arXiv subjects

Karl Aberer

Publications and source records attributed to Karl Aberer.

At least 37 records · Page 2Linked to original sources

arXiv2021

Legal Transformer Models May Not Always Help

Deep learning-based Natural Language Processing methods, especially transformers, have achieved impressive performance in the last few years. Applying those state-of-the-art NLP methods to legal activities to automate or simplify some simple work is of great value. This work investigates the value of domain adaptive pre-training and language adapters in legal NLP tasks. By comparing the performance of language models with domain adaptive pre-training on different tasks and different dataset splits, we show that domain adaptive pre-training is only helpful with low-resource downstream tasks, thus far from being a panacea. We also benchmark the performance of adapters in a typical legal NLP task and show that they can yield similar performance to full model tuning with much smaller training costs. As an additional result, we release LegalRoBERTa, a RoBERTa model further pre-trained on legal corpora.

cs.CL↗

arXiv2021

Direction is what you need: Improving Word Embedding Compression in Large Language Models

The adoption of Transformer-based models in natural language processing (NLP) has led to great success using a massive number of parameters. However, due to deployment constraints in edge devices, there has been a rising interest in the compression of these models to improve their inference time and memory footprint. This paper presents a novel loss objective to compress token embeddings in the Transformer-based models by leveraging an AutoEncoder architecture. More specifically, we emphasize the importance of the direction of compressed embeddings with respect to original uncompressed embeddings. The proposed method is task-agnostic and does not require further language modeling pre-training. Our method significantly outperforms the commonly used SVD-based matrix-factorization approach in terms of initial language model Perplexity. Moreover, we evaluate our proposed approach over SQuAD v1.1 dataset and several downstream tasks from the GLUE benchmark, where we also outperform the baseline in most scenarios. Our code is public.

cs.CL↗

arXiv2021

Tactical Reframing of Online Disinformation Campaigns Against The Istanbul Convention

In March 2021, Turkey withdrew from The Istanbul Convention, a human-rights treaty that addresses violence against women, citing issues with the convention's implicit recognition of sexual and gender minorities. In this work, we trace disinformation campaigns related to the Istanbul Convention and its associated Turkish law that circulate on divorced men's rights Facebook groups. We find that these groups adjusted the narrative and focus of the campaigns to appeal to a larger audience, which we refer to as "tactical reframing." Initially, the men organized in a grass-roots manner to campaign against the Turkish law that was passed to codify the convention, focusing on one-sided custody of children and indefinite alimony. Later, they reframed their campaign and began attacking the Istanbul Convention, highlighting its acknowledgment of homosexuality. This case study highlights how disinformation campaigns can be used to weaponize homophobia in order to limit the rights of women. To the best of our knowledge, this is the first case study that analyzes a narrative reframing in the context of a disinformation campaign on social media.

cs.CY↗

arXiv2021

On Representation Learning for Scientific News Articles Using Heterogeneous Knowledge Graphs

In the era of misinformation and information inflation, the credibility assessment of the produced news is of the essence. However, fact-checking can be challenging considering the limited references presented in the news. This challenge can be transcended by utilizing the knowledge graph that is related to the news articles. In this work, we present a methodology for creating scientific news article representations by modeling the directed graph between the scientific news articles and the cited scientific publications. The network used for the experiments is comprised of the scientific news articles, their topic, the cited research literature, and their corresponding authors. We implement and present three different approaches: 1) a baseline Relational Graph Convolutional Network (R-GCN), 2) a Heterogeneous Graph Neural Network (HetGNN) and 3) a Heterogeneous Graph Transformer (HGT). We test these models in the downstream task of link prediction on the: a) news article - paper links and b) news article - article topic links. The results show promising applications of graph neural network approaches in the domains of knowledge tracing and scientific news credibility assessment.

cs.SI↗

arXiv2021

A Dataset of State-Censored Tweets

Many governments impose traditional censorship methods on social media platforms. Instead of removing it completely, many social media companies, including Twitter, only withhold the content from the requesting country. This makes such content still accessible outside of the censored region, allowing for an excellent setting in which to study government censorship on social media. We mine such content using the Internet Archive's Twitter Stream Grab. We release a dataset of 583,437 tweets by 155,715 users that were censored between 2012-2020 July. We also release 4,301 accounts that were censored in their entirety. Additionally, we release a set of 22,083,759 supplemental tweets made up of all tweets by users with at least one censored tweet as well as instances of other users retweeting the censored user. We provide an exploratory analysis of this dataset. Our dataset will not only aid in the study of government censorship but will also aid in studying hate speech detection and the effect of censorship on social media users. The dataset is publicly available at https://doi.org/10.5281/zenodo.4439509

cs.SI↗

arXiv2021

Ephemeral Astroturfing Attacks: The Case of Fake Twitter Trends

We uncover a previously unknown, ongoing astroturfing attack on the popularity mechanisms of social media platforms: ephemeral astroturfing attacks. In this attack, a chosen keyword or topic is artificially promoted by coordinated and inauthentic activity to appear popular, and, crucially, this activity is removed as part of the attack. We observe such attacks on Twitter trends and find that these attacks are not only successful but also pervasive. We detected over 19,000 unique fake trends promoted by over 108,000 accounts, including not only fake but also compromised accounts, many of which remained active and continued participating in the attacks. Trends astroturfed by these attacks account for at least 20% of the top 10 global trends. Ephemeral astroturfing threatens the integrity of popularity mechanisms on social media platforms and by extension the integrity of the platforms.

cs.CR↗

arXiv2020

SciLens News Platform: A System for Real-Time Evaluation of News Articles

We demonstrate the SciLens News Platform, a novel system for evaluating the quality of news articles. The SciLens News Platform automatically collects contextual information about news articles in real-time and provides quality indicators about their validity and trustworthiness. These quality indicators derive from i) social media discussions regarding news articles, showcasing the reach and stance towards these articles, and ii) their content and their referenced sources, showcasing the journalistic foundations of these articles. Furthermore, the platform enables domain-experts to review articles and rate the quality of news sources. This augmented view of news articles, which combines automatically extracted indicators and domain-expert reviews, has provably helped the platform users to have a better consensus about the quality of the underlying articles. The platform is built in a distributed and robust fashion and runs operationally handling daily thousands of news articles. We evaluate the SciLens News Platform on the emerging topic of COVID-19 where we highlight the discrepancies between low and high-quality news outlets based on three axes, namely their newsroom activity, evidence seeking and social engagement. A live demonstration of the platform can be found here: http://scilens.epfl.ch.

cs.IR↗

arXiv2020

Spoken dialect identification in Twitter using a multi-filter architecture

This paper presents our approach for SwissText & KONVENS 2020 shared task 2, which is a multi-stage neural model for Swiss German (GSW) identification on Twitter. Our model outputs either GSW or non-GSW and is not meant to be used as a generic language identifier. Our architecture consists of two independent filters where the first one favors recall, and the second one filter favors precision (both towards GSW). Moreover, we do not use binary models (GSW vs. not-GSW) in our filters but rather a multi-class classifier with GSW being one of the possible labels. Our model reaches F1-score of 0.982 on the test set of the shared task.

cs.CL↗

arXiv2020

Upgrading the Newsroom: An Automated Image Selection System for News Articles

We propose an automated image selection system to assist photo editors in selecting suitable images for news articles. The system fuses multiple textual sources extracted from news articles and accepts multilingual inputs. It is equipped with char-level word embeddings to help both modeling morphologically rich languages, e.g. German, and transferring knowledge across nearby languages. The text encoder adopts a hierarchical self-attention mechanism to attend more to both keywords within a piece of text and informative components of a news article. We extensively experiment with our system on a large-scale text-image database containing multimodal multilingual news articles collected from Swiss local news media websites. The system is compared with multiple baselines with ablation studies and is shown to beat existing text-image retrieval methods in a weakly-supervised learning setting. Besides, we also offer insights on the advantage of using multiple textual sources and multilingual data.

cs.MM↗

arXiv2020

Can Celebrities Burst Your Bubble?

Polarization is a growing, global problem. As such, many social media based solutions have been proposed in order to reduce it. In this study, we propose a new solution that recommends topics to celebrities to encourage them to join a polarized debate and increase exposure to contrarian content - bursting the filter bubble. Using a state-of-the art model that quantifies the degree of polarization, this paper makes a first attempt to empirically answer the question: Can celebrities burst filter bubbles? We use a case study to analyze how people react when celebrities are involved in a controversial topic and conclude with a list possible research directions.

cs.SI↗

arXiv2019

On Node Features for Graph Neural Networks

Graph neural network (GNN) is a deep model for graph representation learning. One advantage of graph neural network is its ability to incorporate node features into the learning process. However, this prevents graph neural network from being applied into featureless graphs. In this paper, we first analyze the effects of node features on the performance of graph neural network. We show that GNNs work well if there is a strong correlation between node features and node labels. Based on these results, we propose new feature initialization methods that allows to apply graph neural network to non-attributed graphs. Our experimental results show that the artificial features are highly competitive with real features.

cs.LG↗

arXiv2019

Aligning Multilingual Word Embeddings for Cross-Modal Retrieval Task

In this paper, we propose a new approach to learn multimodal multilingual embeddings for matching images and their relevant captions in two languages. We combine two existing objective functions to make images and captions close in a joint embedding space while adapting the alignment of word embeddings between existing languages in our model. We show that our approach enables better generalization, achieving state-of-the-art performance in text-to-image and image-to-text retrieval task, and caption-caption similarity task. Two multimodal multilingual datasets are used for evaluation: Multi30k with German and English captions and Microsoft-COCO with English and Japanese captions.

cs.CL↗

arXiv2019

Parallel Computation of Graph Embeddings

Graph embedding aims at learning a vector-based representation of vertices that incorporates the structure of the graph. This representation then enables inference of graph properties. Existing graph embedding techniques, however, do not scale well to large graphs. We therefore propose a framework for parallel computation of a graph embedding using a cluster of compute nodes with resource constraints. We show how to distribute any existing embedding technique by first splitting a graph for any given set of constrained compute nodes and then reconciling the embedding spaces derived for these subgraphs. We also propose a new way to evaluate the quality of graph embeddings that is independent of a specific inference task. Based thereon, we give a formal bound on the difference between the embeddings derived by centralised and parallel computation. Experimental results illustrate that our approach for parallel computation scales well, while largely maintaining the embedding quality.

cs.LG↗

arXiv2019

SciLens: Evaluating the Quality of Scientific News Articles Using Social Media and Scientific Literature Indicators

This paper describes, develops, and validates SciLens, a method to evaluate the quality of scientific news articles. The starting point for our work are structured methodologies that define a series of quality aspects for manually evaluating news. Based on these aspects, we describe a series of indicators of news quality. According to our experiments, these indicators help non-experts evaluate more accurately the quality of a scientific news article, compared to non-experts that do not have access to these indicators. Furthermore, SciLens can also be used to produce a completely automated quality score for an article, which agrees more with expert evaluators than manual evaluations done by non-experts. One of the main elements of SciLens is the focus on both content and context of articles, where context is provided by (1) explicit and implicit references on the article to scientific literature, and (2) reactions in social media referencing the article. We show that both contextual elements can be valuable sources of information for determining article quality. The validation of SciLens, done through a combination of expert and non-expert annotation, demonstrates its effectiveness for both semi-automatic and automatic quality evaluation of scientific news.

cs.IR↗

arXiv2019

Selection Bias in News Coverage: Learning it, Fighting it

News entities must select and filter the coverage they broadcast through their respective channels since the set of world events is too large to be treated exhaustively. The subjective nature of this filtering induces biases due to, among other things, resource constraints, editorial guidelines, ideological affinities, or even the fragmented nature of the information at a journalist's disposal. The magnitude and direction of these biases are, however, widely unknown. The absence of ground truth, the sheer size of the event space, or the lack of an exhaustive set of absolute features to measure make it difficult to observe the bias directly, to characterize the leaning's nature and to factor it out to ensure a neutral coverage of the news. In this work, we introduce a methodology to capture the latent structure of media's decision process on a large scale. Our contribution is multi-fold. First, we show media coverage to be predictable using personalization techniques, and evaluate our approach on a large set of events collected from the GDELT database. We then show that a personalized and parametrized approach not only exhibits higher accuracy in coverage prediction, but also provides an interpretable representation of the selection bias. Last, we propose a method able to select a set of sources by leveraging the latent representation. These selected sources provide a more diverse and egalitarian coverage, all while retaining the most actively covered events.

cs.SI↗

arXiv2019

A Dynamic Embedding Model of the Media Landscape

Information about world events is disseminated through a wide variety of news channels, each with specific considerations in the choice of their reporting. Although the multiplicity of these outlets should ensure a variety of viewpoints, recent reports suggest that the rising concentration of media ownership may void this assumption. This observation motivates the study of the impact of ownership on the global media landscape and its influence on the coverage the actual viewer receives. To this end, the selection of reported events has been shown to be informative about the high-level structure of the news ecosystem. However, existing methods only provide a static view into an inherently dynamic system, providing underperforming statistical models and hindering our understanding of the media landscape as a whole. In this work, we present a dynamic embedding method that learns to capture the decision process of individual news sources in their selection of reported events while also enabling the systematic detection of large-scale transformations in the media landscape over prolonged periods of time. In an experiment covering over 580M real-world event mentions, we show our approach to outperform static embedding methods in predictive terms. We demonstrate the potential of the method for news monitoring applications and investigative journalism by shedding light on important changes in programming induced by mergers and acquisitions, policy changes, or network-wide content diffusion. These findings offer evidence of strong content convergence trends inside large broadcasting groups, influencing the news ecosystem in a time of increasing media ownership concentration.

cs.SI↗

arXiv2019

Weakly Supervised Active Learning with Cluster Annotation

In this work, we introduce a novel framework that employs cluster annotation to boost active learning by reducing the number of human interactions required to train deep neural networks. Instead of annotating single samples individually, humans can also label clusters, producing a higher number of annotated samples with the cost of a small label error. Our experiments show that the proposed framework requires 82% and 87% less human interactions for CIFAR-10 and EuroSAT datasets respectively when compared with the fully-supervised training while maintaining similar performance on the test set.

cs.LG↗

arXiv2018

Polisis: Automated Analysis and Presentation of Privacy Policies Using Deep Learning

Privacy policies are the primary channel through which companies inform users about their data collection and sharing practices. These policies are often long and difficult to comprehend. Short notices based on information extracted from privacy policies have been shown to be useful but face a significant scalability hurdle, given the number of policies and their evolution over time. Companies, users, researchers, and regulators still lack usable and scalable tools to cope with the breadth and depth of privacy policies. To address these hurdles, we propose an automated framework for privacy policy analysis (Polisis). It enables scalable, dynamic, and multi-dimensional queries on natural language privacy policies. At the core of Polisis is a privacy-centric language model, built with 130K privacy policies, and a novel hierarchy of neural-network classifiers that accounts for both high-level aspects and fine-grained details of privacy practices. We demonstrate Polisis' modularity and utility with two applications supporting structured and free-form querying. The structured querying application is the automated assignment of privacy icons from privacy policies. With Polisis, we can achieve an accuracy of 88.4% on this task. The second application, PriBot, is the first freeform question-answering system for privacy policies. We show that PriBot can produce a correct answer among its top-3 results for 82% of the test questions. Using an MTurk user study with 700 participants, we show that at least one of PriBot's top-3 answers is relevant to users for 89% of the test questions.

cs.CL↗