Structure learning of graphical models for task-oriented robot grasping

In the collective imaginaries a robot is a human like machine as any androids in science fiction. However the type of robots that you will encounter most frequently are machinery that do work that is too dangerous, boring or onerous. Most of the robots in the world are of this type. They can be found in auto, medical, manufacturing and space industries. Therefore a robot is a system that contains sensors, control systems, manipulators, power supplies and software all working together to perform a task. The development and use of such a system is an active area of research and one of the main problems is the development of interaction skills with the surrounding environment, which include the ability to grasp objects. To perform this task the robot needs to sense the environment and acquire the object informations, physical attributes that may influence a grasp. Humans can solve this grasping problem easily due to their past experiences, that is why many researchers are approaching it from a machine learning perspective finding grasp of an object using information of already known objects. But humans can select the best grasp amongst a vast repertoire not only considering the physical attributes of the object to grasp but even to obtain a certain effect. This is why in our case the study in the area of robot manipulation is focused on grasping and integrating symbolic tasks with data gained through sensors. The learning model is based on Bayesian Network to encode the statistical dependencies between the data collected by the sensors and the symbolic task. This data representation has several advantages. It allows to take into account the uncertainty of the real world, allowing to deal with sensor noise, encodes notion of causality and provides an unified network for learning. Since the network is actually implemented and based on the human expert knowledge, it is very interesting to implement an automated method to learn the structure as in the future more tasks and object features can be introduced and a complex network design based only on human expert knowledge can become unreliable. Since structure learning algorithms presents some weaknesses, the goal of this thesis is to analyze real data used in the network modeled by the human expert, implement a feasible structure learning approach and compare the results with the network designed by the expert in order to possibly enhance it.

Comparative study of the different models for the calculation of BLEVEs overpressure effects

The BLEVE, acronym for Boiling Liquid Expanding Vapour Explosion, is one of the most dangerous accidents that can occur in pressure vessels. It can be defined as an explosion resulting from the failure of a vessel containing a pressure liquefied gas stored at a temperature significantly above its boiling point at atmospheric pressure. This phenomenon frequently appears when a vessel is engulfed by a fire: the heat causes the internal pressure to raise and the mechanical proprieties of the wall to decrease, with the consequent rupture of the tank and the instantaneous release of its whole content. After the breakage, the vapour outflows and expands and the liquid phase starts boiling due to the pressure drop. The formation and propagation of a distructive schock wave may occur, together with the ejection of fragments, the generation of a fireball if the stored fluid is flammable and immediately ignited or the atmospheric dispersion of a toxic cloud if the fluid contained inside the vessel is toxic. Despite the presence of many studies on the BLEVE mechanism, the exact causes and conditions of its occurrence are still elusive. In order to better understand this phenomenon, in the present study first of all the concept and definition of BLEVE are investigated. A historical analysis of the major events that have occurred over the past 60 years is described. A research of the principal causes of this event, including the analysis of the substances most frequently involved, is presented too. Afterwards a description of the main effects of BLEVEs is reported, focusing especially on the overpressure. Though the major aim of the present thesis is to contribute, with a comparative analysis, to the validation of the main models present in the literature for the calculation and prediction of the overpressure caused by BLEVEs. In line with this purpose, after a short overview of the available approaches, their ability to reproduce the trend of the overpressure is investigated. The overpressure calculated with the different models is compared with values deriving from events happened in the past and ad-hoc experiments, focusing the attention especially on medium and large scale phenomena. The ability of the models to consider different filling levels of the reservoir and different substances is analyzed too. The results of these calculations are extensively discussed. Finally some conclusive remarks are reported.

Intermediate linguistic task fine-tuning on Multi-lingual models

State-of-the-art NLP systems are generally based on the assumption that the underlying models are provided with vast datasets to train on. However, especially when working in multi-lingual contexts, datasets are often scarce, thus more research should be carried out in this field. This thesis investigates the benefits of introducing an additional training step when fine-tuning NLP models, named Intermediate Training, which could be exploited to augment the data used for the training phase. The Intermediate Training step is applied by training models on NLP tasks that are not strictly related to the target task, aiming to verify if the models are able to leverage the learned knowledge of such tasks. Furthermore, in order to better analyze the synergies between different categories of NLP tasks, experimentations have been extended also to Multi-Task Training, in which the model is trained on multiple tasks at the same time.

Bio-QA-GNN: Reasoning with Language Models and Knowledge Graphs for Interpretable Biomedical Question Answering

Nowadays the idea of injecting world or domain-specific structured knowledge into pre-trained language models (PLMs) is becoming an increasingly popular approach for solving problems such as biases, hallucinations, huge architectural sizes, and explainability lack—critical for real-world natural language processing applications in sensitive fields like bioinformatics. One recent work that has garnered much attention in Neuro-symbolic AI is QA-GNN, an end-to-end model for multiple-choice open-domain question answering (MCOQA) tasks via interpretable text-graph reasoning. Unlike previous publications, QA-GNN mutually informs PLMs and graph neural networks (GNNs) on top of relevant facts retrieved from knowledge graphs (KGs). However, taking a more holistic view, existing PLM+KG contributions mainly consider commonsense benchmarks and ignore or shallowly analyze performances on biomedical datasets. This thesis start from a propose of a deep investigation of QA-GNN for biomedicine, comparing existing or brand-new PLMs, KGs, edge-aware GNNs, preprocessing techniques, and initialization strategies. By combining the insights emerged in DISI's research, we introduce Bio-QA-GNN that include a KG. Working with this part has led to an improvement in state-of-the-art of MCOQA model on biomedical/clinical text, largely outperforming the original one (+3.63\% accuracy on MedQA). Our findings also contribute to a better understanding of the explanation degree allowed by joint text-graph reasoning architectures and their effectiveness on different medical subjects and reasoning types. Codes, models, datasets, and demos to reproduce the results are freely available at: \url{https://github.com/disi-unibo-nlp/bio-qagnn}.