Towards Trustworthy Deception Detection: Benchmarking Model Robustness across Domains, Modalities, and Languages

Evaluating model robustness is critical when developing trustworthy models not only to gain deeper understanding of model behavior, strengths, and weaknesses, but also to develop future models that are generalizable and robust across expected environments a model may encounter in deployment. In this paper we present a framework for measuring model robustness for an important but difficult text classification task - deceptive news detection. We evaluate model robustness to out-of-domain data, modality-specific features, and languages other than English. Our investigation focuses on three type of models: LSTM models trained on multiple datasets(Cross-Domain), several fusion LSTM models trained with images and text and evaluated with three state-of-the-art embeddings, BERT ELMo, and GloVe (Cross-Modality), and character-level CNN models trained on multiple languages (Cross-Language). Our analyses reveal a significant drop in performance when testing neural models on out-of-domain data and non-English languages that may be mitigated using diverse training data. We find that with additional image content as input, ELMo embeddings yield significantly fewer errors compared to BERT orGLoVe. Most importantly, this work not only carefully analyzes deception model robustness but also provides a framework of these analyses that can be applied to new models or extended datasets in the future.

Evaluating Deception Detection Model Robustness To Linguistic Variation

With the increasing use of machine-learning driven algorithmic judgements, it is critical to develop models that are robust to evolving or manipulated inputs. We propose an extensive analysis of model robustness against linguistic variation in the setting of deceptive news detection, an important task in the context of misinformation spread online. We consider two prediction tasks and compare three state-of-the-art embeddings to highlight consistent trends in model performance, high confidence misclassifications, and high impact failures. By measuring the effectiveness of adversarial defense strategies and evaluating model susceptibility to adversarial attacks using character- and word-perturbed text, we find that character or mixed ensemble models are the most effective defenses and that character perturbation-based attack tactics are more successful.

Measure Utility, Gain Trust: Practical Advice for XAI Researcher

Research into the explanation of machine learning models, i.e., explainable AI (XAI), has seen a commensurate exponential growth alongside deep artificial neural networks throughout the past decade. For historical reasons, explanation and trust have been intertwined. However, the focus on trust is too narrow, and has led the research community astray from tried and true empirical methods that produced more defensible scientific knowledge about people and explanations. To address this, we contribute a practical path forward for researchers in the XAI field. We recommend researchers focus on the utility of machine learning explanations instead of trust. We outline five broad use cases where explanations are useful and, for each, we describe pseudo-experiments that rely on objective empirical measurements and falsifiable hypotheses. We believe that this experimental rigor is necessary to contribute to scientific knowledge in the field of XAI.

Evaluating Neural Machine Comprehension Model Robustness to Noisy Inputs and Adversarial Attacks

We evaluate machine comprehension models' robustness to noise and adversarial attacks by performing novel perturbations at the character, word, and sentence level. We experiment with different amounts of perturbations to examine model confidence and misclassification rate, and contrast model performance in adversarial training with different embedding types on two benchmark datasets. We demonstrate improving model performance with ensembling. Finally, we analyze factors that effect model behavior under adversarial training and develop a model to predict model errors during adversarial attacks.

Fishing for Clickbaits in Social Images and Texts with Linguistically-Infused Neural Network Models

This paper presents the results and conclusions of our participation in the Clickbait Challenge 2017 on automatic clickbait detection in social media. We first describe linguistically-infused neural network models and identify informative representations to predict the level of clickbaiting present in Twitter posts. Our models allow to answer the question not only whether a post is a clickbait or not, but to what extent it is a clickbait post e.g., not at all, slightly, considerably, or heavily clickbaity using a score ranging from 0 to 1. We evaluate the predictive power of models trained on varied text and image representations extracted from tweets. Our best performing model that relies on the tweet text and linguistic markers of biased language extracted from the tweet and the corresponding page yields mean squared error (MSE) of 0.04, mean absolute error (MAE) of 0.16 and R2 of 0.43 on the held-out test data. For the binary classification setup (clickbait vs. non-clickbait), our model achieved F1 score of 0.69. We have not found that image representations combined with text yield significant performance improvement yet. Nevertheless, this work is the first to present preliminary analysis of objects extracted using Google Tensorflow object detection API from images in clickbait vs. non-clickbait Twitter posts. Finally, we outline several steps to improve model performance as a part of the future work.