2 |
WebSRC: A Dataset for Web-Based Structural Reading Comprehension ...
|
|
|
|
BASE
|
|
Show details
|
|
3 |
Semantic Categorization of Social Knowledge for Commonsense Question Answering ...
|
|
|
|
BASE
|
|
Show details
|
|
4 |
ESTER: A Machine Reading Comprehension Dataset for Reasoning about Event Semantic Relations ...
|
|
|
|
BASE
|
|
Show details
|
|
5 |
CrossVQA: Scalably Generating Benchmarks for Systematically Testing VQA Generalization ...
|
|
|
|
BASE
|
|
Show details
|
|
7 |
PAQ: 65 Million Probably-Asked Questions and What You Can Do With Them ...
|
|
|
|
BASE
|
|
Show details
|
|
8 |
Enhancing Multiple-choice Machine Reading Comprehension by Punishing Illogical Interpretations ...
|
|
|
|
BASE
|
|
Show details
|
|
9 |
Mapping probability word problems to executable representations ...
|
|
|
|
BASE
|
|
Show details
|
|
10 |
Contrastive Domain Adaptation for Question Answering using Limited Text Corpora ...
|
|
|
|
BASE
|
|
Show details
|
|
11 |
Smoothing Dialogue States for Open Conversational Machine Reading ...
|
|
|
|
BASE
|
|
Show details
|
|
12 |
How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answering ...
|
|
|
|
BASE
|
|
Show details
|
|
14 |
FiD-Ex: Improving Sequence-to-Sequence Models for Extractive Rationale Generation ...
|
|
|
|
BASE
|
|
Show details
|
|
15 |
Improving Question Answering Model Robustness with Synthetic Adversarial Data Generation ...
|
|
|
|
Abstract:
Anthology paper link: https://aclanthology.org/2021.emnlp-main.696/ Abstract: Despite recent progress, state-of-the-art question answering models remain vulnerable to a variety of adversarial attacks. While dynamic adversarial data collection, in which a human annotator tries to write examples that fool a model-in-the-loop, can improve model robustness, this process is expensive which limits the scale of the collected data. In this work, we are the first to use synthetic adversarial data generation to make question answering models more robust to human adversaries. We develop a data generation pipeline that selects source passages, identifies candidate answers, generates questions, then finally filters or re-labels them to improve quality. Using this approach, we amplify a smaller human-written adversarial dataset to a much larger set of synthetic question-answer pairs. By incorporating our synthetic data, we improve the state-of-the-art on the AdversarialQA dataset by 3.7F1 and improve model generalisation ...
|
|
Keyword:
Computational Linguistics; Machine Learning; Machine Learning and Data Mining; Natural Language Processing; Question-Answering Systems
|
|
URL: https://dx.doi.org/10.48448/f04n-c312 https://underline.io/lecture/37811-improving-question-answering-model-robustness-with-synthetic-adversarial-data-generation
|
|
BASE
|
|
Hide details
|
|
16 |
Back-Training excels Self-Training at Unsupervised Domain Adaptation of Question Generation and Passage Retrieval ...
|
|
|
|
BASE
|
|
Show details
|
|
17 |
Zero-Shot Dialogue State Tracking via Cross-Task Transfer ...
|
|
|
|
BASE
|
|
Show details
|
|
18 |
Case-based Reasoning for Natural Language Queries over Knowledge Bases ...
|
|
|
|
BASE
|
|
Show details
|
|
19 |
Will this Question be Answered? Question Filtering via Answer Model Distillation for Efficient Question Answering ...
|
|
|
|
BASE
|
|
Show details
|
|
20 |
Data-QuestEval: A Referenceless Metric for Data-to-Text Semantic Evaluation ...
|
|
|
|
BASE
|
|
Show details
|
|
|
|