Netherlands
algorithm engineer
Modeling long-term dependencies in high-resolution data
Mitigating Spurious Correlations in Weakly Supervised Semantic Segmentation via Cross-architecture Consistency Regularization.
Machine Learning for the Quantified Self
Detecting and analyzing the behavior of Bio-inspired Robot，maximizing robotic efficiency in locomotion within real world contexts. Evolution algorithms； Pose estimation； Video embedding；
Li-ion battery defect detection Analyzing and testing different detection models and assessing the quality of these detection models. Focus on using newest algorithms with a custom dataset to detect the surface flaws of the battery. Enhanced the model performance through the incorporation of advanced techniques such as attention mechanism, new feature pyramid, the optimized model was then successfully deployed in the real environment.