Dean Ninalga

Dean Ninalga

Detecting transphobia, homophobia, and various other forms of hate speech is difficult. Signals can vary depending on factors such as language, culture, geographical region, and the particular online platform. Here, we present a joint multilingual (M-L) and language-specific (L-S) approach to homophobia and transphobic hate speech detection (HSD). M-L models are needed to catch words, phrases, and concepts that are less common or missing in a particular language and subsequently overlooked by L-S models. Nonetheless, L-S models are better situated to understand the cultural and linguistic context of the users who typically write in a particular language. Here we construct a simple and successful way to merge the M-L and L-S approaches through simple weight interpolation in such a way that is interpretable and data-driven. We demonstrate our system on task A of the 'Shared Task on Homophobia/Transphobia Detection in social media comments' dataset for homophobia and transphobic HSD. Our system achieves the best results in three of five languages and achieves a 0.997 macro average F1-score on Malayalam texts.

Dean Ninalga

This paper presents a state-of-the-art solution to the LongEval CLEF 2023 Lab Task 2: LongEval-Classification. The goal of this task is to improve and preserve the performance of sentiment analysis models across shorter and longer time periods. Our framework feeds date-prefixed textual inputs to a pre-trained language model, where the timestamp is included in the text. We show date-prefixed samples better conditions model outputs on the temporal context of the respective texts. Moreover, we further boost performance by performing self-labeling on unlabeled data to train a student model. We augment the self-labeling process using a novel augmentation strategy leveraging the date-prefixed formatting of our samples. We demonstrate concrete performance gains on the LongEval-Classification evaluation set over non-augmented self-labeling. Our framework achieves a 2nd place ranking with an overall score of 0.6923 and reports the best Relative Performance Drop (RPD) of -0.0656 over the short evaluation set.

Dean Ninalga

Depression is debilitating, and not uncommon. Indeed, studies of excessive social media users show correlations with depression, ADHD, and other mental health concerns. Given that there is a large number of people with excessive social media usage, then there is a significant population of potentially undiagnosed users and posts that they create. In this paper, we propose a depression severity detection system using a semi-supervised learning technique to predict if a post is from a user who is experiencing severe, moderate, or low (non-diagnostic) levels of depression. Namely, we use a trained model to classify a large number of unlabelled social media posts from Reddit, then use these generated labels to train a more powerful classifier. We demonstrate our framework on Detecting Signs of Depression from Social Media Text - LT-EDI@RANLP 2023 shared task, where our framework ranks 3rd overall.

Dean Ninalga

Recent work suggests that interpolating between the weights of two specialized language models can transfer knowledge between tasks in a way that multi-task learning cannot. However, very few have explored interpolation between more than two models, where each has a distinct knowledge base. In this paper, we introduce Derivative Free Weight-space Ensembling (DFWE), a new few-sample task transfer approach for open-domain dialogue. Our framework creates a set of diverse expert language models trained using a predefined set of source tasks. Next, we finetune each of the expert models on the target task, approaching the target task from several distinct knowledge bases. Finally, we linearly interpolate between the model weights using a gradient-free-optimization algorithm, to efficiently find a good interpolation weighting. We demonstrate the effectiveness of the method on FETA-Friends outperforming the standard pretrain-finetune approach.

Milad Leyli-Abadi, Jérôme Picault, Antoine Marot et al.

This paper addresses the growing computational challenges of power grid simulations, particularly with the increasing integration of renewable energy sources like wind and solar. As grid operators must analyze significantly more scenarios in near real-time to prevent failures and ensure stability, traditional physical-based simulations become computationally impractical. To tackle this, a competition was organized to develop AI-driven methods that accelerate power flow simulations by at least an order of magnitude while maintaining operational reliability. This competition utilized a regional-scale grid model with a 30\% renewable energy mix, mirroring the anticipated near-future composition of the French power grid. A key contribution of this work is through the use of LIPS (Learning Industrial Physical Systems), a benchmarking framework that evaluates solutions based on four critical dimensions: machine learning performance, physical compliance, industrial readiness, and generalization to out-of-distribution scenarios. The paper provides a comprehensive overview of the Machine Learning for Physical Simulation (ML4PhySim) competition, detailing the benchmark suite, analyzing top-performing solutions that outperformed traditional simulation methods, and sharing key organizational insights and best practices for running large-scale AI competitions. Given the promising results achieved, the study aims to inspire further research into more efficient, scalable, and sustainable power network simulation methodologies.