Vector Institute Researchers Present 24 Papers at Virtual ICML 2020
The Vector Institute for Artificial Intelligence announces that its researchers are preparing for the 2020 International Conference on Machine Learning (ICML), which will be held virtually from July 13 to July 18. As one of the world's premier machine learning conferences, ICML 2020 accepted 1,088 papers from 4,990 submissions, resulting in a 21.8% acceptance rate. Vector Institute faculty and affiliates contributed significantly, with 19 papers from faculty members and five from affiliates accepted, comprising 2.21% of the total accepted works. The article highlights several key research areas presented by Vector researchers, including Angular Visual Hardness for improving model calibration, causal modeling for fairness in dynamical systems, convex representation learning for generalized invariance, and efficient training methods for energy-based models. These contributions underscore the institute's active role in advancing machine learning methodologies, particularly in areas of model fairness, robustness, and efficiency. The shift to a virtual format reflects broader adaptations within the academic community during this period.
Wire timeline
Vector Institute Researchers Present 24 Papers at Virtual ICML 2020
The Vector Institute for Artificial Intelligence announces that its researchers are preparing for the 2020 International Conference on Machine Learning (ICML), which will be held virtually from July 13 to July 18. As one of the world's premier machine learning conferences, ICML 2020 accepted 1,088 papers from 4,990 submissions, resulting in a 21.8% acceptance rate. Vector Institute faculty and affiliates contributed significantly, with 19 papers from faculty members and five from affiliates accepted, comprising 2.21% of the total accepted works. The article highlights several key research areas presented by Vector researchers, including Angular Visual Hardness for improving model calibration, causal modeling for fairness in dynamical systems, convex representation learning for generalized invariance, and efficient training methods for energy-based models. These contributions underscore the institute's active role in advancing machine learning methodologies, particularly in areas of model fairness, robustness, and efficiency. The shift to a virtual format reflects broader adaptations within the academic community during this period.
Vector Institute for Artificial Intelligence