Language models can be manipulated by adversarial attacks, which introduce subtle perturbations to input data. While recent attack methods can achieve a relatively high attack success rate (ASR), we’ve observed that the generated adversarial examples have a different data distribution compared with the original examples. Specifically, these adversarial examples exhibit reduced confidence levels and greater divergence from the training data distribution. Consequently, they are easy to detect using straightforward detection methods, diminishing the efficacy of such attacks. To address this issue, we propose a Distribution-Aware Adversarial Attack (DA³) method. DA³ considers the distribution shifts of adversarial examples to improve attacks’ effectiveness under detection methods. We further design a novel evaluation metric, the Non-detectable Attack Success Rate (NASR), which integrates both ASR and detectability for the attack task. We conduct experiments on four widely used datasets to validate the attack effectiveness and transferability of adversarial examples generated by DA³ against both the white-box BERT-base and RoBERTa-base models and the black-box LLaMA2-7b model.

Voice-driven software systems are in abundance. However, language models that power these systems are traditionally trained on fluent, written text corpora. Hence there can be a misalignment between the inherent disfluency of transcribed spoken content and the fluency of the written training data. Furthermore, gold-standard disfluency annotations of various complexities for incremental training can be expensive to collect. So, we propose in this paper a Disfluency Augmented Curriculum Learning (DACL) approach to tackle the complex structure of disfluent sentences and generate fluent texts from them, by using Curriculum Learning (CL) coupled with our synthetically augmented disfluent texts of various levels. DACL harnesses the tiered structure of our generated synthetic disfluent data using CL, by training the model on basic samples (i.e. more fluent) first before training it on more complex samples (i.e. more disfluent). In contrast to the random data exposure paradigm, DACL focuses on a simple-to-complex learning process. We comprehensively evaluate DACL on Switchboard Penn Treebank-3 and compare it to the state-of-the-art disfluency removal models. Our model surpasses existing techniques in word-based precision (by up to 1%) and has shown favorable recall and F1 scores.

Spoken content is abundant – including podcasts, meeting transcripts, and TikTok-like short videos. And yet, many important tasks like summarization are often designed for written content rather than the looser, noiser, and more disfluent style of spoken content. Hence, we aim in this paper to quantify the impact of disfluency on spoken content summarization. Do disfluencies negatively impact the quality of summaries generated by existing approaches? And if so, to what degree? Coupled with these goals, we also investigate two methods towards improving summarization in the presence of such disfluencies. We find that summarization quality does degrade with an increase in these disfluencies and that a combination of multiple disfluency types leads to even greater degradation. Further, our experimental results show that naively removing disfluencies and augmenting with special tags can worsen the summarization when used for testing, but that removing disfluencies for fine-tuning yields the best results. We make the code available at https://github.com/mariateleki/Quantifying-Impact-Disfluency.

A key component of modern conversational systems is the Dialogue State Tracker (or DST), which models a user’s goals and needs. Toward building more robust and reliable DSTs, we introduce a prompt-based learning approach to automatically generate effective adversarial examples to probe DST models. Two key characteristics of this approach are: (i) it only needs the output of the DST with no need for model parameters, and (ii) it can learn to generate natural language utterances that can target any DST. Through experiments over state-of-the-art DSTs, the proposed framework leads to the greatest reduction in accuracy and the best attack success rate while maintaining good fluency and a low perturbation ratio. We also show how much the generated adversarial examples can bolster a DST through adversarial training. These results indicate the strength of prompt-based attacks on DSTs and leave open avenues for continued refinement.

Pre-trained Language Models are widely used in many important real-world applications. However, recent studies show that these models can encode social biases from large pre-training corpora and even amplify biases in downstream applications. To address this challenge, we propose Co²PT, an efficient and effective *debias-while-prompt tuning* method for mitigating biases via counterfactual contrastive prompt tuning on downstream tasks. Our experiments conducted on three extrinsic bias benchmarks demonstrate the effectiveness of Co²PT on bias mitigation during the prompt tuning process and its adaptability to existing upstream debiased language models. These findings indicate the strength of Co²PT and provide promising avenues for further enhancement in bias mitigation on downstream tasks.

Question Generation (QG) is a fundamental NLP task for many downstream applications. Recent studies on open-book QG, where supportive answer-context pairs are provided to models, have achieved promising progress. However, generating natural questions under a more practical closed-book setting that lacks these supporting documents still remains a challenge. In this work, we propose a new QG model for this closed-book setting that is designed to better understand the semantics of long-form abstractive answers and store more information in its parameters through contrastive learning and an answer reconstruction module. Through experiments, we validate the proposed QG model on both public datasets and a new WikiCQA dataset. Empirical results show that the proposed QG model outperforms baselines in both automatic evaluation and human evaluation. In addition, we show how to leverage the proposed model to improve existing question-answering systems. These results further indicate the effectiveness of our QG model for enhancing closed-book question-answering tasks.

This paper presents six document classification models using the latest transformer encoders and a high-performing ensemble model for a task of offensive language identification in social media. For the individual models, deep transformer layers are applied to perform multi-head attentions. For the ensemble model, the utterance representations taken from those individual models are concatenated and fed into a linear decoder to make the final decisions. Our ensemble model outperforms the individual models and shows up to 8.6% improvement over the individual models on the development set. On the test set, it achieves macro-F1 of 90.9% and becomes one of the high performing systems among 85 participants in the sub-task A of this shared task. Our analysis shows that although the ensemble model significantly improves the accuracy on the development set, the improvement is not as evident on the test set.

Free text data from social media is now widely used in natural language processing research, and one of the most common machine learning tasks performed on this data is classification. Generally speaking, performances of supervised classification algorithms on social media datasets are lower than those on texts from other sources, but recently-proposed transformer-based models have considerably improved upon legacy state-of-the-art systems. Currently, there is no study that compares the performances of different variants of transformer-based models on a wide range of social media text classification datasets. In this paper, we benchmark the performances of transformer-based pre-trained models on 25 social media text classification datasets, 6 of which are health-related. We compare three pre-trained language models, RoBERTa-base, BERTweet and ClinicalBioBERT in terms of classification accuracy. Our experiments show that RoBERTa-base and BERTweet perform comparably on most datasets, and considerably better than ClinicalBioBERT, even on health-related datasets.

We present a transformer-based sarcasm detection model that accounts for the context from the entire conversation thread for more robust predictions. Our model uses deep transformer layers to perform multi-head attentions among the target utterance and the relevant context in the thread. The context-aware models are evaluated on two datasets from social media, Twitter and Reddit, and show 3.1% and 7.0% improvements over their baselines. Our best models give the F1-scores of 79.0% and 75.0% for the Twitter and Reddit datasets respectively, becoming one of the highest performing systems among 36 participants in this shared task.