Temporal dynamics of benthic assemblages along a gradient of ocean acidification at a CO2 vent’s system of the Ischia Island

Global warming and ocean acidification, due to rising atmospheric levels of CO2, represent an actual threat to terrestrial and marine environments. Since Industrial Revolution, in less of 250 years, pH of surface seawater decreased on average of 0.1 unit, and is expected to further decreases of approximately 0.3-0.4 units by the end of this century. Naturally acidified marine areas, such as CO2 vent systems at the Ischia Island, allow to study acclimatation and adaptation of individual species as well as the structure of communities, and ecosystems to OA. The main aim of this thesis was to study how hard bottom sublittoral benthic assemblages changed trough time along a pH gradient. For this purpose, the temporal dynamics of mature assemblages established on artificial substrates (volcanic tiles) over a 3 year- period were analysed. Our results revealed how composition and dynamics of the community were altered and highly simplified at different level of seawater acidification. In fact, extreme low values of pH (approximately 6.9), affected strongly the assemblages, reducing diversity both in terms of taxa and functional groups, respect to lower acidification levels (mean pH 7.8) and ambient conditions (8.1 unit). Temporal variation was observed in terms of species composition but not in functional groups. Variability was related to species belonging to the same functional group, suggesting the occurrence of functional redundancy. Therefore, the analysis of functional groups kept information on the structure, but lost information on species diversity and dynamics. Decreasing in ocean pH is only one of many future global changes that will occur at the end of this century (increase of ocean temperature, sea level rise, eutrophication etc.). The interaction between these factors and OA could exacerbate the community and ecosystem effects showed by this thesis.

Continual learning on nano drones

Miniaturized flying robotic platforms, called nano-drones, have the potential to revolutionize the autonomous robots industry sector thanks to their very small form factor. The nano-drones’ limited payload only allows for a sub-100mW microcontroller unit for the on-board computations. Therefore, traditional computer vision and control algorithms are too computationally expensive to be executed on board these palm-sized robots, and we are forced to rely on artificial intelligence to trade off accuracy in favor of lightweight pipelines for autonomous tasks. However, relying on deep learning exposes us to the problem of generalization since the deployment scenario of a convolutional neural network (CNN) is often composed by different visual cues and different features from those learned during training, leading to poor inference performances. Our objective is to develop and deploy and adaptation algorithm, based on the concept of latent replays, that would allow us to fine-tune a CNN to work in new and diverse deployment scenarios. To do so we start from an existing model for visual human pose estimation, called PULPFrontnet, which is used to identify the pose of a human subject in space through its 4 output variables, and we present the design of our novel adaptation algorithm, which features automatic data gathering and labeling and on-device deployment. We therefore showcase the ability of our algorithm to adapt PULP-Frontnet to new deployment scenarios, improving the R2 scores of the four network outputs, with respect to an unknown environment, from approximately [−0.2, 0.4, 0.0,−0.7] to [0.25, 0.45, 0.2, 0.1]. Finally we demonstrate how it is possible to fine-tune our neural network in real time (i.e., under 76 seconds), using the target parallel ultra-low power GAP 8 System-on-Chip on board the nano-drone, and we show how all adaptation operations can take place using less than 2mWh of energy, a small fraction of the available battery power.