Search results for: Olivier Teytaud

Items from 1 to 20 out of 58 results

chapter

Boosting Active Learning to Optimality: A Tractable Monte-Carlo, Billiard-Based Algorithm

Philippe Rolet, Michèle Sebag, Olivier Teytaud

Lecture Notes in Computer Science > Machine Learning and Knowledge Discovery in Databases > Regular Papers > 302-317

This paper focuses on Active Learning with a limited number of queries; in application domains such as Numerical Engineering, the size of the training set might be limited to a few dozen or hundred examples due to computational constraints. Active Learning under bounded resources is formalized as a finite horizon Reinforcement Learning problem, where the sampling strategy aims at minimizing the expectation...

chapter

Algorithms (X, sigma, eta): Quasi-random Mutations for Evolution Strategies

Anne Auger, Mohammed Jebalia, Olivier Teytaud

Lecture Notes in Computer Science > Artificial Evolution > Advances > 296-307

Randomization is an efficient tool for global optimization. We here define a method which keeps : – the order 0 of evolutionary algorithms (no gradient) ; – the stochastic aspect of evolutionary algorithms ; – the efficiency of so-called ”low-dispersion” points ; and which ensures under mild assumptions global convergence with linear convergence rate. We use i) sampling...

chapter

Robust Optimizers for Nonlinear Programming in Approximate Dynamic Programming

Olivier Teytaud, Sylvain Gelly

Lecture Notes in Electrical Engineering > Informatics in Control, Automation and Robotics > Intelligent Control Systems and Optimization > 95-106

Many stochastic dynamic programming tasks in continuous action-spaces are tackled through discretization. We here avoid discretization; then, approximate dynamic programming (ADP) involves (i) many learning tasks, performed here by Support Vector Machines, for Bellman-function-regression (ii) many non-linear-optimization tasks for action-selection, for which we compare many algorithms. We include...

chapter

Learning opening books in partially observable games: Using random seeds in Phantom Go

Tristan Cazenave, Jialin Liu, Fabien Teytaud, Olivier Teytaud

2016 IEEE Conference on Computational Intelligence and Games (CIG) > 1 - 7

2016 IEEE Conference on Computational Intelligence and Games (CIG)

Many artificial intelligences (AIs) are randomized. One can be lucky or unlucky with the random seed; we quantify this effect and show that, maybe contrarily to intuition, this is far from being negligible. Then, we apply two different existing algorithms for selecting good seeds and good probability distributions over seeds. This mainly leads to learning an opening book. We apply this to Phantom...

chapter

Biasing Monte-Carlo Simulations through RAVE Values

Arpad Rimmel, Fabien Teytaud, Olivier Teytaud

Lecture Notes in Computer Science > Computers and Games > 59-68

The Monte-Carlo Tree Search algorithm has been successfully applied in various domains. However, its performance heavily depends on the Monte-Carlo part. In this paper, we propose a generic way of improving the Monte-Carlo simulations by using RAVE values, which already strongly improved the tree part of the algorithm. We prove the generality and efficiency of our approach by showing improvements...

chapter

Conditioning, Halting Criteria and Choosing λ

Olivier Teytaud

Lecture Notes in Computer Science > Artificial Evolution > Theory in GA and ES > 196-206

We show the convergence of 1 + λ-ES with standard step-size update-rules on a large family of fitness functions without any convexity assumption or quasi-convexity assumptions ([3,6]). The result provides a rule for choosing λ and shows the consistency of halting criteria based on thresholds on the step-size. The family of functions under work is defined through a condition-number that generalizes...

chapter

Multigrid Convergence and Surface Area Estimation

David Coeurjolly, Frédéric Flin, Olivier Teytaud, Laure Tougne

Lecture Notes in Computer Science > Geometry, Morphology, and Computational Imaging > 101-119

Surface area of discrete objects is an important feature for model-based image analysis. In this article, we present a theoretical framework in order to prove multigrid convergence of surface area estimators based on discrete normal vector field integration. The paper details an algorithm which is optimal in time and multigrid convergent to estimate the surface area and a very efficient algorithm...

chapter

Kernel Based Image Classification

Olivier Teytaud, David Sarrut

Lecture Notes in Computer Science > Artificial Neural Networks — ICANN 2001 > Kernel Methods > 369-375

In this study, we consider low-level image classification, with several machine learning algorithms adapted to high dimension problems: kernel-based algorithms. The first is Support Vector Machines (SVM), the second is Bayes Point Machines (BPM). We compare these algorithms based on strong mathematical results and nice geometrical arguments in a feature space to the simplest algorithm we could imagine...

chapter

Complexity Bounds for Batch Active Learning in Classification

Philippe Rolet, Olivier Teytaud

Lecture Notes in Computer Science > Machine Learning and Knowledge Discovery in Databases > Regular Papers > 293-305

Active learning [1] is a branch of Machine Learning in which the learning algorithm, instead of being directly provided with pairs of problem instances and their solutions (their labels), is allowed to choose, from a set of unlabeled data, which instances to query. It is suited to settings where labeling instances is costly. This paper analyzes the speed-up of batch (parallel) active learning compared...

chapter

Bandit-Based Genetic Programming

Jean-Baptiste Hoock, Olivier Teytaud

Lecture Notes in Computer Science > Genetic Programming > Posters > 268-277

We consider the validation of randomly generated patterns in a Monte-Carlo Tree Search program. Our bandit-based genetic programming (BGP) algorithm, with proved mathematical properties, outperformed a highly optimized handcrafted module of a well-known computer-Go program with several world records in the game of Go.

chapter

Adaptive Noisy Optimization

Philippe Rolet, Olivier Teytaud

Lecture Notes in Computer Science > Applications of Evolutionary Computation > EvoSTOC Contributions > 592-601

In this paper, adaptive noisy optimization on variants of the noisy sphere model is considered, i.e. optimization in which the same algorithm is able to adapt to several frameworks, including some for which no bound has never been derived. Incidentally, bounds derived by [16] for noise quickly decreasing to zero around the optimum are extended to the more general case of a positively lower-bounded...

chapter

A Statistical Learning Perspective of Genetic Programming

Nur Merve Amil, Nicolas Bredeche, Christian Gagné, Sylvain Gelly, more

Lecture Notes in Computer Science > Genetic Programming > Posters > 327-338

This paper proposes a theoretical analysis of Genetic Programming (GP) from the perspective of statistical learning theory, a well grounded mathematical toolbox for machine learning. By computing the Vapnik-Chervonenkis dimension of the family of programs that can be inferred by a specific setting of GP, it is proved that a parsimonious fitness ensures universal consistency. This means that the empirical...

chapter

On the Ultimate Convergence Rates for Isotropic Algorithms and the Best Choices Among Various Forms of Isotropy

Olivier Teytaud, Sylvain Gelly, Jérémie Mary

Lecture Notes in Computer Science > Parallel Problem Solving from Nature - PPSN IX > Theory > 32-41

In this paper, we show universal lower bounds for isotropic algorithms, that hold for any algorithm such that each new point is the sum of one already visited point plus one random isotropic direction multiplied by any step size (whenever the step size is chosen by an oracle with arbitrarily high computational power). The bound is 1–O(1/d) for the constant in the linear convergence (i.e. the constant...

chapter

Bandit-Based Estimation of Distribution Algorithms for Noisy Optimization: Rigorous Runtime Analysis

Philippe Rolet, Olivier Teytaud

Lecture Notes in Computer Science > Learning and Intelligent Optimization > Main Track (Regular Papers) > 97-110

We show complexity bounds for noisy optimization, in frameworks in which noise is stronger than in previously published papers[19]. We also propose an algorithm based on bandits (variants of [16]) that reaches the bound within logarithmic factors. We emphasize the differences with empirical derived published algorithms. Complete mathematical proofs can be found in [26].

chapter

Consistency Modifications for Automatically Tuned Monte-Carlo Tree Search

Vincent Berthier, Hassen Doghmen, Olivier Teytaud

Lecture Notes in Computer Science > Learning and Intelligent Optimization > Main Track (Regular Papers) > 111-124

Monte-Carlo Tree Search algorithms (MCTS [4,6]), including upper confidence trees (UCT [9]), are known for their impressive ability in high dimensional control problems. Whilst the main testbed is the game of Go, there are increasingly many applications [13,12,7]; these algorithms are now widely accepted as strong candidates for high-dimensional control applications. Unfortunately, it is known that...

chapter

On the Parallel Speed-Up of Estimation of Multivariate Normal Algorithm and Evolution Strategies

Fabien Teytaud, Olivier Teytaud

Lecture Notes in Computer Science > Applications of Evolutionary Computing > EvoNUM Contributions > 655-664

Motivated by parallel optimization, we experiment EDA-like adaptation-rules in the case of λ large. The rule we use, essentially based on estimation of multivariate normal algorithm, is (i) compliant with all families of distributions for which a density estimation algorithm exists (ii) simple (iii) parameter-free (iv) better than current rules in this framework of λ large. The speed-up as a function...

chapter

Creating an Upper-Confidence-Tree Program for Havannah

Fabien Teytaud, Olivier Teytaud

Lecture Notes in Computer Science > Advances in Computer Games > 65-74

Monte-Carlo Tree Search and Upper Confidence Bounds provided huge improvements in computer-Go. In this paper, we test the generality of the approach by experimenting on the game, Havannah, which is known for being especially difficult for computers. We show that the same results hold, with slight differences related to the absence of clearly known patterns for the game of Havannah, in spite of the...

chapter

Segmentation and Length Estimation of 3D Discrete Curves

David Coeurjolly, Isabelle Debled-Rennesson, Olivier Teytaud

Lecture Notes in Computer Science > Digital and Image Geometry > Multigrid Convergence > 299-317

We propose in this paper an arithmetical definition of 3-D discrete lines as well as an efficient construction algorithm. From this notion, an algorithm of 3-D discrete lines segmentation has been developed. It is then used to calculate the length of a discrete curve. Ap roof of the multigrid convergence of length estimators is presented.

chapter

Association Rule Interestingness: Measure and Statistical Validation

Stephane Lallich, Olivier Teytaud, Elie Prudhomme

Studies in Computational Intelligence > Quality Measures in Data Mining > Rule quality and validation > 251-275

chapter

Continuous Upper Confidence Trees

Adrien Couëtoux, Jean-Baptiste Hoock, Nataliya Sokolovska, Olivier Teytaud, more

Lecture Notes in Computer Science > Learning and Intelligent Optimization > Main Track (Regular Papers) > 433-445

Upper Confidence Trees are a very efficient tool for solving Markov Decision Processes; originating in difficult games like the game of Go, it is in particular surprisingly efficient in high dimensional problems. It is known that it can be adapted to continuous domains in some cases (in particular continuous action spaces). We here present an extension of Upper Confidence Trees to continuous stochastic...

Publication date

Set your own date range

Publication type

book (46)
article (12)

Keywords

GAMES (7)
COMPUTERS (4)
FUZZY MARKUP LANGUAGE (4)
MONTE CARLO METHODS (4)
ADAPTIVE ASSESSMENT (3)
ALGORITHM DESIGN AND ANALYSIS (3)
COMPUTER GO (3)
FUZZY ONTOLOGY (3)
GAME OF GO (3)
ADAPTATION MODELS (2)
ADAPTIVE SYSTEMS (2)
COMPUTATIONAL MODELING (2)
EDUCATIONAL INSTITUTIONS (2)
EVOLUTION STRATEGIES (2)
GAME THEORY (2)
MONTE-CARLO TREE SEARCH (2)
NASH EQUILIBRIUM (2)
NOISY OPTIMIZATION (2)
ONTOLOGIES (2)
RUNTIME ANALYSIS (2)
TIN (2)
VC-DIMENSION (2)
3D DISCRETE LINE (1)
ADDITIVE NOISE (1)
ALGORITHM SELECTION (1)
APPLICATIONS (1)
ARTIFICIAL INTELLIGENCE (1)
ASH (1)
ASSEMBLY (1)
ATMOSPHERIC MODELING (1)
BANDIT PROBLEMS (1)
BANDIT-BASED MONTE-CARLO PLANNING (1)
BAYES METHODS (1)
BELIEF STATE ESTIMATION (1)
BLACK-BOX NOISY OPTIMIZATION (1)
BOOSTING (1)
COMPLEXITY (1)
COMPUTER GAMES (1)
COMPUTER SCIENCE (1)
CONTEXT (1)
CONVERGENCE RATIO (1)
CONVERGENCE SPEED (1)
DECIDABILITY (1)
DECISION TREES (1)
DERANDOMIZATION (1)
DIFFERENTIAL EVOLUTION (1)
DISCRETE CURVE (1)
DISCRETE OPTIMIZATION (1)
DISCRETIZATION (1)
DYNAMIC PROGRAMMING (1)
ERROR ANALYSIS (1)
ESTIMATION (1)
EVOLUTION ALGORITHMS (1)
EXPENSIVE OPTIMIZATION (1)
EXPLORATION POLICY (1)
FUZZ IEEE 2009 (1)
FUZZY INFERENCE MECHANISM (1)
FUZZY SYSTEM (1)
FUZZY SYSTEMS (1)
GAME GO (1)
GENETICS (1)
HALTING PROBLEM (1)
HUMANS (1)
IEEE INTERNATIONAL CONFERENCE (1)
INTELLIGENT AGENT (1)
ITEM RESPONSE THEORY (1)
KNOWLEDGE MANAGEMENT (1)
KOLMOGOROV’S EXTENSION THEOREM (1)
LENGTH ESTIMATION (1)
MATHEMATICAL MODEL (1)
MATIYASEVICH MACHINES (1)
MCTS (1)
MINES GAME (1)
MONTE CARLO (1)
MONTE CARLO TREE SEARCH (1)
MONTE CARLO TREE SEARCH (MCTS) (1)
NO-FREE-LUNCH (1)
NODE(TAG=, PARTS=[NODE(TAG=I, PARTS=[Q])])-LEARNING (1)
NOGO (1)
NOISE (1)
NOISE MEASUREMENT (1)
ONTOLOGY (1)
OPTIMIZATION (1)
PARAMETER ESTIMATION (1)
PARTIALLY OBSERVABLE MARKOV DECISION PROCESSES (1)
PHANTOMS (1)
PORTFOLIOS (1)
PRAGMATICS (1)
PROBABILITY (1)
PROBABILITY DISTRIBUTION (1)
RECOMMENDATION POLICY (1)
RESAMPLING (1)
SEGMENTATION (1)
SIGN CONDITIONS (1)
SIGN PATTERNS (1)
SIMPLE REGRET (1)
SPECIAL ISSUES AND SECTIONS (1)
SPHERE FUNCTION (1)
STANDARDS (1)
TRACES (1)
more

Data set

Springer (35)
ieee (19)
Elsevier (4)

INFONA - science communication portal

Search results for: Olivier Teytaud

Boosting Active Learning to Optimality: A Tractable Monte-Carlo, Billiard-Based Algorithm

Algorithms (X, sigma, eta): Quasi-random Mutations for Evolution Strategies

Robust Optimizers for Nonlinear Programming in Approximate Dynamic Programming

Learning opening books in partially observable games: Using random seeds in Phantom Go

Biasing Monte-Carlo Simulations through RAVE Values

Conditioning, Halting Criteria and Choosing λ

Multigrid Convergence and Surface Area Estimation

Kernel Based Image Classification

Complexity Bounds for Batch Active Learning in Classification

Bandit-Based Genetic Programming

Adaptive Noisy Optimization

A Statistical Learning Perspective of Genetic Programming

On the Ultimate Convergence Rates for Isotropic Algorithms and the Best Choices Among Various Forms of Isotropy

Bandit-Based Estimation of Distribution Algorithms for Noisy Optimization: Rigorous Runtime Analysis

Consistency Modifications for Automatically Tuned Monte-Carlo Tree Search

On the Parallel Speed-Up of Estimation of Multivariate Normal Algorithm and Evolution Strategies

Creating an Upper-Confidence-Tree Program for Havannah

Segmentation and Length Estimation of 3D Discrete Curves

Association Rule Interestingness: Measure and Statistical Validation

Continuous Upper Confidence Trees

Filter options

Publication date

Publication type

Keywords

Data set

Journal

INFONA - science communication portal

Search results for: Olivier Teytaud

Add recipient

Sending message cancelled

Are you sure you want to cancel sending this message?

Send message

Filter options

Publication date

Date range setting

Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.

Publication type

Keywords

Data set

Journal

Reporting an error / abuse

Sending the report failed

Accessibility options