1,721,009 research outputs found
An embedded end-to-end voice assistant
No full textVoice assistants are spreading in various environments, such as houses and cars, bringing the possibility of controlling heterogeneous Internet of Things devices with simple voice commands. However, massive use of the cloud connection for speech processing requires an efficient and robust Internet connection and raises concerns in terms of privacy. Therefore, we propose an end-to-end solution able to work totally offline, based on a system architecture combining different Deep Learning models to implement all the steps of the speech elaboration process. Being interested in targeting the Italian language, we exploited the transfer learning paradigm, which allows leveraging models trained in English on large datasets and fine-tuning them to the target language on a smaller dataset. The proposed system architecture is configurable and easily extensible to other languages. Experimental results in an automotive application use case show that our solution outperforms the other embedded models and achieves performance comparable to state-of-the-art cloud-connected solutions for Automatic Speech Recognition. Moreover, overall latency is significantly reduced by eliminating the need to connect to the cloud
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Building a Pipeline for Efficient Production of Synthetic Datasets for Improving RL in Automated Driving
No full textOnline deep reinforcement learning training poses challenges due to its length and instability, despite the development of learning algorithms targeted to overcome these issues. Offline learning has emerged as a potential solution, but it reintroduces the issue of dataset production, which is resource-consuming and challenging even in simulation environments. This paper investigates efficient dataset creation for offline learning in the context of automated driving. Our proposed solution is a pipeline based on the CARLA simulator, which offers a wide variety in terms of car models, weather conditions, and environments. The pipeline aims to produce high-quality datasets for pre-training, training, and fine-tuning models, targeting improved training speed and reduced divergence. By leveraging CARLA's level of realism, we address the resource-intensive nature of dataset production, providing researchers and car manufacturers with a valuable tool for advancing the development of robust automated driving systems
Appropriate Similarity Measures for Author Cocitation Analysis
Full text linkWe provide a number of new insights into the methodological discussion about author cocitation analysis. We first argue that the use of the Pearson correlation for measuring the similarity between authors’ cocitation profiles is not very satisfactory. We then discuss what kind of similarity measures may be used as an alternative to the Pearson correlation. We consider three similarity measures in particular. One is the well-known cosine. The other two similarity measures have not been used before in the bibliometric literature. Finally, we show by means of an example that our findings have a high practical relevance.information science;Pearson correlation;cosine;similarity measure;author cocitation analysis
Benchmarking Microcontrollers with Ultra-Low Resolution Images Classification
No full textThe advancement of edge devices equipped with specialized hardware accelerators, data caches, and microcontroller units (MCUs) has brought the deployment and execution of Deep Neural Network (DNN) models closer to users and real-world sensor systems. This paper explores the potential of various specialized MCUs across four real-world applications (waste classification, presence detection, miniature robot detection, and sign language interpretation). We evaluate three well-known MCU (STM32H7, Arduino, and MAX78000) comparing their inference time and power/energy consumption on four ultra-low resolution image-classification datasets with varying input and task complexities. Our findings indicate that all MCUs deliver excellent performance in terms of inference time and energy consumption for real-time applications, with the MAX78000 outperforming the others in all the metrics considered, thanks to the use of a DNN accelerator
Leveraging Neural Architecture Search for Structural Health Monitoring on Resource-Constrained Devices
No full textIn recent decades signal processing incorporated the capabilities offered by Deep Learning (DL) models, especially for complex tasks. DL models demand significant memory, power, and computational resources, posing challenges for Microcontroller Units (MCUs) with limited capacities. The possibility to run models directly on the edge device is key in connectivity-limited scenarios such as Structural Health Monitoring (SHM). For those scenarios, it is necessary to use Tiny Machine Learning techniques to reduces computational requirements. This study focuses on the impact of the extended version of the state-of-the-art Neural Architecture Search (NAS) tool, μNAS, for SHM applications, targeting four commonly used MCUs. Our assessment is based on the Z24 Bridge benchmark dataset, a common dataset for SHM we employed to train and evaluate models. We then discuss if the models found fit the constraints of the MCUs and the possible tradeoffs between error rate and model computational requirements. We also offer a comparison with the Raspberry Pi 4 Model B to highlight μNAS’s capability in achieving high accuracy with higher computing capabilities. The obtained results are promising, as the found models satisfy the given constraints both in term of accuracy and memory footprint
Exploring Decision Transformer for Highway Automated Driving
No full textThe evolution of Automated Driving Functions (ADFs) is contingent upon the effective implementation of Decision-Making (DM), context perception, and predictive vehicle control. Conventional Deep Reinforcement Learning (DRL) methodologies frequently prove inadequate in dynamic settings, largely due to their inherent limitations in addressing real-time DM and assigning long-term credit. DRL via sequence modeling represents a promising avenue for addressing these challenges by combining the strengths of Attention-based architectures, such Transformer, and DRL. The integration of self-attention mechanisms with offline DRL enables long-term credit assignment, fine-tuning and prevent continuous interaction with the environment, mitigating risks related to real-world simulations and trial-and-error approaches. This paper examines the potential of Decision Transformer (DT) within the AD domain. A DT model was implemented and trained within the highway-env simulation environment. To do so, an offline RL dataset was constructed using a pre-trained Deep Q-Network (DQN) agent. The model was evaluated by comparing its performance against that of the pre-trained DQN and a random agent. Results demonstrated that the DT model exhibited superior DM capabilities, with higher average returns and longer episode durations than DQN. These findings highlight the potential of Transformer-based DRL in AD
Quantization of MobileNetV2 for Resource-Constrained Microcontrollers
No full textThe proliferation of IoT and edge computing devices demands for efficient deployment strategies due to their limited computational capabilities and energy constraints. This paper investigates the application of deep learning models in such resource-constrained environments, focusing on the quantization of the MobileNetV2 architecture. We evaluate three primary quantization techniques: dynamic quantization, static quantization, and quantization-aware training. Using a subset of the Vehicle Make and Model Recognition dataset, specifically the Most Stolen Vehicles in the US in 2017, we compare our results with a previous study, highlighting the advancements achieved through MobileNetV2 and its quantization process. Deployment on two STM32 boards, L4 and H7 series, demonstrates the effectiveness of the quantized MobileNetV2 model in achieving efficient, low-power, and low-latency execution on low-power MCUs
- …
