Divergence measures based on the Shannon entropy

By Jianhua Lin. 1991

How researchers see this paper

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

In Information Theory

Citations in the corpus

Women also Snowboard: Overcoming Bias in Captioning Models (2018, methodology): “To formalize this notion of fairness in our captioning systems, we report the outcome type divergence between genders by measuring the Jensen-Shannon [47] divergence between Correct/Incorrect/Other outcomes for Men and Women.”
Adversarially Learned Inference (2016, methodology): “This can be shown using the same proof sketch as in the original GAN paper (Goodfellow et al. , 2014).”
Hacking smart machines with smarter ones: How to extract meaningful data from machine learning classifiers (2013, methodology): “this step can be simply assimilated into the training phase of the meta-classiﬁer. In other cases, as the example in Section 3.1, we will discuss a ﬁlter realized with the Kullback-Leibler divergence [43]. 3 Case studies In this section we provide two examples of attacks performed according with the”
Learning to select data for transfer learning with Bayesian Optimization (2017, background): “Our features are: • Jensen-Shannon divergence (Lin, 1991): Jensen-Shannon divergence is a smoothed, symmetric variant of D KL that has been successfully used for domain adaptation (Plank and van Noord, 2011; Remus, 2012).”
A Divergence Minimization Perspective on Imitation Learning Methods (2019, methodology): “We presented f -MAX, a generalization of AIRL [1] based on f -divergence [14].”
Covert Communication Over Noisy Channels: A Resolvability Perspective (2015, methodology): “The performance of the test may be captured by computing the Jensen-Shannon divergence [32]”
MSˆ2: Multi-Document Summarization of Medical Studies (2021, methodology): “If we consider this distribution for the gold summary (Pij) and the generated summary (Qij), we can compute the Jensen-Shannon Distance (JSD) (Lin, 1991),”
Kernel-Based Object Tracking (2003, background): “The Bhattacharyya coefficient is a divergence-type measure [49] which has a straightforward geometric interpretation .”
On the parameter space of generative lexicalized statistical parsing models (2004, background): “The era of modern statistical parsing did not truly begin until many years after Booth and Thompson’s work, with the work of Magerman and Marcus (1991) and Black, Lafferty, and Roukos (1992). Magerman and Marcus (1991) developed a probabilistic, k-best chart parser called Pearl. This model attempted”
Learning to model relatedness for news recommendation (2011, methodology): “Due to the zero-probability problem, we use the Jensen-Shannon divergence [19]:”
The Case for Evaluating Causal Models Using Interventional Measures and Empirical Data (2019, background): “A single TVD value is computed for each causal effect, which can then be aggregated for comparison.”
Alignment-free genome comparison with feature frequency profiles (FFP) and optimal resolutions (2009, unknown): “See Lin (15) for a further description.”
Similarity Measures for Text Document Clustering (2008, methodology): “As for the averaged Kullback-Leibler divergence, Lin gave a clear explanation in [11].”