge: | Cross-lingual alignment is the meaningful similarity of representations across languages in multilingual language models. |
| Approach: | They propose a taxonomy of methods to improve cross-lingual alignment . they argue that an effective trade-off between language-neutral and language-specific information is key . |
| Outcome: | The proposed methods can be applied to encoder models and encoder-decoder-only models . they show that language-neutral and language-specific information is key . |
Butterfly Effects in Frame Semantic Parsing: impact of data processing on model ranking (C18-1)
Copied to clipboard
| Challenge: | a common contribution to computational linguistics research is a new model for a specific task. |
| Approach: | They propose an open-source standardized processing pipeline for frame semantic parsing . they propose a standard evaluation resource that can be shared and reused for robust comparison . |
| Outcome: | The proposed model can be shared and reused for robust model comparison. |
Please, Don’t Forget the Difference and the Confidence Interval when Seeking for the State-of-the-Art Status (2022.lrec-1)
Copied to clipboard
| Challenge: | comparing NLP systems by performance has become an essential question . comparing systems by performing performance criterion is criticized for allowing chance to determine superiority . |
| Approach: | They propose to use bootstrap confidence intervals instead of state-of-the-art status and statistical significance testing to compare NLP system performance. |
| Outcome: | The bootstrap confidence intervals are used to compare NLP system performance . the bootstrap test is more accurate than state-of-the-art status and statistical significance testing . |
Are All Languages Equally Hard to Language-Model? (N18-2)
Copied to clipboard
| Challenge: | a fair comparison of language models is tricky because of the size of the corpora and the variability of orthographic systems. |
| Approach: | They propose a framework for fair cross-linguistic comparison of language models . they show that in some languages, textual expression is harder to predict with n-gram models compared to LSTM models based on translated text . |
| Outcome: | The proposed framework is based on translated text and language models on 21 languages. |
Benchmarking Long-tail Generalization with Likelihood Splits (2023.findings-eacl)
Copied to clipboard
| Challenge: | Existing methods to test out of distribution generalization have their own strengths and weaknesses. |
| Approach: | They propose a method to create challenging benchmarks that require generalizing to the tail of the distribution by re-splitting existing datasets. |
| Outcome: | The proposed approach can be customized to construct meaningful splits for a wide range of tasks. |
Rethinking the Evaluation of Alignment Methods: Insights into Diversity, Generalisation, and Safety (2026.eacl-srw)
Copied to clipboard
Denis Janiak, Julia Moska, Dawid Motyka, Karolina Seweryn, Paweł Walkowiak, Bartosz Żuk, Arkadiusz Janz
| Challenge: | Existing studies focus on individual techniques or specific dimensions, lacking a holistic assessment of the inherent trade-offs. |
| Approach: | They propose a framework that compares LLM alignment methods across five axes . they use a validated LLM-as-judge prompt to compare the results . |
| Outcome: | The proposed framework compares LLM alignment methods across factuality, safety, conciseness, proactivity, diversity and safety axes . it provides insights into trade-offs of common alignment methods, guiding the development of more balanced and reliable LLMs. |
Sensitivity to Input Order: Evaluation of an Incremental and Memory-Limited Bayesian Cross-Situational Word Learning Model (C18-1)
Copied to clipboard
| Challenge: | Existing word learning models are insensitive to input order effects, but they are noisy and only provide ambiguous information. |
| Approach: | They propose a Bayesian cross-situational word learning model with an incremental memory-limited algorithm for predicting input order effects. |
| Outcome: | The proposed model performs well on corpus data while being insensitive to input order effects. |