Probabilistic Methods for Deep Learning
Deadline for manuscript submissions: closed (1 October 2021) | Viewed by 30302
Interests: statistical machine learning; probabilistic modeling; Bayesian statistics; deep learning; generative models; approximate inference
The umbrella of techniques known as deep learning has had empirical success across a variety of predictive modeling tasks. Consequently, there is hope that deep learning can catalyze progress in medicine, the sciences, and other domains of consequence. Yet, many deep learning techniques are ill-equipped for these new settings in which safety and transparency are crucial for their success. For instance, neural networks have been shown to be overconfident, which could lead to them being unduly trusted to make a medical diagnosis.
Combining deep learning with probabilistic and statistical methodologies is one potential way to overcome—or at least ameliorate—these shortcomings. A probabilistic approach can quantify a network’s uncertainty, allowing for more informed down-stream decision making. Of course, this is a non-trivial pursuit, as deep learning incurs computational and analytical difficulties that do not plague more traditional models. Adapting deep learning so that its robustness and uncertainty can be quantified without sacrificing predictive power is an open and challenging problem.
In this Special Issue, we aim to highlight work at the intersection of deep learning, probabilistic modeling, and statistical inference. In particular, we welcome work on Bayesian neural networks, deep latent variable models, deep ensembles, networks with statistical guarantees (e.g., via conformal inference), and probabilistic understanding of neural networks (e.g., via infinite limits).
Dr. Eric Nalisnick
Dr. Dustin Tran
Manuscript Submission Information
Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.
Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Entropy is an international peer-reviewed open access monthly journal published by MDPI.
Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.
- deep learning
- neural networks
- probabilistic modeling
- Bayesian statistics
- statistical inference
- uncertainty quantification