Adam Popowicz
Papers
1
Total Citations
14
H-Index
1
About
Adam Popowicz is a researcher focused on the reliability and safety of deep neural networks, with a particular emphasis on out-of-distribution (OOD) detection. His most-cited work, "Detection of Out-of-Distribution Samples Using Binary Neuron Activation Patterns" (2023, 14 citations), addresses a critical limitation of DNN classifiers: their inability to reliably identify novel or unseen inputs. Popowicz introduces a novel method that leverages binary neuron activation patterns to distinguish OOD samples from in-distribution data, offering a computationally efficient and interpretable solution. This contribution is especially vital for safety-critical applications, such as autonomous systems and medical diagnostics, where misclassifying an unknown input can have severe consequences. By tackling the challenge of OOD detection, Popowicz helps bridge the gap between high-performing models and their real-world deployment. His work underscores a commitment to building more robust and trustworthy AI systems, making him a notable voice in the ongoing effort to enhance neural network reliability.
Research Focus
Key Achievements
Top Papers
- 1