|
The 3rd Workshop on Tractable Probabilistic Modeling (TPM 2019) @ ICML, Long Beach, California.
https://sites.google.com/view/icmltpm2019/home
----------------------------------------------------------------
Important Dates
Paper submission deadline: May 10, 2019 AOE (UTC-12:00h)
Notification to authors: May 15, 2019
Camera ready version: May 31, 2019 AOE (UTC-12:00h)
Workshop Date: June 14/15, 2019 (TBA)
----------------------------------------------------------------
Tractable probabilistic modeling (TPM) is concerned with the inherent trade-off between the expressivity of the probabilistic models and the complexity of performing various types of inference on them, as well as learning them from data. Traditional topics in this area include efficient learning of probabilistic models, exact inference, and approximate routines with guarantees. Relevant model classes include low- and bounded-treewidth PGM, determinantal point processes, exchangeable probabilistic models, arithmetic circuits, sum-product networks, cutset networks, probabilistic sentential decision diagrams, and more. Successful real-world applications of such models comprise: image classification, completion and generation, scene understanding, activity recognition, language and speech modeling, bioinformatics, collaborative filtering, verification and diagnosis of physical systems.
This year's workshop will focus especially on bringing together researchers working on the different fronts and communities of TPM. We especially encourage submissions highlighting the challenges and opportunities for tractable inference and modeling within the rising field of probabilistic programming and the neural probabilistic modeling community, recently achieving impressive successes in many application fields.
Here is a non-exhaustive list of possible venues. Any other work relevant to the TPM community will be highly appreciated.
Tractable inference with neural probabilistic models
Challenges in tractable probabilistic programming
New tractable representations in discrete, continuous and hybrid domains
Tractable models and explainable AI
Learning algorithms for tractable probabilistic models
Theoretical and empirical analysis of tractable modeling
Approximate inference algorithms with guarantees on approximation quality
Applications of tractable probabilistic modeling
------------------------------------------------------------------------------------------
Submission Instructions:
Recently published research papers can be submitted as they were accepted.
Original papers (up to 8 pages, not including references) and abstracts (up to 2 pages, not including references) are required to follow the same style guidelines of ICML 2019
All submissions must be electronic, in the above format and submitted through easychair (link below).
Reviewing for TPM 2019 is single-blind and we also encourage authors to share code and data for reproducibility.
Submission Link: https://easychair.org/conferences/?conf=tpm2019
--------------------------------------------------------------------------------------------
Organizers:
Daniel Lowd (University of Oregon)
Tahrima Rahman (University of Texas, Dallas)
Antonio Vergari (Max-Planck-Insitute for Intelligent Systems, University of California, Los Angeles)
Alejandro Molina (TU Darmstadt)
Pedro Domingos (University of Washington)
|
