People constantly use language to learn about the world. Computational linguists have capitalized on this fact to build large language models (LLMs) that acquire co-occurrence-based knowledge from language corpora. LLMs achieve impressive performance on many tasks, but the robustness of their world knowledge has been questioned. Here, we ask: do LLMs acquire generalized knowledge about real-world events? Using curated sets of minimal sentence pairs (n=1215), we tested whether LLMs are more likely to generate plausible event descriptions compared to their implausible counterparts. We found that LLMs systematically distinguish possible and impossible events (The teacher bought the laptop vs. The laptop bought the teacher) but fall short of human performance when distinguishing likely and unlikely events (The nanny tutored the boy vs. The boy tutored the nanny). In follow-up analyses, we show that (i) LLM scores are driven by both plausibility and surface-level sentence features, (ii) LLMs generalize well across syntactic sentence variants (active vs passive) but less well across semantic sentence variants (synonymous sentences), (iii) some, but not all LLM deviations from ground-truth labels align with crowdsourced human judgments, and (iv) explicit event plausibility information emerges in middle LLM layers and remains high thereafter. Overall, our analyses reveal a gap in LLMs' event knowledge, highlighting their limitations as generalized knowledge bases. We conclude by speculating that the differential performance on impossible vs. unlikely events is not a temporary setback but an inherent property of LLMs, reflecting a fundamental difference between linguistic knowledge and world knowledge in intelligent systems.

Selecting the number of topics in LDA models is considered to be a difficult task, for which alternative approaches have been proposed. The performance of the recently developed singular Bayesian information criterion (sBIC) is evaluated and compared to the performance of alternative model selection criteria. The sBIC is a generalization of the standard BIC that can be implemented to singular statistical models. The comparison is based on Monte Carlo simulations and carried out for several alternative settings, varying with respect to the number of topics, the number of documents and the size of documents in the corpora. Performance is measured using different criteria which take into account the correct number of topics, but also whether the relevant topics from the DGPs are identified. Practical recommendations for LDA model selection in applications are derived.

We describe a Physics-Informed Neural Network (PINN) that simulates the flow induced by the astronomical tide in a synthetic port channel, with dimensions based on the Santos - S\~ao Vicente - Bertioga Estuarine System. PINN models aim to combine the knowledge of physical systems and data-driven machine learning models. This is done by training a neural network to minimize the residuals of the governing equations in sample points. In this work, our flow is governed by the Navier-Stokes equations with some approximations. There are two main novelties in this paper. First, we design our model to assume that the flow is periodic in time, which is not feasible in conventional simulation methods. Second, we evaluate the benefit of resampling the function evaluation points during training, which has a near zero computational cost and has been verified to improve the final model, especially for small batch sizes. Finally, we discuss some limitations of the approximations used in the Navier-Stokes equations regarding the modeling of turbulence and how it interacts with PINNs.

Task-oriented dialogue (TOD) systems have been applied in a range of domains to support human users to achieve specific goals. Systems are typically constructed for a single domain or language and do not generalise well beyond this. Their extension to other languages in particular is restricted by the lack of available training data for many of the world's languages. To support work on Natural Language Understanding (NLU) in TOD across multiple languages and domains simultaneously, we constructed MULTI3NLU++, a multilingual, multi-intent, multi-domain dataset. MULTI3NLU++ extends the English-only NLU++ dataset to include manual translations into a range of high, medium and low resource languages (Spanish, Marathi, Turkish and Amharic), in two domains (banking and hotels). MULTI3NLU++ inherits the multi-intent property of NLU++, where an utterance may be labelled with multiple intents, providing a more realistic representation of a user's goals and aligning with the more complex tasks that commercial systems aim to model. We use MULTI3NLU++ to benchmark state-of-the-art multilingual language models as well as Machine Translation and Question Answering systems for the NLU task of intent detection for TOD systems in the multilingual setting. The results demonstrate the challenging nature of the dataset, particularly in the low-resource language setting.

Quantum computing is a promising paradigm based on quantum theory for performing fast computations. Quantum algorithms are expected to surpass their classical counterparts in terms of computational complexity for certain tasks, including machine learning. In this paper, we design, implement, and evaluate three hybrid quantum k-Means algorithms, exploiting different degree of parallelism. Indeed, each algorithm incrementally leverages quantum parallelism to reduce the complexity of the cluster assignment step up to a constant cost. In particular, we exploit quantum phenomena to speed up the computation of distances. The core idea is that the computation of distances between records and centroids can be executed simultaneously, thus saving time, especially for big datasets. We show that our hybrid quantum k-Means algorithms can be more efficient than the classical version, still obtaining comparable clustering results.

This project explores the feasibility of remote patient monitoring based on the analysis of 3D movements captured with smartwatches. We base our analysis on the Kinematic Theory of Rapid Human Movement. We have validated our research in a real case scenario for stroke rehabilitation at the Guttmann Institute5 (neurorehabilitation hospital), showing promising results. Our work could have a great impact in remote healthcare applications, improving the medical efficiency and reducing the healthcare costs. Future steps include more clinical validation, developing multi-modal analysis architectures (analysing data from sensors, images, audio, etc.), and exploring the application of our technology to monitor other neurodegenerative diseases.

Federated Learning (FL) has emerged as a promising distributed learning paradigm with an added advantage of data privacy. With the growing interest in having collaboration among data owners, FL has gained significant attention of organizations. The idea of FL is to enable collaborating participants train machine learning (ML) models on decentralized data without breaching privacy. In simpler words, federated learning is the approach of ``bringing the model to the data, instead of bringing the data to the mode''. Federated learning, when applied to data which is partitioned vertically across participants, is able to build a complete ML model by combining local models trained only using the data with distinct features at the local sites. This architecture of FL is referred to as vertical federated learning (VFL), which differs from the conventional FL on horizontally partitioned data. As VFL is different from conventional FL, it comes with its own issues and challenges. In this paper, we present a structured literature review discussing the state-of-the-art approaches in VFL. Additionally, the literature review highlights the existing solutions to challenges in VFL and provides potential research directions in this domain.

Generative adversarial networks are a promising tool for image generation in the astronomy domain. Of particular interest are conditional generative adversarial networks (cGANs), which allow you to divide images into several classes according to the value of some property of the image, and then specify the required class when generating new images. In the case of images from Imaging Atmospheric Cherenkov Telescopes (IACTs), an important property is the total brightness of all image pixels (image size), which is in direct correlation with the energy of primary particles. We used a cGAN technique to generate images similar to whose obtained in the TAIGA-IACT experiment. As a training set, we used a set of two-dimensional images generated using the TAIGA Monte Carlo simulation software. We artificiallly divided the training set into 10 classes, sorting images by size and defining the boundaries of the classes so that the same number of images fall into each class. These classes were used while training our network. The paper shows that for each class, the size distribution of the generated images is close to normal with the mean value located approximately in the middle of the corresponding class. We also show that for the generated images, the total image size distribution obtained by summing the distributions over all classes is close to the original distribution of the training set. The results obtained will be useful for more accurate generation of realistic synthetic images similar to the ones taken by IACTs.

This paper presents The Shared Task on Euphemism Detection for the Third Workshop on Figurative Language Processing (FigLang 2022) held in conjunction with EMNLP 2022. Participants were invited to investigate the euphemism detection task: given input text, identify whether it contains a euphemism. The input data is a corpus of sentences containing potentially euphemistic terms (PETs) collected from the GloWbE corpus (Davies and Fuchs, 2015), and are human-annotated as containing either a euphemistic or literal usage of a PET. In this paper, we present the results and analyze the common themes, methods and findings of the participating teams

A significant level of stigma and inequality exists in mental healthcare, especially in under-served populations, which spreads through collected data. When not properly accounted for, machine learning (ML) models learned from data can reinforce the structural biases already present in society. Here, we present a systematic study of bias in ML models designed to predict depression in four different case studies covering different countries and populations. We find that standard ML approaches show regularly biased behaviors. However, we show that standard mitigation techniques, and our own post-hoc method, can be effective in reducing the level of unfair bias. We provide practical recommendations to develop ML models for depression risk prediction with increased fairness and trust in the real world. No single best ML model for depression prediction provides equality of outcomes. This emphasizes the importance of analyzing fairness during model selection and transparent reporting about the impact of debiasing interventions.