Results 1  10
of
32
Using advice to transfer knowledge acquired in one reinforcement learning task to another
 In Proceedings of the Sixteenth European Conference on Machine Learning
, 2005
"... Abstract. We present a method for transferring knowledge learned in one task to a related task. Our problem solvers employ reinforcement learning to acquire a model for one task. We then transform that learned model into advice for a new task. A human teacher provides a mapping from the old task to ..."
Abstract

Cited by 50 (11 self)
 Add to MetaCart
Abstract. We present a method for transferring knowledge learned in one task to a related task. Our problem solvers employ reinforcement learning to acquire a model for one task. We then transform that learned model into advice for a new task. A human teacher provides a mapping from the old task to the new task to guide this knowledge transfer. Advice is incorporated into our problem solver using a knowledgebased support vector regression method that we previously developed. This advicetaking approach allows the problem solver to refine or even discard the transferred knowledge based on its subsequent experiences. We empirically demonstrate the effectiveness of our approach with two games from the RoboCup soccer simulator: KeepAway and BreakAway. Our results demonstrate that a problem solver learning to play BreakAway using advice extracted from KeepAway outperforms a problem solver learning without the benefit of such advice. 1
Exact 1Norm support vector machines via unconstrained convex differentiable minimization
 Data Mining Institute, Computer Sciences Department, University of Wisconsin
, 2005
"... Support vector machines utilizing the 1norm, typically set up as linear programs (Mangasarian, 2000; Bradley and Mangasarian, 1998), are formulated here as a completely unconstrained minimization of a convex differentiable piecewisequadratic objective function in the dual space. The objective fu ..."
Abstract

Cited by 20 (1 self)
 Add to MetaCart
Support vector machines utilizing the 1norm, typically set up as linear programs (Mangasarian, 2000; Bradley and Mangasarian, 1998), are formulated here as a completely unconstrained minimization of a convex differentiable piecewisequadratic objective function in the dual space. The objective function, which has a Lipschitz continuous gradient and contains only one additional finite parameter, can be minimized by a generalized Newton method and leads to an exact solution of the support vector machine problem. The approach here is based on a formulation of a very general linear program as an unconstrained minimization problem and its application to support vector machine classification problems. The present approach which generalizes both (Mangasarian, 2004) and (Fung and Mangasarian, 2004) is also applied to nonlinear approximation where a minimal number of nonlinear kernel functions are utilized to approximate a function from a given number of function values. 1.
Semisupervised regression with order preferences
, 2006
"... Following a discussion on the general form of regularization for semisupervised learning, we propose a semisupervised regression algorithm. It is based on the assumption that we have certain order preferences on unlabeled data (e.g., point x1 has a larger target value than x2). Semisupervised lea ..."
Abstract

Cited by 11 (1 self)
 Add to MetaCart
(Show Context)
Following a discussion on the general form of regularization for semisupervised learning, we propose a semisupervised regression algorithm. It is based on the assumption that we have certain order preferences on unlabeled data (e.g., point x1 has a larger target value than x2). Semisupervised learning consists of enforcing the order preferences as regularization in a risk minimization framework. The optimization problem can be effectively solved by a linear program. Experiments show that the proposed semisupervised regression outperforms standard regression. 1 Semisupervised learning as regularization on unlabeled data Semisupervised learning works when its assumption on unlabeled data, often expressed as regularization, fits the reality of the problem domain. In this paper we first generalize the regularization formulation of some common semisupervised learning approaches, namely manifold regularization, semisupervised support vector machines, and multiview learning [1, 2, 3]. Regularization for each individual approach is not new. However these approaches have been studied largely in isolation. Our general form serves as a bridge to connect them, and to inspire novel semisupervised approaches. As an example of the latter, we propose a novel algorithm for semisupervised regression. The proposed regression algorithm is able to incorporate domain knowledge about the relative order of target values on unlabeled points. It thus differs from, and complements, existing semisupervised regression methods, which do not use such domain knowledge but require multiple views [4, 5]. Let us review the three common semisupervised learning methods. Manifold regularization [6, 7] generalizes
Simpler knowledgebased support vector machines
 In Proceedings of the TwentyThird International Conference on Machine Learning
, 2006
"... If appropriately used, prior knowledge can significantly improve the predictive accuracy of learning algorithms or reduce the amount of training data needed. In this paper we introduce a simple method to incorporate prior knowledge in support vector machines by modifying the hypothesis space rathe ..."
Abstract

Cited by 11 (0 self)
 Add to MetaCart
If appropriately used, prior knowledge can significantly improve the predictive accuracy of learning algorithms or reduce the amount of training data needed. In this paper we introduce a simple method to incorporate prior knowledge in support vector machines by modifying the hypothesis space rather than the optimization problem. The optimization problem is amenable to solution by the constrained concave convex procedure, which finds a local optimum. The paper discusses different kinds of prior knowledge and demonstrates the applicability of the approach in some characteristic experiments. 1.
A simple and effective method for incorporating advice into kernel methods
 In AAAI
, 2006
"... We propose a simple mechanism for incorporating advice (prior knowledge), in the form of simple rules, into supportvector methods for both classification and regression. Our approach is based on introducing inequality constraints associated with datapoints that match the advice. These constrained d ..."
Abstract

Cited by 7 (5 self)
 Add to MetaCart
We propose a simple mechanism for incorporating advice (prior knowledge), in the form of simple rules, into supportvector methods for both classification and regression. Our approach is based on introducing inequality constraints associated with datapoints that match the advice. These constrained datapoints can be standard examples in the training set, but can also be unlabeled data in a semisupervised, advicetaking approach. Our new approach is simpler to implement and more efficiently solved than the knowledgebased support vector classification methods of Fung, Mangasarian and Shavlik (2002; 2003) and the knowledgebased support vector regression method of Mangasarian, Shavlik, and Wild (2004), while performing approximately as well as these more complex approaches. Experiments using our new approach on a synthetic task and a reinforcementlearning problem within the RoboCup soccer simulator show that our advicetaking method can significantly outperform a method without advice and perform similarly to prior advicetaking, supportvector machines.
Transfer Learning via Advice Taking
"... Abstract The goal of transfer learning is to speed up learning in a new task by transferring knowledge from one or more related source tasks. We describe a transfer method in which a reinforcement learner analyzes its experience in the source task and learns rules to use as advice in the target task ..."
Abstract

Cited by 7 (2 self)
 Add to MetaCart
(Show Context)
Abstract The goal of transfer learning is to speed up learning in a new task by transferring knowledge from one or more related source tasks. We describe a transfer method in which a reinforcement learner analyzes its experience in the source task and learns rules to use as advice in the target task. The rules, which are learned via inductive logic programming, describe the conditions under which an action is successful in the source task. The advicetaking algorithm used in the target task allows a reinforcement learner to benefit from rules even if they are imperfect. A humanprovided mapping describes the alignment between the source and target tasks, and may also include advice about the differences between them. Using three tasks in the RoboCup simulated soccer domain, we demonstrate that this transfer method can speed up reinforcement learning substantially. 1
2007) Incorporating Prior Domain Knowledge into Inductive Machine Learning Its implementation in contemporary capital markets
"... The paper reviews the recent developments of incorporating prior domain knowledge into inductive machine learning, and proposes a guideline that incorporates prior domain knowledge in three key issues of inductive machine learning algorithms: consistency, generalization and convergence. With respect ..."
Abstract

Cited by 6 (0 self)
 Add to MetaCart
(Show Context)
The paper reviews the recent developments of incorporating prior domain knowledge into inductive machine learning, and proposes a guideline that incorporates prior domain knowledge in three key issues of inductive machine learning algorithms: consistency, generalization and convergence. With respect to each issue, this paper gives some approaches to improve the performance of the inductive machine learning algorithms and discusses the risks of incorporating domain knowledge. As a case study, a hierarchical modelling method, VQSVM, is proposed and tested over some imbalanced data sets with various imbalance ratios and various numbers of subclasses. 1.
Generative Prior Knowledge for Discriminative Classification
"... We present a novel framework for integrating prior knowledge into discriminative classifiers. Our framework allows discriminative classifiers such as Support Vector Machines (SVMs) to utilize prior knowledge specified in the generative setting. The dual objective of fitting the data and respecting p ..."
Abstract

Cited by 5 (0 self)
 Add to MetaCart
We present a novel framework for integrating prior knowledge into discriminative classifiers. Our framework allows discriminative classifiers such as Support Vector Machines (SVMs) to utilize prior knowledge specified in the generative setting. The dual objective of fitting the data and respecting prior knowledge is formulated as a bilevel program, which is solved (approximately) via iterative application of secondorder cone programming. To test our approach, we consider the problem of using WordNet (a semantic database of English language) to improve lowsample classification accuracy of newsgroup categorization. WordNet is viewed as an approximate, but readily available source of background knowledge, and our framework is capable of utilizing it in a flexible way. 1.
Rotational prior knowledge for SVMs
 In Proceedings of the Sixteenth European Conference on Machine Learning
, 2005
"... Abstract. Incorporation of prior knowledge into the learning process can significantly improve lowsample classification accuracy. We show how to introduce prior knowledge into linear support vector machines in form of constraints on the rotation of the normal to the separating hyperplane. Such know ..."
Abstract

Cited by 4 (1 self)
 Add to MetaCart
(Show Context)
Abstract. Incorporation of prior knowledge into the learning process can significantly improve lowsample classification accuracy. We show how to introduce prior knowledge into linear support vector machines in form of constraints on the rotation of the normal to the separating hyperplane. Such knowledge frequently arises naturally, e.g., as inhibitory and excitatory influences of input variables. We demonstrate that the generalization ability of rotationallyconstrained classifiers is improved by analyzing their VC and fatshattering dimensions. Interestingly, the analysis shows that largemargin classification framework justifies the use of stronger prior knowledge than the traditional VC framework. Empirical experiments with text categorization and political party affiliation prediction confirm the usefulness of rotational prior knowledge. 1
Automating the ILP Setup Task: Converting User Advice about Specific Examples into General Background Knowledge
 International Conference on Inductive Logic Programming
, 2010
"... Abstract. Inductive Logic Programming (ILP) provides an effective method of learning logical theories given a set of positive examples, a set of negative examples, a corpus of background knowledge, and specification of a search space (e.g., via mode definitions) from which to compose the theories. W ..."
Abstract

Cited by 3 (3 self)
 Add to MetaCart
(Show Context)
Abstract. Inductive Logic Programming (ILP) provides an effective method of learning logical theories given a set of positive examples, a set of negative examples, a corpus of background knowledge, and specification of a search space (e.g., via mode definitions) from which to compose the theories. While specifying positive and negative examples is relatively straightforward, composing effective background knowledge and searchspace definition requires detailed understanding of many aspects of the ILP process and limits the usability of ILP. We introduce two techniques to automate the use of ILP for a nonILP expert. These techniques include automatic generation of background knowledge from usersupplied information in the form of a simple relevance language, used to describe important aspects of specific training examples, and an iterativedeepeningstyle search process. Keywords: Advice Taking, Human Teaching of Machines. 1