Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 5 of 5 for “"Label complexity"”.

  1. Active learning under the Bernstein condition for general losses

    … loss exponential improvement is achieved in label complexity over passive learning.

    uvic Repository record for Active learning under the Bernstein condition for general losses (opens in a new tab)

  2. OBSERVING DYNAMIC EMOTIONAL FACES FROM MANY GENERATIONS

    … age differences in emotion perception is a label selection task that presents static images of actors expressing emotion. More recently, dynamic stimuli have been developed; they portray realistic facial expressions that are more comparable to what one sees in real life. The use of dynamic …

    wku-diss Repository record for OBSERVING DYNAMIC EMOTIONAL FACES FROM MANY GENERATIONS (opens in a new tab)

  3. SEQUENTIAL DECISION MAKING FOR ACTIVE LEARNING AND INFERENCE IN ONLINE SETTINGS

    … settings where the hypothesis space is large and labeled data are expensive. Examples include unusual activities in surveillance feedings, target search among large areas, frauds in financial transactions, attacks and intrusions in communication and computer networks, anomalies in infrastructures …

    cornell Repository record for SEQUENTIAL DECISION MAKING FOR ACTIVE LEARNING AND INFERENCE IN ONLINE SETTINGS (opens in a new tab)

  4. Stochastic Optimization and Learning: An Adaptive and Resource-Efficient Approach

    … incurs a bounded regret and has an order-optimal label complexity. We also study the problem of noisy group testing under unknown noise models, which is contrast to the existing studies that assume perfect knowledge of probabilistic model of the noise. We propose a novel algorithm that is agnostic …

    cornell Repository record for Stochastic Optimization and Learning: An Adaptive and Resource-Efficient Approach (opens in a new tab)

  5. Online and active learning of big networks: theory and algorithms

    … data-driven applications. On the other hand, the labels of the nodes in big networks are scarce. It is urgent to optimize the process by which the labels are collected, because it is unrealistic to get labels of every node. The objective of my research is to develop algorithms for big network …

    uiuc Repository record for Online and active learning of big networks: theory and algorithms (opens in a new tab)