Conference
Algorithmic Learning Theory
About: Algorithmic Learning Theory is an academic conference. The conference publishes majorly in the area(s): Learnability & Time complexity. Over the lifetime, 1210 publications have been published by the conference receiving 19851 citations.
Topics: Learnability, Time complexity, Algorithmic learning theory, Upper and lower bounds, Computer science
Papers published on a yearly basis
Papers
8 Oct 2005
TL;DR: The Hilbert-Schmidt Independence Criterion (HSIC) as mentioned in this paper is based on the eigenspectrum of covariance operators in reproducing kernel Hilbert spaces (RKHSs).
Abstract: We propose an independence criterion based on the eigenspectrum of covariance operators in reproducing kernel Hilbert spaces (RKHSs), consisting of an empirical estimate of the Hilbert-Schmidt norm of the cross-covariance operator (we term this a Hilbert-Schmidt Independence Criterion, or HSIC) This approach has several advantages, compared with previous kernel-based independence criteria First, the empirical estimate is simpler than any other kernel dependence test, and requires no user-defined regularisation Second, there is a clearly defined population quantity which the empirical estimate approaches in the large sample limit, with exponential convergence guaranteed between the two: this ensures that independence tests based on HSIC do not suffer from slow learning rates Finally, we show in the context of independent component analysis (ICA) that the performance of HSIC is competitive with that of previously published kernel-based criteria, and of other recently published ICA methods
1,548 citations
1 Oct 2007
TL;DR: This work describes a technique for comparing distributions without the need for density estimation as an intermediate step, which relies on mapping the distributions into a reproducing kernel Hilbert space.
Abstract: We describe a technique for comparing distributions without the need for density estimation as an intermediate step. Our approach relies on mapping the distributions into a reproducing kernel Hilbert space. Applications of this technique can be found in two-sample tests, which are used for determining whether two sets of observations arise from the same distribution, covariate shift correction, local learning, measures of independence, and density estimation.
1,056 citations
1 Feb 1991
TL;DR: A discussion of the feasibility of extending the RLGG framework to allow for the invention of new predicates and the possible relationship between Algorithmic Complexity theory and Probably-Approximately-Correct Learning is discussed.
Abstract: A new research area, Inductive Logic Programming, is presently emerging. While inheriting various positive characteristics of the parent subjects of Logic Programming and Machine Learning, it is hoped that the new area will overcome many of the limitations of its forebears. The background to present developments within this area is discussed and various goals and aspirations for the increasing body of researchers are identified. Inductive Logic Programming needs to be based on sound principles from both Logic and Statistics. On the side of statistical justification of hypotheses we discuss the possible relationship between Algorithmic Complexity theory and Probably-Approximately-Correct (PAC) Learning. In terms of logic we provide a unifying framework for Muggleton and Buntine’s Inverse Resolution (IR) and Plotkin’s Relative Least General Generalisation (RLGG) by rederiving RLGG in terms of IR. This leads to a discussion of the feasibility of extending the RLGG framework to allow for the invention of new predicates, previously discussed only within the context of IR.
849 citations
29 Oct 2012
TL;DR: The question of the optimality of Thompson Sampling for solving the stochastic multi-armed bandit problem is answered positively for the case of Bernoulli rewards by providing the first finite-time analysis that matches the asymptotic rate given in the Lai and Robbins lower bound for the cumulative regret.
Abstract: The question of the optimality of Thompson Sampling for solving the stochastic multi-armed bandit problem had been open since 1933. In this paper we answer it positively for the case of Bernoulli rewards by providing the first finite-time analysis that matches the asymptotic rate given in the Lai and Robbins lower bound for the cumulative regret. The proof is accompanied by a numerical comparison with other optimal policies, experiments that have been lacking in the literature until now for the Bernoulli case.
638 citations
3 Oct 2009
TL;DR: The main result is that the required exploration-exploitation trade-offs are qualitatively different, in view of a general lower bound on the simple regret in terms of the cumulative regret.
Abstract: We consider the framework of stochastic multi-armed bandit problems and study the possibilities and limitations of strategies that perform an online exploration of the arms. The strategies are assessed in terms of their simple regret, a regret notion that captures the fact that exploration is only constrained by the number of available rounds (not necessarily known in advance), in contrast to the case when the cumulative regret is considered and when exploitation needs to be performed at the same time.We believe that this performance criterion is suited to situations when the cost of pulling an arm is expressed in terms of resources rather than rewards. We discuss the links between the simple and the cumulative regret. The main result is that the required exploration-exploitation trade-offs are qualitatively different, in view of a general lower bound on the simple regret in terms of the cumulative regret.
554 citations
Performance Metrics
| Year | Papers |
|---|---|
| 2021 | 70 |
| 2020 | 114 |
| 2019 | 78 |
| 2018 | 85 |
| 2017 | 34 |
| 2016 | 31 |