Publications

A core endeavour in current affective computing and social signal processing research is the construction of datasets embedding suitable ground truths to foster machine learning methods. This practice brings up hitherto overlooked intricacies. In this paper, we consider causal factors potentially arising when human raters evaluate the affect fluctuations of subjects involved in dyadic interactions and subsequently categorise them in terms of social participation traits. To gauge such factors, we propose an emulator as a statistical approximation of the human rater, and we first discuss the motivations and the rationale behind the approach.The emulator is laid down in the next section as a phenomenological model where the core affect stochastic dynamics as perceived by the rater are captured through an Ornstein-Uhlenbeck process; its parameters are then exploited to infer potential causal effects in the attribution of social traits. Following that, by resorting to a publicly available dataset, the adequacy of the model is evaluated in terms of both human raters' emulation and machine learning predictive capabilities. We then present the results, which are followed by a general discussion concerning findings and their implications, together with advantages and potential applications of the approach.

La presente invenzione riguarda un metodo per stimare automaticamente una posizione conforme della pupilla di un paziente durante l’esecuzione di un esame oftalmico. Il metodo si basa sull’acquisizione di immagini rappresentative della pupilla e sulla loro elaborazione mediante algoritmi di classificazione, comprendenti tecniche di machine learning, al fine di determinare la posizione della pupilla rispetto all’asse ottico di un dispositivo oftalmico o di valutare un parametro di stato della pupilla. L’invenzione riguarda inoltre un dispositivo per esami oftalmici che implementa tale metodo, comprendente un modulo ottico che include uno specchio dicroico configurato per deviare un segnale luminoso rappresentativo della pupilla verso un sensore ottico di acquisizione di immagini, consentendo al contempo ad un ulteriore segnale luminoso rappresentativo della pupilla di propagarsi senza interferenze rilevanti verso componenti ottiche interne del dispositivo oftalmico per l’esecuzione dell’esame di interesse. L’invenzione comprende altresì un kit elettronico collegabile ad un dispositivo oftalmico esistente, che ne consente l’aggiornamento funzionale per l’esecuzione della stima della posizione della pupilla senza alterare le funzionalità diagnostiche originarie. La soluzione proposta migliora l’affidabilità, la ripetibilità e l’usabilità degli esami oftalmici eseguiti da personale specializzato, mantenendo la compatibilità con la strumentazione oftalmica esistente.

The CLIP model has been recently proven to be very effective for a variety of cross-modal tasks, including the evaluation of captions generated from vision-and-language models. In this paper, we propose a new recipe for a contrastive-based evaluation metric for image captioning, namely Positive-Augmented Contrastive learning Score (PAC-S), that in a novel way unifies the learning of a contrastive visual-semantic space with the addition of generated images and text on curated data. Experiments spanning several datasets demonstrate that our new metric achieves the highest correlation with human judgments on both images and videos, outperforming existing reference-based metrics like CIDEr and SPICE and reference-free metrics like CLIP-Score. Finally, we test the system-level correlation of the proposed metric when considering popular image captioning approaches, and assess the impact of employing different cross-modal features. We publicly release our source code and trained models.

Many Machine Learning and Deep Learning algorithms are widely used with remarkable success in scenarios whose benchmark datasets consist of reliable data. However, they often struggle to handle realistic scenarios, particularly those in the financial sector, where available data constantly vary, increase daily, and may contain noise. As a result, we present an overview of the ongoing research at the AImageLab research laboratory of the University of Modena and Reggio Emilia, in collaboration with AxyonAI, focused on exploring Continual Learning methods in the presence of noisy data, with a special focus on noisy labels. To the best of our knowledge, this is a problem that has received limited attention from the scientific community thus far.

A method for localizing terminals is carried out by preparing a plurality of antennas at distinct points of an area to be monitored and acquiring by means of said antennas an identification signal uniquely associated with a terminal present within the area to be monitored. For each antenna, a measurement is made of a received strength signal, RSS, representative of a strength of the identification signal acquired by the antennas and a probability distribution of a position of the terminal with respect to each antenna is generated as a function of the respective RSS. The position of the terminal is determined by maximizing the probability distribution.

Background and objective: The functions of an organism and its biological processes result from the expression of genes and proteins. Therefore quantifying and predicting mRNA and protein levels is a crucial aspect of scientific research. Concerning the prediction of mRNA levels, the available approaches use the sequence upstream and downstream of the Transcription Start Site (TSS) as input to neural networks. The State-of-the-art models (e.g., Xpresso and Basenjii) predict mRNA levels exploiting Convolutional (CNN) or Long Short Term Memory (LSTM) Networks. However, CNN prediction depends on convolutional kernel size, and LSTM suffers from capturing long-range dependencies in the sequence. Concerning the prediction of protein levels, as far as we know, there is no model for predicting protein levels by exploiting the gene or protein sequences. Methods: Here, we exploit a new model type (called Perceiver) for mRNA and protein level prediction, exploiting a Transformer-based architecture with an attention module to attend to long-range interactions in the sequences. In addition, the Perceiver model overcomes the quadratic complexity of the standard Transformer architectures. This work's contributions are 1. DNAPerceiver model to predict mRNA levels from the sequence upstream and downstream of the TSS; 2. ProteinPerceiver model to predict protein levels from the protein sequence; 3. Protein&DNAPerceiver model to predict protein levels from TSS and protein sequences. Results: The models are evaluated on cell lines, mice, glioblastoma, and lung cancer tissues. The results show the effectiveness of the Perceiver-type models in predicting mRNA and protein levels. Conclusions: This paper presents a Perceiver architecture for mRNA and protein level prediction. In the future, inserting regulatory and epigenetic information into the model could improve mRNA and protein level predictions. The source code is freely available at https://github.com/MatteoStefanini/DNAPerceiver.

Image Captioning is the task of translating an input image into a textual description. As such, it connects Vision and Language in a generative fashion, with applications that range from multi-modal search engines to help visually impaired people. Although recent years have witnessed an increase in accuracy in such models, this has also brought increasing complexity and challenges in interpretability and visualization. In this work, we focus on Transformer-based image captioning models and provide qualitative and quantitative tools to increase interpretability and assess the grounding and temporal alignment capabilities of such models. Firstly, we employ attribution methods to visualize what the model concentrates on in the input image, at each step of the generation. Further, we propose metrics to evaluate the temporal alignment between model predictions and attribution scores, which allows measuring the grounding capabilities of the model and spot hallucination flaws. Experiments are conducted on three different Transformer-based architectures, employing both traditional and Vision Transformer-based visual features.

Continual Learning (CL) investigates how to train Deep Networks on a stream of tasks without incurring forgetting. CL settings proposed in literature assume that every incoming example is paired with ground-truth annotations. However, this clashes with many real-world applications: gathering labeled data, which is in itself tedious and expensive, becomes infeasible when data flow as a stream. This work explores Continual Semi-Supervised Learning (CSSL): here, only a small fraction of labeled input examples are shown to the learner. We assess how current CL methods (e.g.: EWC, LwF, iCaRL, ER, GDumb, DER) perform in this novel and challenging scenario, where overfitting entangles forgetting. Subsequently, we design a novel CSSL method that exploits metric learning and consistency regularization to leverage unlabeled examples while learning. We show that our proposal exhibits higher resilience to diminishing supervision and, even more surprisingly, relying only on supervision suffices to outperform SOTA methods trained under full supervision.

Background and objectives: In the latest years, the prediction of gene expression levels has been crucial due to its potential applications in the clinics. In this context, Xpresso and others methods based on Convolutional Neural Networks and Transformers were firstly proposed to this aim. However, all these methods embed data with a standard one-hot encoding algorithm, resulting in impressively sparse matrices. In addition, post-transcriptional regulation processes, which are of uttermost importance in the gene expression process, are not considered in the model.Methods: This paper presents Transformer DeepLncLoc, a novel method to predict the abundance of the mRNA (i.e., gene expression levels) by processing gene promoter sequences, managing the problem as a regression task. The model exploits a transformer-based architecture, introducing the DeepLncLoc method to perform the data embedding. Since DeepLncloc is based on word2vec algorithm, it avoids the sparse matrices problem.Results: Post-transcriptional information related to mRNA stability and transcription factors is included in the model, leading to significantly improved performances compared to the state-of-the-art works. Transformer DeepLncLoc reached 0.76 of R-2 evaluation metric compared to 0.74 of Xpresso.Conclusion: The Multi-Headed Attention mechanisms which characterizes the transformer methodology is suitable for modeling the interactions between DNA's locations, overcoming the recurrent models. Finally, the integration of the transcription factors data in the pipeline leads to impressive gains in predictive power. (C) 2022 Elsevier B.V. All rights reserved.