Entropy and Margin Maximization for Structured Output Learning

By Patrick A. Pletscher, Cheng Soon Ong, Joachim M. Buhmann. 2010

47 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Optimization & Training, Statistical Learning Theory

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Optimization & Training

In Statistical Learning Theory

Citations in the corpus

On Structured Prediction Theory with Calibrated Convex Surrogate Losses (2017, neutral): “, in conditional random fields [25], Φlog(f ,y) := log( ∑ ŷ∈Y exp fŷ)− fy, and their hybrids [38, 21, 22, 41].”
Cross-Modal Supervision for Learning Active Speaker Detection in Video (2016, methodology): “We treat this as a structured output prediction problem [26].”
On the calibration of sensor arrays for pattern recognition using the minimal number of experiments (2014, background): “For example, Adaboost, Support Vector Machines, Neural Networks, Least Squares Support Vector Machines or Logistic Regression can be formulated as Large Margin algorithms [62,63] and Support Vector Machines are related to probabilistic exponential models [64,65,39].”
SEARNN: Training RNNs with Global-Local Losses (2017, unknown): “A slight modiﬁcation allows us to exploit this information more meaningfully: we add information about the full costs in the exponential, following e.g. Pletscher et al. (2010).”
Long-tail learning via logit adjustment (2020, background): “Losses of the general form (11) have also been explored for structured prediction (Pletscher et al., 2010).”
Deep Structured Prediction with Nonlinear Output Transformations (2018, background): “Several works studied the learning problem when inference is hard [13, 26, 40, 17, 33, 23, 43, 35], designing effective approximations.”
Advanced Structured Prediction (2014, methodology): “We follow the approach of Pletscher et al. (2010) in introducing maxmargin methods as it facilitates comparison with Herding.”
Weakly Supervised Object Detection with Posterior Regularization (2014, methodology): “In order to alleviate this restriction, we replace the “hard” max in (4) with the “soft” max as in [12, 23]: ls(w,x ,y) = log ∑”
A Constrained Latent Variable Model for Coreference Resolution (2013, background): “(7) approaches a Kronecker delta function that puts probability 1 on the max-scoring mention j = arg max 0 ≤ k<i w · φ ( i, j ) (assuming no ties), and 0 everywhere else (Pletscher et al., 2010; Samdani et al., 2012).”
A Primal-Dual Message-Passing Algorithm for Approximated Large Scale Structured Prediction (2010, methodology): “Independently, [4, 15] described the connection between structured SVMs loss and CRFs loss.”
Efficient Structured Prediction with Latent Variables for General Graphical Models (2012, methodology): “To control the variance of the log-linear probability model we follow (Hazan & Urtasun, 2010; Pletscher et al., 2010) and introduce a temperature parameter (cid:15) , i.e ., Importantly, (cid:15) deﬁnes an entire family of structured prediction tasks with latent variables.”