We present Partition-wise Gated Recurrent Units (pGRUs) for point-based trajectory classification to detect real-world trawler fishing activities in the ocean. We propose partition-wise activation functions integrated with Gated Recurrent Units that partitions each feature and uses independent parameters to model distinct regions of the feature space. This approach enables us to leverage the benefits of deep learning on a low-dimensional and heterogeneous feature space by mapping the low-dimensional features into another space that can be better separated with piece-wise learnable parameters. We show that the proposed partition-wise activation functions can approximate a wide array of functions, including conventional activations such as sigmoid and hyperbolic tangent. Our experimental results demonstrate that pGRU learns the probability distribution from data and achieves substantial improvements over the state-of-the-art systems on the task of trawler fishing activity detection.

Amachine lifelong learningsystem based ontask rehearsalandmultiple task learning(MTL) is used to sequentially learn a series of medical diagnostic tasks. The representations of successfully learned neural network models of the tasks are stored within a domain knowledge database.Virtual examplesgenerated from these models are relearned, orrehearsed, in parallel with each new task using theηMTL neural network algorithm, a variant of MTL. TheηMTL algorithm employs a separate learning rate,ηkfor each task output,k.ηkvaries as a function of the measure of relatedness between each prior taskkand the new task being learned. Working together, the task rehearsal method andηMTL are able to develop more accurate hypotheses for a new task by selectively transferring knowledge from related tasks in domain knowledge. Coronary artery disease data sets from three real and four fictitious hospitals provide a domain of related and unrelated tasks for testing the system. The experimental results demonstrate the method's ability to sequentially retain and transfer clinical diagnostic knowledge when learning from impoverished training sets.

Artificial neural networks and inductive transfer are used to develop models that predict the discharge (flow rate) of fresh water streams in Nova Scotia from weather data. The objective is to show that transfer can be used to reduce the time and cost associated with collecting large amounts of the data for environmental modeling. The models use two days of weather data to predict the discharge for the following day. The models can be applied to land use, water management and flood predictions for sections of streams where continuous monitoring is not feasible. Models developed using only 180 days of training data with transfer from related streams perform as well on independent test data as models constructed using five years of training data and no transfer.

Fundamental to the problem of lifelong machine learning is how to consolidate the knowledge of a learned task within a long-term memory structure (domain knowledge) without the loss of prior knowledge. We investigate the effect of curriculum, i.e., a selection of tasks and the order in which they are learned, on the consolidation of task knowledge. Relevant background material on knowledge transfer and consolidation using multiple task learning (MTL) neural networks is reviewed. A large MTL network is used as the long-term memory structure and task rehearsal overcomes the stability-plasticity problem and the loss of prior knowledge. Experimental results demonstrate that curriculum has a noticeable effect on the accuracy of consolidated knowledge particularly for the first few tasks that are learned. The results also suggest that, for given set of tasks and training examples, the mean accuracy of consolidated domain knowledge converges to the same level regardless of the curriculum. This is caused by the averaging effect of sequential consolidation.

With a distinction made between two forms of task knowledge transfer, 'representational' and 'functional', eta MTL, a modified version of the multiple task learning (MTL) method of functional (parallel) transfer, is introduced. The eta MTL method employs a separate learning rate, etak, for each task output node k. etak varies as a function of a measure of relatedness, Rk, between the kth task and the primary task of interest. Results of experiments demonstrate the ability of eta MTL to dynamically select the most related source task(s) for the functional transfer of prior domain knowledge. The eta MTL method of learning is nearly equivalent to standard MTL when all parallel tasks are sufficiently related to the primary task, and is similar to single task learning when none of the parallel tasks are related to the primary task.

The X.500 Directory Service provides a powerful mechanism for storing and retrieving information about objects in a distributed computing environment. This requires the functional components of the Directory Service to have knowledge of the structure and representation, or schema, of the information held within the directory. The management of the Directory Schema is a subject requiring further research and development. We identify the three major technical elements required for properly managing the Directory Schema. We then focus on one of these elements; propagation of the schema between functional components of the Directory Service. Three subproblems from within schema propagation are presented, along with several alternative solutions.>