• [Jun 28, 2024] One paper Towards Unbiased Calibration using Meta-Regularization was accepted at Trans. on Machine Learning Research
• [Feb 11, 2024] I am SPC member for IJCAI 2024
• [Mar 1, 2023] I am SPC member for IJCAI 2023
• [Dec 4, 2022] One paper State-Regularized Recurrent Neural Networks to Extract Automata and Explain Predictions was accepted at IEEE Trans. on PAMI
• [Nov 4, 2022] Two papers Calibrating Imbalanced Classifiers with Focal Loss: An Empirical Study and Deploying a Retrieval based Response Model for Task Oriented Dialogues were accepted at EMNLP 2022 Industry Track

Welcome! My name is Cheng Wang. I am currently a Senior Applied Scientist at Amazon. Prior to that, I was a Research Scientist at Machine Learning Group, NEC Laboratories Europe. I received my PhD (Dr.rer.nat) from Hasso-Plattner-Institute, University of Potsdam in 2017. There, I was supervised by Prof. Dr. Christoph Meinel, my research topic was "Multimodal Deep Learning". My research focuses are (1) machine (deep) learning methods, e.g, learning algorithms, sequence-to-sequence learning, representation learning, neural architectures, uncertainty estimation/quantification, model calibration and interpretability; (2) their applications to real-world tasks, e.g., information retrieval, recommender systems, natural language processing, computer vision and multimodal learning. I am particularly interested in sequential models (Transformer/GPT/RNNs/LSTMs etc) and state machines related to them.