دورية أكاديمية

On the Difference between the Information Bottleneck and the Deep Information Bottleneck

التفاصيل البيبلوغرافية
العنوان: On the Difference between the Information Bottleneck and the Deep Information Bottleneck
المؤلفون: Aleksander Wieczorek, Volker Roth
المصدر: Entropy, Vol 22, Iss 2, p 131 (2020)
بيانات النشر: MDPI AG, 2020.
سنة النشر: 2020
المجموعة: LCC:Science
LCC:Astrophysics
LCC:Physics
مصطلحات موضوعية: information bottleneck, markov assumption, markov chain, deep variational information bottleneck, conditional independence, mutual information, Science, Astrophysics, QB460-466, Physics, QC1-999
الوصف: Combining the information bottleneck model with deep learning by replacing mutual information terms with deep neural nets has proven successful in areas ranging from generative modelling to interpreting deep neural networks. In this paper, we revisit the deep variational information bottleneck and the assumptions needed for its derivation. The two assumed properties of the data, X and Y, and their latent representation T, take the form of two Markov chains T − X − Y and X − T − Y . Requiring both to hold during the optimisation process can be limiting for the set of potential joint distributions P ( X , Y , T ) . We, therefore, show how to circumvent this limitation by optimising a lower bound for the mutual information between T and Y: I ( T ; Y ) , for which only the latter Markov chain has to be satisfied. The mutual information I ( T ; Y ) can be split into two non-negative parts. The first part is the lower bound for I ( T ; Y ) , which is optimised in deep variational information bottleneck (DVIB) and cognate models in practice. The second part consists of two terms that measure how much the former requirement T − X − Y is violated. Finally, we propose interpreting the family of information bottleneck models as directed graphical models, and show that in this framework, the original and deep information bottlenecks are special cases of a fundamental IB model.
نوع الوثيقة: article
وصف الملف: electronic resource
اللغة: English
تدمد: 1099-4300
Relation: https://www.mdpi.com/1099-4300/22/2/131; https://doaj.org/toc/1099-4300
DOI: 10.3390/e22020131
URL الوصول: https://doaj.org/article/2591f2272b754ab79d7f7a3b35fd8e09
رقم الأكسشن: edsdoj.2591f2272b754ab79d7f7a3b35fd8e09
قاعدة البيانات: Directory of Open Access Journals
الوصف
تدمد:10994300
DOI:10.3390/e22020131