NLP research has explored different neural model architectures and sizes, datasets, training objectives, and transfer learning techniques. However, the choice of optimizer during training has not been explored as extensively. Typically, some variant of Stochastic Gradient Descent (SGD) is employed, selected among numerous variants, using unclear criteria, often with minimal or no tuning of the optimizer's hyperparameters. Experimenting with five GLUE datasets, two models (DistilBERT and DistilRoBERTa), and seven popular optimizers (SGD, SGD with Momentum, Adam, AdaMax, Nadam, AdamW, and AdaBound), we find that when the hyperparameters of the optimizers are tuned, there is no substantial difference in test performance across the five more elaborate (adaptive) optimizers, despite differences in training loss. Furthermore, tuning just the learning rate is in most cases as good as tuning all the hyperparameters. Hence, we recommend picking any of the best-behaved adaptive optimizers (e.g., Adam) and tuning only its learning rate. When no hyperparameter can be tuned, SGD with Momentum is the best choice.

To effectively optimize Takagi-Sugeno-Kang (TSK) fuzzy systems for regression problems, a mini-batch gradient descent with regularization, DropRule, and AdaBound (MBGD-RDA) algorithm was recently proposed. This paper further proposes FCM-RDpA, which improves MBGD-RDA by replacing the grid partition approach in rule initialization by fuzzy c-means clustering, and AdaBound by Powerball AdaBelief, which integrates recently proposed Powerball gradient and AdaBelief to further expedite and stabilize parameter optimization. Extensive experiments on 22 regression datasets with various sizes and dimensionalities validated the superiority of FCM-RDpA over MBGD-RDA, especially when the feature dimensionality is higher. We also propose an additional approach, FCM-RDpAx, that further improves FCM-RDpA by using augmented features in both the antecedents and consequents of the rules.

Adaptive gradient methods such as Adam have gained extreme popularity due to their success in training complex neural networks and less sensitivity to hyperparameter tuning compared to SGD. However, it has been recently shown that Adam can fail to converge and might cause poor generalization -- this lead to the design of new, sophisticated adaptive methods which attempt to generalize well while being theoretically reliable. In this technical report we focus on AdaBound, a promising, recently proposed optimizer. We present a stochastic convex problem for which AdaBound can provably take arbitrarily long to converge in terms of a factor which is not accounted for in the convergence rate guarantee of Luo et al. (2019). We present a new $O(\sqrt T)$ regret guarantee under different assumptions on the bound functions, and provide empirical results on CIFAR suggesting that a specific form of momentum SGD can match AdaBound's performance while having less hyperparameters and lower computational costs.

Takagi-Sugeno-Kang (TSK) fuzzy systems are very useful machine learning models for regression problems. However, to our knowledge, there has not existed an efficient and effective training algorithm that enables them to deal with big data. Inspired by the connections between TSK fuzzy systems and neural networks, we extend three powerful neural network optimization techniques, i.e., mini-batch gradient descent, regularization, and AdaBound, to TSK fuzzy systems, and also propose a novel DropRule technique specifically for training TSK fuzzy systems. Our final algorithm, mini-batch gradient descent with regularization, DropRule and AdaBound (MBGD-RDA), can achieve fast convergence in training TSK fuzzy systems, and also superior generalization performance in testing. It can be used for training TSK fuzzy systems on datasets of any size; however, it is particularly useful for big datasets, on which currently no other efficient training algorithms exist.