PITCHD 2023 EDITION: THE 4th PITCHD
Machine learning on ultra-low power devices:
approaches for accurate and energy efficient inference.
Mr. Francesco Daghero, PhD candidate DAUIN PoliTO.
Machine Learning (ML) is now central to IoT applications,
offering low latency and privacy benefits through on-device
processing. However, constraints on computational and memory
resources in IoT devices hinder direct ML model deployment. This
presentation explains methods like mixed-precision quantization,
binarization, and adaptive inference, optimizing models for
embedded systems, reducing energy usage, and maintaining
accuracy.
Quantum Computing applications tackling nowadays’ problems:
optimising urban traffic mobility case study
Mr Andrea Marchesin, PhD Candidate, DET, PoliTO
Quantum Computing (QC) offers a revolutionary
approach to solving complex problems beyond traditional
processing systems. A novel research work suggests that
applying QC can improve traffic flow and reduce
congestion in modern cities, bringing benefits to the
environment and citizens' quality of life.
Constrained optimization via controlled multipliers
Mr. Simone Pirrera, PhD Candidate, DAUIN, PoliTO
We propose a control theory-based approach to develop novel
optimization algorithms for constrained non-convex problems.
Specifically, we define a continuous-time system that, under a
properly defined control action, converges to a feasible stationary
point. We design the control action using PI and feedback
linearization, and we conduct theoretical analysis to show method's
convergence. Finally, we demonstrate the method's practical
effectiveness through numerical examples.
Date and Time
Location
Hosts
Registration
-
Add Event to Calendar
- Politecnico di Torino
- Corso Duca degli Abruzzi, 24
- Torino, Piemonte
- Italy
- Room Number: Maxwell Room, DET