Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userEvaluating Saliency Methods for Neural Language Models
تقييم أساليب الرافية لنماذج اللغة العصبية
339
0 0
0.0
(
0
)
Created by
Shamra Editor
Ask ChatGPT about the research
تستخدم أساليب الرالف على نطاق واسع لتفسير تنبؤات الشبكة العصبية، ولكن غالبا ما تعارض أساليب أنواع مختلفة من الأساليب المختلفة حتى في تفسيرات نفس التنبؤ الذي أدلى به نفس النموذج. في هذه الحالات، كيف يمكننا تحديد متى تكون هذه التفسيرات جديرة بالثقة بما يكفي لاستخدامها في التحليلات؟ لمعالجة هذا السؤال، نقوم بإجراء تقييم شامل وكمي لأساليب الرالف في فئة أساسية من نماذج NLP: نماذج اللغة العصبية. نقيم جودة تفسيرات التنبؤ من وجهات نظر اثنين يمثل كل منها خاصية مرغوبة لهذه التفسيرات: المعقولية والإخلاص. يتم إجراء تقييمنا على أربع مجموعات بيانات مختلفة تم بناؤها من الشرح البشري الحالي للاتفاقات النحوية واللالسة، على مستوى الحكم على مستوى الحكم والوثائق. من خلال تقييمنا، حددنا طرق مختلفة من الطرق التي يمكن أن تسفر عن تفسيرات ذات جودة منخفضة. نوصي بأن ينشر العمل المستقبلي لنشر هذه الأساليب إلى نماذج اللغة العصبية صحة تفسيراتها بعناية قبل رسم رؤى.
Saliency methods are widely used to interpret neural network predictions, but different variants of saliency methods often disagree even on the interpretations of the same prediction made by the same model. In these cases, how do we identify when are these interpretations trustworthy enough to be used in analyses? To address this question, we conduct a comprehensive and quantitative evaluation of saliency methods on a fundamental category of NLP models: neural language models. We evaluate the quality of prediction interpretations from two perspectives that each represents a desirable property of these interpretations: plausibility and faithfulness. Our evaluation is conducted on four different datasets constructed from the existing human annotation of syntactic and semantic agreements, on both sentence-level and document-level. Through our evaluation, we identified various ways saliency methods could yield interpretations of low quality. We recommend that future work deploying such methods to neural language models should carefully validate their interpretations before drawing insights.
References used
https://aclanthology.org/
rate research
Read More
High-performance neural language models have obtained state-of-the-art results on a wide range of Natural Language Processing (NLP) tasks. However, results for common benchmark datasets often do not reflect model reliability and robustness when appli
ed to noisy, real-world data. In this study, we design and implement various types of character-level and word-level perturbation methods to simulate realistic scenarios in which input texts may be slightly noisy or different from the data distribution on which NLP systems were trained. Conducting comprehensive experiments on different NLP tasks, we investigate the ability of high-performance language models such as BERT, XLNet, RoBERTa, and ELMo in handling different types of input perturbations. The results suggest that language models are sensitive to input perturbations and their performance can decrease even when small changes are introduced. We highlight that models need to be further improved and that current benchmarks are not reflecting model robustness well. We argue that evaluations on perturbed inputs should routinely complement widely-used benchmarks in order to yield a more realistic understanding of NLP systems' robustness.
This paper takes a first step towards a critical thinking curriculum for neural auto-regressive language models. We introduce a synthetic corpus of deductively valid arguments, and generate artificial argumentative texts to train CRiPT: a critical th
inking intermediarily pre-trained transformer based on GPT-2. Significant transfer learning effects can be observed: Trained on three simple core schemes, CRiPT accurately completes conclusions of different, and more complex types of arguments, too. CRiPT generalizes the core argument schemes in a correct way. Moreover, we obtain consistent and promising results for NLU benchmarks. In particular, CRiPT's zero-shot accuracy on the GLUE diagnostics exceeds GPT-2's performance by 15 percentage points. The findings suggest that intermediary pre-training on texts that exemplify basic reasoning abilities (such as typically covered in critical thinking textbooks) might help language models to acquire a broad range of reasoning skills. The synthetic argumentative texts presented in this paper are a promising starting point for building such a critical thinking curriculum for language models.''
Using data from English cloze tests, in which subjects also self-reported their gender, age, education, and race, we examine performance differences of pretrained language models across demographic groups, defined by these (protected) attributes. We
demonstrate wide performance gaps across demographic groups and show that pretrained language models systematically disfavor young non-white male speakers; i.e., not only do pretrained language models learn social biases (stereotypical associations) -- pretrained language models also learn sociolectal biases, learning to speak more like some than like others. We show, however, that, with the exception of BERT models, larger pretrained language models reduce some the performance gaps between majority and minority groups.
We describe the systems developed by the National Research Council Canada for the Uralic language identification shared task at the 2021 VarDial evaluation campaign. We evaluated two different approaches to this task: a probabilistic classifier explo
iting only character 5-grams as features, and a character-based neural network pre-trained through self-supervision, then fine-tuned on the language identification task. The former method turned out to perform better, which casts doubt on the usefulness of deep learning methods for language identification, where they have yet to convincingly and consistently outperform simpler and less costly classification algorithms exploiting n-gram features.
Successful methods for unsupervised neural machine translation (UNMT) employ cross-lingual pretraining via self-supervision, often in the form of a masked language modeling or a sequence generation task, which requires the model to align the lexical-
and high-level representations of the two languages. While cross-lingual pretraining works for similar languages with abundant corpora, it performs poorly in low-resource and distant languages. Previous research has shown that this is because the representations are not sufficiently aligned. In this paper, we enhance the bilingual masked language model pretraining with lexical-level information by using type-level cross-lingual subword embeddings. Empirical results demonstrate improved performance both on UNMT (up to 4.5 BLEU) and bilingual lexicon induction using our method compared to a UNMT baseline.
suggested questions
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
