Building AI That Knows When Not to Trust Itself
Why uncertainty in deep learning is as important as accuracy.
Posts by Category
blog
System-Level Power Profiling for AI Workloads at ESAIL
Vendor GPU telemetry captures fast on-chip power dynamics, while external board-level measurements reveal true aggregated energy cost, making the latter esse...
Walking Noise
On Layer-Specific Robustness of Neural Architectures against Noisy Computations and Associated Characteristic Learning Dynamics
How much “Brain Damage” can an LLM Tolerate?
An exploration of the effect of RRAM activation write noise on the GEMMA LLM
ZITI Thesis Fair April 2024
A short overview of the ZITI Thesis Fair of Spring Semester 2024.
research
Building AI That Knows When Not to Trust Itself
Why uncertainty in deep learning is as important as accuracy.
System-Level Power Profiling for AI Workloads at ESAIL
Vendor GPU telemetry captures fast on-chip power dynamics, while external board-level measurements reveal true aggregated energy cost, making the latter esse...
Walking Noise
On Layer-Specific Robustness of Neural Architectures against Noisy Computations and Associated Characteristic Learning Dynamics
How much “Brain Damage” can an LLM Tolerate?
An exploration of the effect of RRAM activation write noise on the GEMMA LLM
event
ZITI Thesis Fair April 2024
A short overview of the ZITI Thesis Fair of Spring Semester 2024.
facilities
System-Level Power Profiling for AI Workloads at ESAIL
Vendor GPU telemetry captures fast on-chip power dynamics, while external board-level measurements reveal true aggregated energy cost, making the latter esse...