The paper concerns online linear optimization where at each trial, the player, prior to prediction, receives a hint about the loss function. The hint has a form of a unit vector which is weakly correlated with the loss vector (its angle's cosine with loss vector is at least alpha). The paper shows that: - When the set of feasible actions is strongly convex, there exists an algorithm which gets logarithmic regret (in T). The algorithm is obtained by a reduction to the online learning problem with exp-concave losses. The bound is unimprovable in general, as shown in the Lower Bounds section.

This paper proposes a new neural architecture for reading comprehension. Compared to many other existing neural architectures, this model 1) densely connects all pairs of passage layers and question layers for encoding; 2) uses a component called Bidirectional Attention Connectors (BAC) for connecting any P-layer and Q-layer which employs an FM layer on top of commonly used bi-directional attention. The proposed architecture has been evaluated on four reading comprehension datasets and demonstrates strong empirical results. Overall, although the proposed ideas could be potentially interesting, I think the presented results (in the current presentation format) are not convincing enough. This makes the results less convincing.

We may earn revenue from the products available on this page and participate in affiliate programs. Building a robot at home is more than just a fun activity--it's a hands-on way to explore the exciting world of STEM [Science, Technology, Engineering, and Math]. Whether you're searching for a children's toy robot to inspire curiosity or a more advanced robot-building kit for older kids or teens, like our best overall Sillbird STEM 12-in-1 Education Solar Robot Toy, the best robot kits offer options for all ages and skill levels. Robot building kits offer a perfect blend of creativity and learning, teaching essential skills like coding, problem-solving, and engineering through play. From preschool-friendly robot toys to beginner robotics kits for older children, these sets provide a fantastic introduction to the basics of robotics.

This paper proposed an online learning algorithm for static and dynamic sum-product networks (SPNs), a type of probabilistic model with tractable inference. The authors essentially combine local structure search in SPNs with a hard variant of expectation-maximization [1]. The algorithm maintains empirical covariance estimates of product nodes and leverages statistical dependence tests to decide when to replace a product (factorized distribution) with either a new leaf or a mixture (sum node). The algorithm further includes a pruning mechanism in order to trim over-grown structures. The proposed method is called online Structure Learning with Running Average Update (oSLRAU).

A) My main concern with this paper is with respect to the main results (Theorems 2 and 3). It seems the authors have not put sufficient care to the fact that \partial \hat{Q} in Algorithm 2 is a biased estimator of the true gradient \partial Q. Also, \hat{Q} defined in Line 189 depends on \hat{\pi} which is an estimate of \pi. Thus, a probabilistic proof would require to look at a conditional probability of the estimation of Q depending on the estimation of \pi. B) Regardless of the above, the final high probability statement in Theorems 2 and 3, seem to be missing the union bound of the error probability in Assumption 1.

This paper proposes a new technique for training ensembles of predictors for supervised-learning tasks. Their main insight is to train individual members of the ensemble in a manner such that they specialize on different parts of the dataset reducing redundancy amongst members and better utilizing the capacity of the individual members. The hope is that ensembles formed out of such predictors will perform better than traditional ensembling techniques. The proposed technique explicitly enforces diversity in two ways: 1. inter-model diversity which makes individual models (predictors) different from each other and 2. intra-model diversity which makes predictors choose data points which are not all similar to each other so that they don't specialize in a very narrow region of the data distribution. This is posed as a bipartite graph matching problem which aims to find a matching between samples and models by selecting edges such that the smallest sum of edge costs is chosen (this is inverted to a maximization problem by subtracting from the highest constant cost one can have on the edges.) To avoid degenerate assignments another matching constraint is introduced which restricts the size of samples selected by each model as well.

This can be done in a batch or in an online fashion from the data. The authors provide an online approach for doing this and prove that it can achieve a regret O(1/sqrt(T)) as a function of data size T. 1. I believe the paper is not written very well. For example, the underlying idea of inverse optimization is not clearly explained as in main references such as [1]. Moreover, the problem has been posed in a quite general setup but at the end the authors have focused on the convex and strongly convex setup where they have used the result already in the literature with some minor modifications.

This paper identifies and separates (kernel) linear least-squares regression problems wherein carrying out multiple passes of stochastic gradient descent (SGD) over a training set can yield better statistical error than only a single pass. This is relevant to the core of machine learning theory, and relates to a line of work published at NIPS, ICML, COLT, and similar conferences in the past several years about the statistical error of one-pass, many-pass, and ERM-based learning. The authors focus on regression problems captured, by assumption, by two parameters: alpha, which governs the exponent of a power-law eigenvalue decay, and r, which governs a transformation under which the Hilbert norm of the optimal predictor is bounded. They refer to problems where r (alpha - 1) / (2 * alpha) as "hard". The main result of the paper is to show that for these "hard" problems, multiple SGD passes either achieve (minimax) optimal rates of statistical estimation, or at least improve the rate relative to a single pass. The results are interesting and might address an unanswered core question in machine learning, and the mathematical presentation is clear, with assumptions upfront.

This paper studies online convex optimization in dynamic environments, where one wishes to control the regret with respect to sequences of comparators, as opposed to a static comparator. The complexity of a comparator sequence is measured by its path length P_T, and the goal is to obtain an optimal regret bound simultaneously for all path lengths. A first bound in this setting was established in the pioneering paper [1], which showed that online gradient descent (OGD, projected on the convex compact domain) with step-size 1/sqrt(T) achieves an at most T {1/2} (1 P_T) regret for all comparator sequences. However, there is a gap between this upper bound and the (T (1 P_T)) {1/2} lower bound on worst-case regret established in Theorem 2 of the present paper, which is the first lower bound for this problem. On the other hand, if the path length P_T one wishes to compare against is known in advance, optimally tuning the step-size of OGD with respect to P_T in the OGD regret bound yields optimal (T (1 P_T)) {1/2} regret for this path length.

This paper formulates a framework which unifies several interactive learning problems with a structure such as interactive clustering. Next, the authors show that QBC can be generalized and kernelized to solve the problems in the framework. The consistency and rate of convergence are analyzed. I do not think I could judge the novelty of the theorems and proofs well. Thus, my comments focus on the practicality of the proposed algorithm, which I believe is relevant to its significance and the impact to the field.