Gabriela Ferraro’s research while affiliated with Australian National University and other places

Today, the internet is an integral part of our daily lives, enabling people to be more connected than ever before. However, this greater connectivity and access to information increase exposure to harmful content, such as cyber-bullying and cyber-hatred. Models based on machine learning and natural language offer a way to make online platforms safer by identifying hate speech in web text autonomously. However, the main difficulty is annotating a sufficiently large number of examples to train these models. This paper uses a transfer learning technique to leverage two independent datasets jointly and builds a single representation of hate speech. We build an interpretable two-dimensional visualization tool of the constructed hate speech representation—dubbed the Map of Hate—in which multiple datasets can be projected and comparatively analyzed. The hateful content is annotated differently across the two datasets (racist and sexist in one dataset, hateful and offensive in another). However, the common representation successfully projects the harmless class of both datasets into the same space and can be used to uncover labeling errors (false positives). We also show that the joint representation boosts prediction performances when only a limited amount of supervision is available. These methods and insights hold the potential for safer social media and reduce the need to expose human moderators and annotators to distressing online messaging.

Extracting multiple relations from text sentences is still a challenge for current Open Relation Extraction (Open RE) tasks. In this paper, we develop several Open RE models based on the bidirectional LSTM-CRF (BiLSTM-CRF) neural network and different contextualized word embedding methods. We also propose a new tagging scheme to solve overlapping problems and enhance models' performance. From the evaluation results and comparisons between models, we select the best combination of tagging scheme, word embedder, and BiLSTM-CRF network to achieve an Open RE model with a remarkable extracting ability on multiple-relation sentences.

Neural networks suffer from catastrophic forgetting and are unable to sequentially learn new tasks without guaranteed stationarity in data distribution. Continual learning could be achieved via replay -- by concurrently training externally stored old data while learning a new task. However, replay becomes less effective when each past task is allocated with less memory. To overcome this difficulty, we supplemented replay mechanics with meta-learning for rapid knowledge acquisition. By employing a meta-learner, which \textit{learns a learning rate per parameter per past task}, we found that base learners produced strong results when less memory was available. Additionally, our approach inherited several meta-learning advantages for continual learning: it demonstrated strong robustness to continually learn under the presence of noises and yielded base learners to higher accuracy in less updates.

In this paper we present the problem of a noisy lexical taxonomy and suggest two tasks as potential remedies. The first task is to identify and eliminate incorrect hypernymy links, and the second is to repopulate the taxonomy with new relations. The first task consists of revising the entire taxonomy and returning a Boolean for each assertion of hypernymy between two nouns (e.g. brie is a kind of cheese ). The second task consists of recursively producing a chain of hypernyms for a given noun, until the most general node in the taxonomy is reached (e.g. brie → cheese → food → etc.). In order to achieve these goals, we implemented a hybrid hypernym-detection algorithm that incorporates various intuitions, such as syntagmatic, paradigmatic and morphological association measures as well as lexical patterns. We evaluate these algorithms individually and collectively and report findings in Spanish, English and French.

Background Assisting moderators to triage harmful posts in Internet Support Groups is relevant to ensure its safe use. Automated text classification methods analysing the language expressed in posts of online forums is a promising solution. Methods Natural Language Processing and Machine Learning technologies were used to build a triage post classifier using a dataset from Reachout.com mental health forum for young people. Results When comparing with the state-of-the-art, a solution mainly based on features from lexical resources, received the best classification performance for the crisis posts (52%), which is the most severe class. Six salient linguistic characteristics were found when analysing the crisis post; (1) posts expressing hopelessness, (2) short posts expressing concise negative emotional responses, (3) long posts expressing variations of emotions, (4) posts expressing dissatisfaction with available health services, (5) posts utilising storytelling, and (6) posts expressing users seeking advice from peers during a crisis. Conclusion It is possible to build a competitive triage classifier using features derived only from the textual content of the post. Further research needs to be done in order to translate our quantitative and qualitative findings into features, as it may improve overall performance.

Extracting and formalising legal norms from legal documents is a time-consuming and complex procedure. Therefore, the automatic methods that can accelerate this process are in high demand. In this paper, we address two major questions related to this problem: (i) what are the challenges in formalising legal documents into a machine understandable formalism? (ii) to what extent can the data-driven state-of-the-art approaches developed in the Natural Language Processing (NLP) community be used to automate the normative mining process. The results of our experiments indicate that NLP technologies such as relation extraction and semantic parsing are promising research avenues to advance research in this area.

Learning to learn (L2L) trains a meta-learner to assist the learning of a task-specific base learner. Previously, it was shown that a meta-learner could learn the direct rules to update learner parameters; and that the learnt neural optimiser updated learners more rapidly than handcrafted gradient-descent methods. However, we demonstrate that previous neural optimisers were limited to update learners on one designated dataset. In order to address input-domain heterogeneity, we introduce Multi-Task Learning to Learn (MTL2L), a context aware neural optimiser which self-modifies its optimisation rules based on input data. We show that MTL2L is capable of updating learners to classify on data of an unseen input-domain at the meta-testing phase.

Background: Assisting moderators to triage harmful posts in Internet Support Groups is relevant to ensure its safe use. Automated text classification methods analysing the language expressed in posts of online forums is a promising solution. Methods: Natural Language Processing and Machine Learning technologies were used to build a triage post classifier using a dataset from Reachout.com mental health forum for young people. Results: When comparing with the state-of-the-art, a solution mainly based on features from lexical resources, received the best classification performance for the crisis posts (52%), which is the most severe class. Six salient linguistic characteristics were found when analysing the crisis post; 1) posts expressing hopelessness, 2) short posts expressing concise negative emotional responses, 3) long posts expressing variations of emotions, 4) posts expressing dissatisfaction with available health services, 5) posts utilising storytelling, and 6) posts expressing users seeking advice from peers during a crisis. Conclusion: It is possible to build a competitive triage classifier using features derived only from the textual content of the post. Further research needs to be done in order to translate our quantitative and qualitative findings into features, as it may improve overall performance.