This paper describes the participation of the Genaios team in the monolingual track of Subtask A at SemEval-2024 Task 8. Our best system, LLMixtic, is a Transformer Encoder that mixes token-level probabilistic features extracted from four LLaMA-2 models. We obtained the best results in the official ranking (96.88% accuracy), showing a false positive ratio of 4.38% and a false negative ratio of 1.97% on the test set. We further study LLMixtic through ablation, probabilistic, and attention analyses, finding that (i) performance improves as more LLMs and probabilistic features are included, (ii) LLMixtic puts most attention on the features of the last tokens, (iii) it fails on samples where human text probabilities become consistently higher than for generated text, and (iv) LLMixtic’s false negatives exhibit a bias towards text with newlines.

We introduce PyRater, an open-source Python toolkit designed for analysing corpora annotations. When creating new annotated language resources, probabilistic models of annotation are the state-of-the-art solution for identifying the best annotators, retrieving the gold standard, and more generally separating annotation signal from noise. PyRater offers a unified interface for several such models and includes an API for the addition of new ones. Additionally, the toolkit has built-in functions to read datasets with multiple annotations and plot the analysis outcomes. In this work, we also demonstrate a novel application of PyRater to zero-shot classifiers, where it effectively selects the best-performing prompt. We make PyRater available to the research community.

Data Maps (Swayamdipta, et al. 2020) have emerged as a powerful tool for diagnosing large annotated datasets. Given a model fitted on a dataset, these maps show each data instance from the dataset in a 2-dimensional space defined by a) the model’s confidence in the true class and b) the variability of this confidence. In previous work, confidence and variability are usually computed using training dynamics, which requires the fitting of a strong model to the dataset. In this paper, we introduce a novel approach: Zero-Shot Data Maps based on fast bi-encoder networks. For each data point, confidence on the true label and variability are computed over the members of an ensemble of zero-shot models constructed with different — but semantically equivalent — label descriptions, i.e., textual representations of each class in a given label space. We conduct a comparative analysis of maps compiled using traditional training dynamics and our proposed zero-shot models across various datasets. Our findings reveal that Zero-Shot Data Maps generally match those produced by the traditional method while delivering up to a 14x speedup. The code is available [here](https://github.com/symanto-research/zeroshot-cartography).

We study the problem of building text classifiers with little or no training data, commonly known as zero and few-shot text classification. In recent years, an approach based on neural textual entailment models has been found to give strong results on a diverse range of tasks. In this work, we show that with proper pre-training, Siamese Networks that embed texts and labels offer a competitive alternative. These models allow for a large reduction in inference cost: constant in the number of labels rather than linear. Furthermore, we introduce label tuning, a simple and computationally efficient approach that allows to adapt the models in a few-shot setup by only changing the label embeddings. While giving lower performance than model fine-tuning, this approach has the architectural advantage that a single encoder can be shared by many different tasks.

Emotion Classification is the task of automatically associating a text with a human emotion. State-of-the-art models are usually learned using annotated corpora or rely on hand-crafted affective lexicons. We present an emotion classification model that does not require a large annotated corpus to be competitive. We experiment with pretrained language models in both a zero-shot and few-shot configuration. We build several of such models and consider them as biased, noisy annotators, whose individual performance is poor. We aggregate the predictions of these models using a Bayesian method originally developed for modelling crowdsourced annotations. Next, we show that the resulting system performs better than the strongest individual model. Finally, we show that when trained on few labelled data, our systems outperform fully-supervised models.

According to the self-determination theory, the levels of satisfaction of three basic needs (competence, autonomy and relatedness) have implications on people’s everyday life and career. We benchmark the novel task of automatically detecting those needs on short posts in English, by modelling it as a ternary classification task, and as three binary classification tasks. A detailed manual analysis shows that the latter has advantages in the real-world scenario, and that our best models achieve similar performances as a trained human annotator.

The task of aspect extraction is an important component of aspect-based sentiment analysis. However, it usually requires an expensive human post-processing to ensure quality. In this work we introduce Aspect On, an interactive solution based on online learning that allows users to post-edit the aspect extraction with little effort. The Aspect On interface shows the aspects extracted by a neural model and, given a dataset, annotates its words with the corresponding aspects. Thanks to the online learning, Aspect On updates the model automatically and continuously improves the quality of the aspects displayed to the user. Experimental results show that Aspect On dramatically reduces the number of user clicks and effort required to post-edit the aspects extracted by the model.

In this paper, we present our participation to the EmoContext shared task on detecting emotions in English textual conversations between a human and a chatbot. We propose four neural systems and combine them to further improve the results. We show that our neural ensemble systems can successfully distinguish three emotions (SAD, HAPPY, and ANGRY) and separate them from the rest (OTHERS) in a highly-imbalanced scenario. Our best system achieved a 0.77 F1-score and was ranked fourth out of 165 submissions.

The polarity classification task aims at automatically identifying whether a subjective text is positive or negative. When the target domain is different from those where a model was trained, we refer to a cross-domain setting. That setting usually implies the use of a domain adaptation method. In this work, we study the single and cross-domain polarity classification tasks from the string kernels perspective. Contrary to classical domain adaptation methods, which employ texts from both domains to detect pivot features, we do not use the target domain for training. Our approach detects the lexical peculiarities that characterise the text polarity and maps them into a domain independent space by means of kernel discriminant analysis. Experimental results show state-of-the-art performance in single and cross-domain polarity classification.

We provide several methods for sentence-alignment of texts with different complexity levels. Using the best of them, we sentence-align the Newsela corpora, thus providing large training materials for automatic text simplification (ATS) systems. We show that using this dataset, even the standard phrase-based statistical machine translation models for ATS can outperform the state-of-the-art ATS systems.