Optimization in machine learning, both theoretical and applied, is presently dominated by first-order gradient methods such as stochastic gradient descent. Second-order optimization methods that involve second-order derivatives and/or second-order statistics of the data have become far less prevalent despite strong theoretical properties, due to their prohibitive computation, memory and communication costs. In an attempt to bridge this gap between theoretical and practical optimization, we present a proof-of-concept distributed system implementation of a second-order preconditioned method (specifically, a variant of full-matrix Adagrad), that along with a few yet critical algorithmic and numerical improvements, provides significant practical gains in convergence on state-of-the-art deep models and gives rise to actual wall-time improvements in practice compared to conventional first-order methods. Our design effectively utilizes the prevalent heterogeneous hardware architecture for training deep models which consists of a multicore CPU coupled with multiple accelerator units. We demonstrate superior performance on very large learning problems in machine translation where our distributed implementation runs considerably faster than existing gradient-based methods.

We present a formulation of deep learning that aims at producing a large margin classifier. The notion of margin, minimum distance to a decision boundary, has served as the foundation of several theoretically profound and empirically suc- cessful results for both classification and regression tasks. However, most large margin algorithms are applicable only to shallow models with a preset feature representation; and conventional margin methods for neural networks only enforce margin at the output layer. Such methods are therefore not well suited for deep networks. In this work, we propose a novel loss function to impose a margin on any chosen set of layers of a deep network (including input and hidden layers). Our formulation allows choosing any lp norm (p >= 1) on the metric measuring the margin. We demonstrate that the decision boundary obtained by our loss has nice properties compared to standard classification loss functions. Specifically, we show improved empirical results on the MNIST, CIFAR-10 and ImageNet datasets on multiple tasks: generalization from small training sets, corrupted labels, and robustness against adversarial perturbations. The resulting loss is general and complementary to existing data augmentation (such as random/adversarial input transform) and regularization techniques (weight decay, dropout, and batch norm).

Imprecise-reward Markov decision processes (IRMDPs) are MDPs in which the reward function is only partially specified (e.g., by some elicitation process). Recent work using minimax regret to solve IRMDPs has shown, despite theoretical intractability, how the set of policies that are nondom- inated w.r.t. reward uncertainty can be exploited to accelerate regret computation. However, the number of nondominated policies is generally so large as to undermine this leverage. In this paper, we show how the quality of the approximation can be improved online by pruning/adding nondominated policies during reward elicitation, while maintain- ing computational tractability. Drawing insights from the POMDP literature, we also develop a new anytime algorithm for constructing the set of nondominated policies with provable (anytime) error bounds.

Most models of utility elicitation in decision support and interactive optimization assume a predefined set of "catalog" features over which user preferences are expressed. However, users may differ in the features over which they are most comfortable expressing their preferences. In this work we consider the problem of feature elicitation: a user's utility function is expressed using features whose definitions (in terms of "catalog" features) are unknown. We cast this as a problem of concept learning, but whose goal is to identify only enough about the concept to enable a good decision to be recommended. We describe computational procedures for identifying optimal alternatives w.r.t. minimax regret in the presence of concept uncertainty; and describe several heuristic query strategies that focus on reduction of relevant concept uncertainty.

The specification of a Markov decision process (MDP) can be difficult. Reward function specification is especially problematic; in practice, it is often cognitively complex and time-consuming for users to precisely specify rewards. This work casts the problem of specifying rewards as one of preference elicitation and aims to minimize the degree of precision with which a reward function must be specified while still allowing optimal or near-optimal policies to be produced. We first discuss how robust policies can be computed for MDPs given only partial reward information using the minimax regret criterion. We then demonstrate how regret can be reduced by efficiently eliciting reward information using bound queries, using regret-reduction as a means for choosing suitable queries. Empirical results demonstrate that regret-based reward elicitation offers an effective way to produce near-optimal policies without resorting to the precise specification of the entire reward function.