A5059337172- Parallel Computing and Optimization Techniques
- Radiation Effects in Electronics
- VLSI and Analog Circuit Testing
- Low-power high-performance VLSI design
- Speech Recognition and Synthesis
- Speech and Audio Processing
- Music and Audio Processing
- Reinforcement Learning in Robotics
- Distributed systems and fault tolerance
- Embedded Systems Design Techniques
- Ferroelectric and Negative Capacitance Devices
- Advanced Bandit Algorithms Research
- Cryptography and Residue Arithmetic
- Building Energy and Comfort Optimization
- Advanced Memory and Neural Computing
- Energy Load and Power Forecasting
- Solar Radiation and Photovoltaics
- Numerical Methods and Algorithms
- Coding theory and cryptography
- Physical Unclonable Functions (PUFs) and Hardware Security
- Smart Grid Energy Management
- Distributed and Parallel Computing Systems
- Algorithms and Data Compression
- IoT and Edge/Fog Computing
Sapienza University of Rome
2021-2024
The L3DAS21 Challenge is aimed at encouraging and fostering collaborative research on machine learning for 3D audio signal processing, with particular focus speech enhancement (SE) sound localization detection (SELD). Alongside the challenge, we release dataset, a 65 hours corpus, accompanied Python API that facilitates data usage results submission stage. Usually, approaches to tasks are based single-perspective Ambisonics recordings or arrays of single-capsule microphones. We propose,...
High-performance embedded systems with powerful processors, specialized hardware accelerators, and advanced software techniques are all key technologies driving the growth of IoT. By combining techniques, it is possible to increase overall reliability safety these by designing architectures that can continue function correctly in event a failure or malfunction. In this work, we fully investigate integration configurable vector acceleration unit fault-tolerant RISC-V Klessydra-fT03 soft core,...
Integer division is key for various applications and often represents the performance bottleneck due to its inherent mathematical properties that limit parallelization.This paper presents a new datadependent variable latency algorithm, derived from classic non-performing restoring method.The proposed technique exploits relationship between number of leading zeros in divisor partial remainder dynamically detect skip those iterations result simple left shift.While similar principle has been...
Reconfigurable processors are hardware architectures that allow for the dynamic configuration of processing resources to optimize performance and power consumption, using partial reconfiguration modify a portion design or update it without affecting entire system. In this work, we present an automatic technique leverages machine learning (ML) algorithms automatically select optimal general-purpose accelerator according workload reconFigure architecture at run-time. The problem is formulated...
Integer division is key for various applications and often represents the performance bottleneck due to its inherent mathematical properties that limit parallelization. This work proposes four 32bit data-dependent-latency schemes, derived from classic non-performing restoring algorithm. The proposed technique exploits relationship between number of leading zeros in divisor partial remainder dynamically detect skip those iterations result a simple left shift. While similar principle has been...
Hyperdimensional Computing (HDC) is a bioinspired learning paradigm, that models neural pattern activities using high-dimensional distributed representations.HDC leverages parallel and simple vector arithmetic operations to combine compare different concepts, enabling cognitive reasoning tasks.The computational efficiency parallelism of this approach make it particularly suited for hardware implementations, especially as lightweight, energy-efficient solution performing tasks on...
The L3DAS21 Challenge is aimed at encouraging and fostering collaborative research on machine learning for 3D audio signal processing, with particular focus speech enhancement (SE) sound localization detection (SELD). Alongside the challenge, we release dataset, a 65 hours corpus, accompanied Python API that facilitates data usage results submission stage. Usually, approaches to tasks are based single-perspective Ambisonics recordings or arrays of single-capsule microphones. We propose,...