Marco Lippi’s research while affiliated with University of Florence and other places

In the context of continual learning, little attention is dedicated to the problem of developing a layer of "concepts", also known as "concept bottleneck", to support the discrimination of higher-level task information, especially when concepts are not supervised. Concept bottleneck discovery in an unsupervised setting is thus largely unexplored, and this paper aims to move a step forward in such direction. We consider a neural network that faces a stream of binary tasks, with no further information on the relationships among them, i.e., no supervisions at the level of concepts. The learning of the concept bottleneck layer is driven by means of a triplet-based criterion, which is instantiated in conjunction with a specifically designed experience replay (concept replay). Such a novel criterion exploits fuzzy Hamming distances to treat vectors of concept probabilities as fuzzy bitstrings, encouraging different concept activations across different tasks, while also adding a regularization effect which pushes probabilities towards crisp values. Despite the lack of concept supervisions, we found that continually learning the streamed tasks in a progressive manner yields the development of inner concepts that are significantly better correlated with the higher-level tasks, compared to the case of joint-offline learning. This result is showcased in an extended experimental activity involving different architectures and newly created (and shared) datasets that are also well-suited to support further investigation of continual learning in concept-based models.

Most of the existing natural language processing systems for legal texts are developed for the English language. Nevertheless, there are several application domains where multiple versions of the same documents are provided in different languages, especially inside the European Union. One notable example is given by Terms of Service (ToS). In this paper, we compare different approaches to the task of detecting potential unfair clauses in ToS across multiple languages. In particular, after developing an annotated corpus and a machine learning classifier for English, we consider and compare several strategies to extend the system to other languages: building a novel corpus and training a novel machine learning system for each language, from scratch; projecting annotations across documents in different languages, to avoid the creation of novel corpora; translating training documents while keeping the original annotations; translating queries at prediction time and relying on the English system only. An extended experimental evaluation conducted on a large, original dataset indicates that the time-consuming task of re-building a novel annotated corpus for each language can often be avoided with no significant degradation in terms of performance.

In this work, we analyse mobile phone variation before and after the 2016 Central Italy earthquake in the affected areas, using Twitter and public reconstruction works data. We create three models and show that Twitter data and the related sentiment on the earthquake, as well as the distribution of emergency houses, can contribute to explaining population variations. Our final Generalised Poisson regression model explains more than 80% of the variance of the population’s variation based on the percentage of negative polarity tweets, the number of emergency houses, the number of negative tweets on the earthquake weighted by the number of residents, number of tweets posted on the earthquake anniversary, the distance from the epicentre and several variables related to public reconstruction works (e.g. school, public housing, hydrological disruption, viability). We found that sentiment on the emergency house can be a proxy for population variation because people who live there did not displace from the crater area. The number of tweets posted during the anniversary day can, instead, indicate negative population variation because the higher the number of tweets, the more people can feel nostalgic after having relocated.

In this paper, we explore the use of machine learning regression algorithms for setup time prediction and we apply them to a real-world scheduling application arising in the color printing industry. As the complexities associated with setup time predictions have received limited attention from the literature, we aim at exploiting a data-driven approach based on machine learning algorithms to enhance the quality of setup time evaluations and narrow the gap between scheduling theory and practice. Using a real-world industrial dataset, we train three different machine learning models: linear regression, random forests, and gradient boosting machines. The experimental results demonstrate that the gradient boosting machine approach obtains the best performance overall, immediately followed by random forests. The accuracy of the obtained predictions shows the effectiveness of the proposed approach in setup time evaluation. The obtained results are particularly significant and valuable due to the versatile nature of the proposed machine learning approaches. These methods can be applied to various scheduling scenarios, making them suitable for integration into scheduling algorithms to potentially improve their accuracy.

We explore the use of residual networks and neural attention for multiple argument mining tasks. We propose a residual architecture that exploits attention, multi-task learning, and makes use of ensemble, without any assumption on document or argument structure. We present an extensive experimental evaluation on five different corpora of user-generated comments, scientific publications, and persuasive essays. Our results show that our approach is a strong competitor against state-of-the-art architectures with a higher computational footprint or corpus-specific design, representing an interesting compromise between generality, performance accuracy and reduced model size.

Recent advances in robotics have allowed the introduction of robots assisting and working together with human subjects. To promote their use and diffusion, intuitive and user-friendly interaction means should be adopted. In particular, gestures have become an established way to interact with robots since they allow to command them in an intuitive manner. In this article, we focus on the problem of gesture recognition in human–robot interaction (HRI). While this problem has been largely studied in the literature, it poses specific constraints when applied to HRI. We propose a framework consisting in a pipeline devised to take into account these specific constraints. We implement the proposed pipeline considering, as an example, an evaluation use case. To this end, we consider standard machine learning algorithms for the classification stage and evaluate their performance considering different performance metrics for a thorough assessment.

In this work, we analyse mobile phone variation before and after the 2016 Central Italy earthquake in the affected areas, using Twitter and public reconstruction works data. We create three models and show that Twitter data and the related sentiment on the earthquake, as well as the distribution emergency houses, can contribute to explaining population variations. Our final Poisson regression model explains more than 80% of the variance of the population's variation based on the percentage of negative polarity tweets, the number of emergency houses, the number of negative tweets on the earthquake weighted by the number of residents, number of tweets posted on the earthquake anniversary, the distance from the epicentre and several variables related to public reconstruction works (e.g., school, public housing, hydrological disruption, viability). We found that sentiment on the emergency house can be a proxy for population variation because people who live there did not displace from the crater area. The number of tweets posted during the anniversary day can, instead, indicate negative population variation because the higher the number of tweets, the more people can feel nostalgic after having relocated.

This paper considers the tasks of detecting and recognizing bearing faults in electric motors from the signals collected from supply currents, using machine learning techniques. In particular, following recent trends in AI, the main point of interest was focused towards interpretable solutions that provide explanations on the decisions taken by the classifiers. For this reason, decision trees were chosen, since they represent a classic machine learning approach which inductively learns tree structures from a collection of observations. Paths along the learnt trees can be easily interpreted as plain classification rules. An extensive experimental comparison shows the strong generalization capabilities of such a classifier. In particular, the present work reports results obtained in a highly challenging scenario, usually overlooked in the literature, where the system is tested on configurations of radial and torsional loads that have not been observed during training. The proposed approach achieves over 90% of accuracy even on this cross-load generalization setting.