Elena Mocanu
YOU?
Author Swipe
View article: Addressing the Collaboration Dilemma in Low-Data Federated Learning via Transient Sparsity
Addressing the Collaboration Dilemma in Low-Data Federated Learning via Transient Sparsity Open
Federated learning (FL) enables collaborative model training across decentralized clients while preserving data privacy, leveraging aggregated updates to build robust global models. However, this training paradigm faces significant challen…
View article: Dynamic Sparse Training versus Dense Training: The Unexpected Winner in Image Corruption Robustness
Dynamic Sparse Training versus Dense Training: The Unexpected Winner in Image Corruption Robustness Open
It is generally perceived that Dynamic Sparse Training opens the door to a new era of scalability and efficiency for artificial neural networks at, perhaps, some costs in accuracy performance for the classification task. At the same time, …
View article: Digital Twin-Empowered Autonomous Driving for E-mobility: Concept, framework, and modeling
Digital Twin-Empowered Autonomous Driving for E-mobility: Concept, framework, and modeling Open
As a disruptive technology in the power and transport sectors, electric mobility (e-mobility) is receiving increasing attention. E-mobility encompasses the electrification of transportation by means of diversified electric vehicles (EVs) i…
View article: E2ENet: Dynamic Sparse Feature Fusion for Accurate and Efficient 3D Medical Image Segmentation
E2ENet: Dynamic Sparse Feature Fusion for Accurate and Efficient 3D Medical Image Segmentation Open
Deep neural networks have evolved as the leading approach in 3D medical image segmentation due to their outstanding performance. However, the ever-increasing model size and computation cost of deep neural networks have become the primary b…
View article: MICROSTRUCTURAL ANALYSIS OF ALLOWANCED CEMENTITIOUS MORTAR WITH DIFERENTS NANOPARTICLES
MICROSTRUCTURAL ANALYSIS OF ALLOWANCED CEMENTITIOUS MORTAR WITH DIFERENTS NANOPARTICLES Open
Nanomaterials are materials with new properties that interact through complex processes at the nanometric scale that include reactions in the quantum field and which still represent an unknown for many researchers. Improving the performanc…
View article: Automatic Noise Filtering with Dynamic Sparse Training in Deep Reinforcement Learning
Automatic Noise Filtering with Dynamic Sparse Training in Deep Reinforcement Learning Open
Tomorrow's robots will need to distinguish useful information from noise when performing different tasks. A household robot for instance may continuously receive a plethora of information about the home, but needs to focus on just a small …
View article: Dynamic Sparse Network for Time Series Classification: Learning What to "see''
Dynamic Sparse Network for Time Series Classification: Learning What to "see'' Open
The receptive field (RF), which determines the region of time series to be ``seen'' and used, is critical to improve the performance for time series classification (TSC). However, the variation of signal scales across and within time serie…
View article: Dynamic Sparse Training for Deep Reinforcement Learning
Dynamic Sparse Training for Deep Reinforcement Learning Open
Deep reinforcement learning (DRL) agents are trained through trial-and-error interactions with the environment. This leads to a long training time for dense neural networks to achieve good performance. Hence, prohibitive computation and me…
View article: Quick and robust feature selection: the strength of energy-efficient sparse training for autoencoders
Quick and robust feature selection: the strength of energy-efficient sparse training for autoencoders Open
Major complications arise from the recent increase in the amount of high-dimensional data, including high computational costs and memory requirements. Feature selection, which identifies the most relevant and informative attributes of a da…
View article: Deep Ensembling with No Overhead for either Training or Testing: The All-Round Blessings of Dynamic Sparsity
Deep Ensembling with No Overhead for either Training or Testing: The All-Round Blessings of Dynamic Sparsity Open
The success of deep ensembles on improving predictive performance, uncertainty estimation, and out-of-distribution robustness has been extensively studied in the machine learning literature. Albeit the promising results, naively training m…
View article: Dynamic Sparse Training for Deep Reinforcement Learning
Dynamic Sparse Training for Deep Reinforcement Learning Open
Deep reinforcement learning (DRL) agents are trained through trial-and-error interactions with the environment. This leads to a long training time for dense neural networks to achieve good performance. Hence, prohibitive computation and me…
View article: FreeTickets: Accurate, Robust and Efficient Deep Ensemble by Training with Dynamic Sparsity.
FreeTickets: Accurate, Robust and Efficient Deep Ensemble by Training with Dynamic Sparsity. Open
Recent works on sparse neural networks have demonstrated that it is possible to train a sparse network in isolation to match the performance of the corresponding dense networks with a fraction of parameters. However, the identification of …
View article: Sparse Training Theory for Scalable and Efficient Agents
Sparse Training Theory for Scalable and Efficient Agents Open
A fundamental task for artificial intelligence is learning. Deep Neural Networks have proven to cope perfectly with all learning paradigms, i.e. supervised, unsupervised, and reinforcement learning. Nevertheless, traditional deep learning …
View article: Quick and Robust Feature Selection: the Strength of Energy-efficient\n Sparse Training for Autoencoders
Quick and Robust Feature Selection: the Strength of Energy-efficient\n Sparse Training for Autoencoders Open
Major complications arise from the recent increase in the amount of\nhigh-dimensional data, including high computational costs and memory\nrequirements. Feature selection, which identifies the most relevant and\ninformative attributes of a…
View article: Effectiveness of neural language models for word prediction of textual mammography reports
Effectiveness of neural language models for word prediction of textual mammography reports Open
Radiologists are required to write free paper text reports for breast screenings in order to assign cancer diagnoses in a later step. The current procedure requires considerable time and needs efficiency. In this paper, to streamline the w…
View article: On-Line Building Energy Optimization Using Deep Reinforcement Learning
On-Line Building Energy Optimization Using Deep Reinforcement Learning Open
Unprecedented high volumes of data are becoming available with the growth of the advanced metering infrastructure. These are expected to benefit planning and operation of the future power systems and to help customers transition from a pas…
View article: One-Shot Learning using Mixture of Variational Autoencoders: a Generalization Learning approach
One-Shot Learning using Mixture of Variational Autoencoders: a Generalization Learning approach Open
Deep learning, even if it is very successful nowadays, traditionally needs very large amounts of labeled data to perform excellent on the classification task. In an attempt to solve this problem, the one-shot learning paradigm, which makes…
View article: Enabling Cooperative Behavior for Building Demand Response Based on Extended Joint Action Learning
Enabling Cooperative Behavior for Building Demand Response Based on Extended Joint Action Learning Open
This paper explores the use of distributed intelligence to assist the integration of the demand as a flexible resource, to mitigate the emerging uncertainty in the power system, while fulfilling the customer's local needs, i.e., comfort ma…
View article: On-line Building Energy Optimization using Deep Reinforcement Learning
On-line Building Energy Optimization using Deep Reinforcement Learning Open
Unprecedented high volumes of data are becoming available with the growth of the advanced metering infrastructure. These are expected to benefit planning and operation of the future power system, and to help the customers transition from a…
View article: Deep learning methods for on-line flexibility prediction and optimal resource allocation in smart buildings
Deep learning methods for on-line flexibility prediction and optimal resource allocation in smart buildings Open
Unprecedented high volume of data is available with the upward growth of the advanced metering infrastructure. Because the built environment is the largest user of electricity, a deeper look at building energy consumption holds promise for…
View article: Big IoT data mining for real-time energy disaggregation in buildings (extended abstract)
Big IoT data mining for real-time energy disaggregation in buildings (extended abstract) Open
In the smart grid context, the identification and prediction of building energy flexibility is a challenging open question. In this paper, we propose a hybrid approach to address this problem. It combines sparse smart meters with deep lear…
View article: Medium Voltage DC Power Systems on Ships: An Offline Parameter Estimation for Tuning the Controllers’ Linearizing Function
Medium Voltage DC Power Systems on Ships: An Offline Parameter Estimation for Tuning the Controllers’ Linearizing Function Open
Future shipboard power systems using Medium Voltage Direct (MVDC) technology will be based on a widespread use of power converters for interfacing generating systems and loads with the main DC bus. Such a heavy exploitation makes the volta…
View article: A topological insight into restricted Boltzmann machines (extented abstract)
A topological insight into restricted Boltzmann machines (extented abstract) Open
Restricted Boltzmann Machines (RBMs) and models derived from them have been successfully used as basic building blocks in deep neural networks for automatic features extraction, unsupervised weights initialization, but also as standalone m…
View article: Big IoT data mining for real-time energy disaggregation in buildings
Big IoT data mining for real-time energy disaggregation in buildings Open
In the smart grid context, the identification and prediction of building energy flexibility is a challenging open question, thus paving the way for new optimized behaviors from the demand side. At the same time, the latest smart meters dev…
View article: A topological insight into restricted Boltzmann machines
A topological insight into restricted Boltzmann machines Open
Restricted Boltzmann Machines (RBMs) and models derived from them have been successfully used as basic building blocks in deep artificial neural networks for automatic features extraction, unsupervised weights initialization, but also as d…
View article: Energy disaggregation for real-time building flexibility detection
Energy disaggregation for real-time building flexibility detection Open
Energy is a limited resource which has to be managed wisely, taking into account both supply-demand matching and capacity constraints in the distribution grid. One aspect of the smart energy management at the building level is given by the…
View article: Big) data analytics in smart grids
Big) data analytics in smart grids Open
Prediction of building energy consumption is a fundamental problem in the smart grid context. Unprecedented high volumes of data and information are available with the upward growth of the smart metering infrastructure. Therefore, we devel…