Decentralized load balancing for highly irregular search problems

In this paper, we present a distributed computing framework for problems characterized by a highly irregular search tree, whereby no reliable workload prediction is available. The framework is based on a peer-to-peer computing environment and dynamic load balancing. The system allows for dynamic resource aggregation, does not depend on any specific meta-computing middleware and is suitable for large-scale, multi-domain, heterogeneous environments, such as computational Grids. Dynamic load balancing policies based on global statistics are known to provide optimal load balancing performance, while randomized techniques provide high scalability. The proposed method combines both advantages and adopts distributed job-pools and a randomized polling technique. The framework has been successfully adopted in a parallel search algorithm for subgraph mining and evaluated on a molecular compounds dataset. The parallel application has shown good calability and close-to linear speedup in a distributed network of workstations.

Processing of regular and irregular past tense morphology in highly proficient second language learners of English: a self-paced reading study

Dual-system models suggest that English past tense morphology involves two processing routes: rule application for regular verbs and memory retrieval for irregular verbs (Pinker, 1999). In second language (L2) processing research, Ullman (2001a) suggested that both verb types are retrieved from memory, but more recently Clahsen and Felser (2006) and Ullman (2004) argued that past tense rule application can be automatised with experience by L2 learners. To address this controversy, we tested highly proficient Greek-English learners with naturalistic or classroom L2 exposure compared to native English speakers in a self-paced reading task involving past tense forms embedded in plausible sentences. Our results suggest that, irrespective to the type of exposure, proficient L2 learners of extended L2 exposure apply rule-based processing.

Lesions impairing regular versus irregular past tense production

We investigated selective impairments in the production of regular and irregular past tense by examining language performance and lesion sites in a sample of twelve stroke patients. A disadvantage in regular past tense production was observed in six patients when phonological complexity was greater for regular than irregular verbs, and in three patients when phonological complexity was closely matched across regularity. These deficits were not consistently related to grammatical difficulties or phonological errors but were consistently related to lesion site. All six patients with a regular past tense disadvantage had damage to the left ventral pars opercularis (in the inferior frontal cortex), an area associated with articulatory sequencing in prior functional imaging studies. In addition, those that maintained a disadvantage for regular verbs when phonological complexity was controlled had damage to the left ventral supramarginal gyrus (in the inferior parietal lobe), an area associated with phonological short-term memory. When these frontal and parietal regions were spared in patients who had damage to subcortical (n = 2) or posterior temporo-parietal regions (n = 3), past tense production was relatively unimpaired for both regular and irregular forms. The remaining (12th) patient was impaired in producing regular past tense but was significantly less accurate when producing irregular past tense. This patient had frontal, parietal, subcortical and posterior temporo-parietal damage, but was distinguished from the other patients by damage to the left anterior temporal cortex, an area associated with semantic processing. We consider how our lesion site and behavioral observations have implications for theoretical accounts of past tense production.