With the rapid advancement of neural language models, the deployment of over-parameterized models has surged, increasing the need for interpretable explanations comprehensible to human inspectors. Existing post-hoc interpretability methods, which often focus on unigram features of single input textual instances, fail to capture the models' decision-making process fully. Additionally, many methods do not differentiate between decisions based on spurious correlations and those based on a holistic understanding of the input. Our paper introduces DISCO, a novel method for discovering global, rule-based explanations by identifying causal n-gram associations with model predictions. This method employs a scalable sequence mining technique to extract relevant text spans from training data, associate them with model predictions, and conduct causality checks to distill robust rules that elucidate model behavior. These rules expose potential overfitting and provide insights into misleading feature combinations. We validate DISCO through extensive testing, demonstrating its superiority over existing methods in offering comprehensive insights into complex model behaviors. Our approach successfully identifies all shortcuts manually introduced into the training data (100% detection rate on the MultiRC dataset), resulting in an 18.8% regression in model performance -- a capability unmatched by any other method. Furthermore, DISCO supports interactive explanations, enabling human inspectors to distinguish spurious causes in the rule-based output. This alleviates the burden of abundant instance-wise explanations and helps assess the model's risk when encountering out-of-distribution (OOD) data.

Building a corpus become an important topic in natural language processing (NLP) and especially for low resource languages (ex: AD), due to the importance that the corpus plays in the development of several tools, such as: Machine Translation Babaali and Salem [2022], Part of speech tagging Chiche and Yitagesu [2022], Named entities recognition Jarrar et al. [2022], etc. in particular with the emergence of techniques based on statistics, machine learning and deep learning. Who exploits this mass of information to develop, train and evaluate models. However, building a corpus is not an easy task Bakari et al. [2016]; it is extremely time-consuming and requires a lot of work, for the good reason that the volume and quality of the corpus are two important parameters. Despite the recent emergence of techniques that consume fewer resources, such as few-shot learning Tunstall et al. [2022]. Over the last few years, a lot of studies in NLP have focused on languages or variants of languages called low resources Mengoni and Santucci [2023]. This change of direction is mainly due to the emergence of social media such as Facebook, Twitter, RenRen, LinkedIn, Google+, and Tuenti, as a means of communication where people exchange messages and comments.

Automatic summarization has consistently attracted attention, due to its versatility and wide application in various downstream tasks. Despite its popularity, we find that annotation efforts have largely been disjointed, and have lacked common terminology. Consequently, it is challenging to discover existing resources or identify coherent research directions. To address this, we survey a large body of work spanning 133 datasets in over 100 languages, creating a novel ontology covering sample properties, collection methods and distribution. With this ontology we make key observations, including the lack in accessible high-quality datasets for low-resource languages, and the field's over-reliance on the news domain and on automatically collected distant supervision. Finally, we make available a web interface that allows users to interact and explore our ontology and dataset collection, as well as a template for a summarization data card, which can be used to streamline future research into a more coherent body of work.

While piano music has become a significant area of study in Music Information Retrieval (MIR), there is a notable lack of datasets for piano solo music with text labels. To address this gap, we present PIAST (PIano dataset with Audio, Symbolic, and Text), a piano music dataset. Utilizing a piano-specific taxonomy of semantic tags, we collected 9,673 tracks from YouTube and added human annotations for 2,023 tracks by music experts, resulting in two subsets: PIAST-YT and PIAST-AT. Both include audio, text, tag annotations, and transcribed MIDI utilizing state-of-the-art piano transcription and beat tracking models. Among many possible tasks with the multi-modal dataset, we conduct music tagging and retrieval using both audio and MIDI data and report baseline performances to demonstrate its potential as a valuable resource for MIR research.

The first piece of AI-generated video I ever made moved me to tears – tears of laughter. Given the chance to fool around with Runway AI's Gen-3 Alpha, I dropped in an image of an eagle carrying off a wolf. Moments later, the picture sprang into life. Except the bird only had one leg – and its plummeting prey sprouted wings from its tail and morphed into a wolf-headed goose. It was weird and hilarious.

On Friday, TriStar Pictures released Here, a 50 million Robert Zemeckis-directed film that used real-time generative AI face transformation techniques to portray actors Tom Hanks and Robin Wright across a 60-year span, marking one of Hollywood's first full-length features built around AI-powered visual effects. The film adapts a 2014 graphic novel set primarily in a New Jersey living room across multiple time periods. Rather than cast different actors for various ages, the production used AI to modify Hanks' and Wright's appearances throughout. The de-aging technology comes from Metaphysic, a visual effects company that creates real time face swapping and aging effects. During filming, the crew watched two monitors simultaneously: one showing the actors' actual appearances and another displaying them at whatever age the scene required.

Nowadays, it's easy to get fooled by AI content on the web. Whether it's a picture of the Pope sporting a puffy Balenciaga jacket or Trump getting tackled and arrested, these AI-generated images appear super realistic (as long as you're not looking too close), so it can be hard to separate fact from fiction. This is because AI doesn't really understand context in the cultural or historical sense. While some AI-generated images are harmless and do not spread misinformation, others, especially ones involving celebrities or politicians, can cause a great deal of damage and brain rot. Heck, I consider myself to be a relatively tech-savvy person and even I've been fooled once or twice.

Short video platforms, such as YouTube, Instagram, or TikTok, are used by billions of users globally. These platforms expose users to harmful content, ranging from clickbait or physical harms to misinformation or online hate. Yet, detecting harmful videos remains challenging due to an inconsistent understanding of what constitutes harm and limited resources and mental tolls involved in human annotation. As such, this study advances measures and methods to detect harm in video content. First, we develop a comprehensive taxonomy for online harm on video platforms, categorizing it into six categories: Information, Hate and harassment, Addictive, Clickbait, Sexual, and Physical harms. Next, we establish multimodal large language models as reliable annotators of harmful videos. We analyze 19,422 YouTube videos using 14 image frames, 1 thumbnail, and text metadata, comparing the accuracy of crowdworkers (Mturk) and GPT-4-Turbo with domain expert annotations serving as the gold standard. Our results demonstrate that GPT-4-Turbo outperforms crowdworkers in both binary classification (harmful vs. harmless) and multi-label harm categorization tasks. Methodologically, this study extends the application of LLMs to multi-label and multi-modal contexts beyond text annotation and binary classification. Practically, our study contributes to online harm mitigation by guiding the definitions and identification of harmful content on video platforms.

The unchecked spread of digital information, combined with increasing political polarization and the tendency of individuals to isolate themselves from opposing political viewpoints, has driven researchers to develop systems for automatically detecting political bias in media. This trend has been further fueled by discussions on social media. We explore methods for categorizing bias in US news articles, comparing rule-based and deep learning approaches. The study highlights the sensitivity of modern self-learning systems to unconstrained data ingestion, while reconsidering the strengths of traditional rule-based systems. Applying both models to left-leaning (CNN) and right-leaning (FOX) news articles, we assess their effectiveness on data beyond the original training and test sets.This analysis highlights each model's accuracy, offers a framework for exploring deep-learning explainability, and sheds light on political bias in US news media. We contrast the opaque architecture of a deep learning model with the transparency of a linguistically informed rule-based model, showing that the rule-based model performs consistently across different data conditions and offers greater transparency, whereas the deep learning model is dependent on the training set and struggles with unseen data.

Leveraging generative AI (for example, Large Language Models) for language understanding within robotics opens up possibilities for LLM-driven robot end-user development (EUD). Despite the numerous design opportunities it provides, little is understood about how this technology can be utilized when constructing robot program logic. In this paper, we outline the background in capturing natural language end-user intent and summarize previous use cases of LLMs within EUD. Taking the context of filmmaking as an example, we explore how a cinematography practitioner's intent to film a certain scene can be articulated using natural language, captured by an LLM, and further parametrized as low-level robot arm movement. We explore the capabilities of an LLM interpreting end-user intent and mapping natural language to predefined, cross-modal data in the process of iterative program development. We conclude by suggesting future opportunities for domain exploration beyond cinematography to support language-driven robotic camera navigation.