Search

The first language has been considered the main source of transfer for the acquisition of further languages for a long time. However, studies by linguists such as Hufeisen (Tertiarsprachen. Theorien, modelle, methode. Tubingen, Stauffenburg, 169–183, 1998), Cenoz and Jessner (The english in Europe. The acquisition of a third language. Multilingual Matters, Clevedon, p. 9, 2000), Cenoz (Cross-linguistic influence in third language acquisition: psycholinguistic perspectives. Multilingual Matters, Clevedon, pp. 8–19, 2001), and Hammarberg (Cross-linguistic influence in third language acquisition: psycholinguistic perspectives. Multilingual Matters, Clevedon, pp. 21–41, 2001) found that the L3 learner has already acquired one second language (L2) and thus this knowledge plays a role in the acquisition of other foreign languages. The research aim of our whole project is to account for the ways and paths of the activation of prior languages in the multilingual acquisition of English. The foci of the present study are the negative transfer phenomena of the second language (L2) German in the third language (L3) acquisition of English. In the present study, we set forth the possibility of negative transfer in L3 interlanguage based on the ‘L2 status factor’ and attempt to explain whether these occurrences of L2 negative transfer can be documented in the L3 written data from learners with different L1 s and at different L3 proficiency levels.

In this paper we investigate using multiple mappings for transfer learning in reinforcement learning tasks. We propose two different transfer learning algorithms that are able to manipulate multiple inter-task mappings for both model-learning and model-free reinforcement learning algorithms. Both algorithms incorporate mechanisms to select the appropriate mappings, helping to avoid the phenomenon of negative transfer. The proposed algorithms are evaluated in the Mountain Car and Keepaway domains. Experimental results show that the use of multiple inter-task mappings can significantly boost the performance of transfer learning methodologies, relative to using a single mapping or learning without transfer.

In real applications of inductive learning, labeled instances are often deficient. The countermeasure is either to ask experts to label informative instances in active learning, or to borrow useful information from abundant labeled instances in the source domain in transfer learning. Due to the high cost of querying experts, it is promising to integrate the two methodologies into a more robust and reliable classification framework to compensate the disadvantages of both methods. Recently, a few research studies have been investigated to integrate the two methods together, which is called transfer active learning. However, when there exist unrelated domains which have different distributions or label assignments, an inevitable problem named negative transfer will happen which leads to degenerated performance. Also, how to avoid selecting unconcerned samples to query is still an open question. To tackle these issues, we propose a hybrid algorithm for active learning with the help of transfer learning by adopting a divergence measure to measure the similarities between different domains, so that the negative effects can be alleviated. To avoid querying irrelevant instances, we also present an adaptive strategy that is able to eliminate unnecessary instances in the input space and models in the model space. Extensive experiments on both synthetic and real data sets show that our algorithm is able to query less instances and converges faster than the state-of-the-art methods.

This chapter relates issues of comprehensible input and comprehensible output to an increasingly prominent field: second-language pragmatics, where the intended meanings often go beyond the literal ones. The chapter will take a close look at what comprehensibility of language at the level of pragmatics actually entails. In looking at both the comprehension and production of pragmatic material, we will consider briefly the negotiation of meaning and conversational repairs—modifications made to the interactional structure of discourse and to words, sounds, and syntax for the purpose of communicating pragmatics. We will first look at what comprehensible input means with regard to pragmatics, whether through language (e.g. lexical items—words and phrases, syntax—e.g. verb tenses, or discourse), through gestures, or through silence. Then we will consider what comprehensible output entails in order that the addressees interpret the intended pragmatics correctly. Finally, we will reflect on the implications for both the learning and teaching of L2 pragmatics in lights of these comprehensibility issues.

Transfer learning techniques have witnessed a significant development in real applications where the knowledge from previous tasks are required to reduce the high cost of inquiring the labeled information for the target task. However, how to avoid negative transfer which happens due to different distributions of tasks in heterogeneous environment is still a open problem. In order to handle this kind of issue, we propose a Compact Coding method for Hyperplane Classifiers (CCHC) under a two-level framework in inductive transfer learning setting. Unlike traditional methods, we measure the similarities among tasks from the macro level perspective through minimum encoding. Particularly speaking, the degree of the similarity is represented by the relevant code length of the class boundary of each source task with respect to the target task. In addition, informative parts of the source tasks are adaptively selected in the micro level viewpoint to make the choice of the specific source task more accurate. Extensive experiments show the effectiveness of our algorithm in terms of the classification accuracy in both UCI and text data sets.

We propose a novel unsupervised transfer learning framework that utilises unlabelled auxiliary data to quantify and select the most relevant transferrable knowledge for recognising a target object class from the background given very limited training target samples. Unlike existing transfer learning techniques, our method does not assume that auxiliary data are labelled, nor the relationships between target and auxiliary classes are known a priori. Our unsupervised transfer learning is formulated by a novel kernel adaptation transfer (KAT) learning framework, which aims to (a) extract general knowledge about how more structured objects are visually distinctive from cluttered background regardless object class, and (b) more importantly, perform selective transfer of knowledge extracted from the auxiliary data to minimise negative knowledge transfer suffered by existing methods. The effectiveness and efficiency of the proposed approach is demonstrated by performing one-class object recognition (object vs. background) task using the Caltech256 dataset.

Domain adaptation (DA) using labeled data from related source domains comes in handy when the labeled patterns of a target domain are scarce. Nevertheless, it is worth noting that when the predictive distribution P(y|x) of the domains differs, which establishes Negative Transfer [19], DA approaches generally fail to perform well. Taking this cue, the Predictive Distribution Matching SVM (PDM-SVM) is proposed to learn a robust classifier in the target domain (referred to as the target classifier) by leveraging the labeled data from only the relevant regions of multiple sources. In particular, a k-nearest neighbor graph is iteratively constructed to identify the regions of relevant source labeled data where the predictive distribution maximally aligns with that of the target data. Predictive distribution matching regularization is then introduced to leverage these relevant source labeled data for training the target classifier. In addition, progressive transduction is adopted to infer the label of target unlabeled data for estimating the predictive distribution of the target domain. Finally, extensive experiments are conducted to illustrate the impact of Negative Transfer on several existing state-of-the-art DA methods, and demonstrate the improved performance efficacy of our proposed PDM-SVM on the commonly used multi-domain Sentiment and Reuters datasets.

This research addresses how an Intelligent Language Tutoring System can effectively help to solve a practical problem of transfer in students' second language learning of Chinese. Our empirical data shows that the problem of transfer accounts for most of the errors observed in the linguistic output of English-speaking students in their study of Chinese. This accords with views of other experts on transfer in the field of second language learning, such as Selinker [15], Cornu [6], Sheen [16] and Cowan [7]. A technique of mixed grammar of Chinese and English is used to tackle the problem of transfer. In this paper, we describe the importance of transfer, explain the data that has been collected, present an overview of the three main models, demonstrate the technique that we use to handle errors of transfer and finally discuss how our system is going to be evaluated.

The literature on learning in relation to age is heterogeneous and does not appear to have developed systematically. In this chapter, an effort is made to examine the themes or major purposes of the individual studies and to organize them into a relatively few categories. It must be recognized, however, that the categories are not mutually exclusive and that several studies could easily have been placed in more than one category. While the review of the literature is not all-inclusive, it is sufficiently comprehensive to represent the scope of interest in a particular type of problem. Except as indicated, these studies did not seem to control for possible age differences in the non-cognitive factors in learning. This lack of control has been one of the major problems in the field. Studies specifically relevant to the issue of non-cognitive contribution to age differences in learnings, were discussed in Chapter 3.

One of the most striking manifestations of rigidity among psychologists is their persistent clinging to discredited concepts, such as “rigidity.” Consideration of such a construct is likely to generate a discussion of syntactics rather than semantics: a study of the language habits of people who use the concept, rather than a study of events to which the concept refers. One of the most imaginative and clear-thinking contributors to this topic, Raymond Cattell, has called attention to “the confusion that will persist so long as some psychologists fail to recognize that in using the same term ‘rigidity’ they are assuming a single characteristic or process where, in fact, there are several.” (Cattell and Tiner, 1949, p. 321).