José María Luna-Romera

Autonomous vehicles are equipped with complimentary sensors to perceive the environment accurately. Deep learning models have proven to be the most effective approach for computer vision problems. Therefore, in autonomous driving, it is essential to design reliable networks to fuse data from different sensors. In this work, we develop a novel data fusion architecture using camera and LiDAR data for object detection in autonomous driving. Given the sparsity of LiDAR data, developing multi-modal fusion models is a challenging task. Our proposal integrates an efficient LiDAR sparse-to-dense completion network into the pipeline of object detection models, achieving a more robust performance at different times of the day. The Waymo Open Dataset has been used for the experimental study, which is the most diverse detection benchmark in terms of weather and lighting conditions. The depth completion network is trained with the KITTI depth dataset, and transfer learning is used to obtain dense maps on Waymo. With the enhanced LiDAR data and the camera images, we explore early and middle fusion approaches using popular object detection models. The proposed data fusion network provides a significant improvement compared to single-modal detection at all times of the day, and outperforms previous approaches that upsample depth maps with classical image processing algorithms. Our multi-modal and multi-source approach achieves a 1.5, 7.5, and 2.1 mean AP increase at day, night, and dawn/dusk, respectively, using four different object detection meta-architectures.

The morphology of nanoparticles governs their properties for a range of important applications. Thus, the ability to statistically correlate this key particle performance parameter is paramount in achieving accurate control of nanoparticle properties. Among several effective techniques for morphological characterization of nanoparticles, transmission electron microscopy (TEM) can provide a direct, accurate characterization of the details of nanoparticle structures and morphology at atomic resolution. However, manually analyzing a large number of TEM images is laborious. In this work, we demonstrate an efficient, robust and highly automated unsupervised machine learning method for the metrology of nanoparticle systems based on TEM images. Our method not only can achieve statistically significant analysis, but it is also robust against variable image quality, imaging modalities, and particle dispersions. The ability to efficiently gain statistically significant particle metrology is critical in advancing precise particle synthesis and accurate property control.

Modern energy systems collect high volumes of data that can provide valuable information about energy consumption. Electric companies can now use historical data to make informed decisions on energy production by forecasting the expected demand. Many deep learning models have been proposed to deal with these types of time series forecasting problems. Deep neural networks, such as recurrent or convolutional, can automatically capture complex patterns in time series data and provide accurate predictions. In particular, Temporal Convolutional Networks (TCN) are a specialised architecture that has advantages over recurrent networks for forecasting tasks. TCNs are able to extract long-term patterns using dilated causal convolutions and residual blocks, and can also be more efficient in terms of computation time. In this work, we propose a TCN-based deep learning model to improve the predictive performance in energy demand forecasting. Two energy-related time series with data from Spain have been studied: the national electric demand and the power demand at charging stations for electric vehicles. An extensive experimental study has been conducted, involving more than 1900 models with different architectures and parametrisations. The TCN proposal outperforms the forecasting accuracy of Long Short-Term Memory (LSTM) recurrent networks, which are considered the state-of-the-art in the field.

Modern energy systems collect high volumes of data that can provide valuable information about energy consumption. Electric companies can now use historical data to make informed decisions on energy production by forecasting the expected demand. Many deep learning models have been proposed to deal with these type of time series forecasting problems. Deep neural networks, such as recurrent or convolutional, can automatically capture complex patterns in time series data and provide accurate predictions. In particular, Temporal Convolutional Networks (TCN) are a specialised architecture that has advantages over recurrent networks for forecasting tasks. TCNs are able to extract long-term patterns using dilated causal convolutions and residual blocks, and can also be more efficient in terms of computation time. In this work, we propose a TCN-based deep learning model to improve the predictive performance in energy demand forecasting. Two energy-related time series with data from Spain have been studied: the national electric demand, and the power demand at charging stations for electric vehicles. An extensive experimental study has been conducted, involving more than 1900 models with different architectures and parametrisations. The TCN proposal outperforms the forecasting accuracy of Long Short-Term Memory (LSTM) recurrent networks, which are considered the state-of-the-art in the field.

Unemployment in Spain is one of the biggest concerns of its inhabitants. Its unemployment rate is the second highest in the European Union, and in the second quarter of 2018 there is a 15.2% unemployment rate, some 3.4 million unemployed. Construction is one of the activity sectors that have suffered the most from the economic crisis. In addition, the economic crisis affected in different ways to the labour market in terms of occupation level or location. The aim of this paper is to discover how the labour market is organised taking into account the jobs that workers get during two periods: 2011-2013, which corresponds to the economic crisis period, and 2014-2016, which was a period of economic recovery. The data used are official records of the Spanish administration corresponding to 1.9 and 2.4 million job placements, respectively. The labour market was analysed by applying unsupervised machine learning techniques to obtain a clear and structured information on the employment generation process and the underlying labour mobility. We have applied two clustering methods with two different technologies, and the results indicate that there were some movements in the Spanish labour market which have changed the physiognomy of some of the jobs. The analysis reveals the changes in the labour market: the crisis forces greater geographical mobility and favours the subsequent emergence of new job sources. Nevertheless, there still exist some clusters that remain stable despite the crisis. We may conclude that we have achieved a characterisation of some important groups of workers in Spain. The methodology used, being supported by Big Data techniques, would serve to analyse any alternative job market.