Dr. Sotirios Chatzis has recently presented a paper in ICML 2013, the most significant annual international conference on machine learning, gathering scientists from the top machine learning groups from the US and worldwide.
The presented paper entitled "Infinite Markov-Switching Maximum Entropy Discrimination Machines," introduces a nonparametric Bayesian Maximum Entropy Discrimination classifier for classification of long sequences, which remain at the same state for quite some time (state-persistence), with applications in video understanding. The proposed method has outperformed state-of-the-art competitors in several real-world scenarios, mostly dealing with semantic video analysis.
Presentation in the leading Machine Learning conference
Dr. Sotirios Chatzis has recently presented a paper in ICML 2013, the most significant annual international conference on machine learning, gathering scientists from the top machine learning groups from the US and worldwide.
The presented paper entitled "Infinite Markov-Switching Maximum Entropy Discrimination Machines," introduces a nonparametric Bayesian Maximum Entropy Discrimination classifier for classification of long sequences, which remain at the same state for quite some time (state-persistence), with applications in video understanding. The proposed method has outperformed state-of-the-art competitors in several real-world scenarios, mostly dealing with semantic video analysis.
