More Like this

1. Deep ReLU networks have surprisingly few activation patterns

   Hanin, B.; Rolnick, D. (December 2019, NeurIPS)
2. Deep ReLU Networks Have Surprisingly Few Activation Patterns

   Hanin, Boris; Rolnick, David (January 2019, Advances in neural information processing systems)
3. Surprisingly Popular Voting with Concentric Rank-Order Models

   https://doi.org/10.1145/3696410.3714707  

   Hosseini, Hadi; Mandal, Debmalya; Puhan, Amrit (April 2025, ACM)
4. Optimal control of differentially flat systems is surprisingly easy

   https://doi.org/10.1016/j.automatica.2023.111404  

   Beaver, Logan E; Malikopoulos, Andreas A (January 2024, Automatica)

   NA (Ed.)

   As we move to increasingly complex cyber–physical systems (CPS), new approaches are needed to plan efficient state trajectories in real-time. In this paper, we propose an approach to significantly reduce the complexity of solving optimal control problems for a class of CPS with nonlinear dynamics. We exploit the property of differential flatness to simplify the Euler–Lagrange equations that arise during optimization, and this simplification eliminates the numerical instabilities that plague optimal control in general. We also present an explicit differential equation that describes the evolution of the optimal state trajectory, and we extend our results to consider both the unconstrained and constrained cases. Furthermore, we demonstrate the performance of our approach by generating the optimal trajectory for a planar manipulator with two revolute joints. We show in simulation that our approach is able to generate the constrained optimal trajectory in 4.5 ms while respecting workspace constraints and switching between a ‘left’ and ‘right’ bend in the elbow joint. 

   more » « less
5. Machine truth serum: a surprisingly popular approach to improving ensemble methods

   https://doi.org/10.1007/s10994-022-06183-y  

   Luo, Tianyi; Liu, Yang (January 2022, Machine Learning)

   Wisdom of the crowd (Surowiecki, 2005a) disclosed a striking fact that the majority voting answer from a crowd is usually more accurate than a few individual experts. The same story is observed in machine learning - ensemble methods (Dietterich, 2000) leverage this idea to exploit multiple machine learning algorithms in various settings e.g., supervised learning and semi-supervised learning to achieve better performance by aggregating the predictions of different algorithms than that obtained from any constituent algorithm alone. Nonetheless, the existing aggregating rule would fail when the majority answer of all the constituent algorithms is more likely to be wrong. In this paper, we extend the idea proposed in Bayesian Truth Serum (Prelec, 2004) that “a surprisingly more popular answer is more likely to be the true answer instead of the majority one” to supervised classification further improved by ensemble final predictions method and semi-supervised classification (e.g., MixMatch (Berthelot et al., 2019)) enhanced by ensemble data augmentations method. The challenge for us is to define or detect when an answer should be considered as being “surprising”. We present two machine learning aided methods which can reveal the truth when the minority instead of majority has the true answer on both settings of supervised and semi-supervised classification problems. We name our proposed method the Machine Truth Serum. Our experiments on a set of classification tasks (image, text, etc.) show that the classification performance can be further improved by applying Machine Truth Serum in the ensemble final predictions step (supervised) and in the ensemble data augmentations step (semi-supervised). 

   more » « less