Prof. Joshua Tenenbaum (CBMM Research Thrust Leader) and Tejas Kulkarni (CBMM Siemens Graduate Fellow) are helping to organize a workshop for NIPS 2015 Workshop on Black Box Learning and Inference, on December 12, 2015. Papers are being accepted now through October 2, 2015.
NIPS 2015 Workshop on Black Box Learning and Inference
December 12, 2015, Montreal, Canada
Submission deadline: October 2, 2015
1. Workshop Overview
Probabilistic models have traditionally co-evolved with tailored algorithms for efficient learning and inference. One of the exciting developments of recent years has been the resurgence of black box methods, which make relatively few assumptions about the model structure, allowing application to broader model families.
In probabilistic programming systems, black box methods have greatly improved the capabilities of inference back ends. Similarly, the design of connectionist models has been simplified by the development of black box frameworks for training arbitrary architectures. These innovations open up opportunities to design new classes of models that smoothly negotiate the transition from low-level features of the data to high-level structured representations that are interpretable and generalize well across examples.
This workshop brings together developers of black box inference technologies, probabilistic programming systems, and connectionist computing frameworks. The goal is to formulate a shared understanding of how black box methods can enable advances in the design of intelligent learning systems.
We invite contributions of abstracts on the following topics, to be presented as posters, contributed talks, or as short spotlight talks:
(a) Inference in probabilistic programming systems and broad model families:
- Variational inference
- Gradient-based methods for parameter estimation
- Metropolis-Hastings variants with efficient rescoring,
- Message passing variants,
- Sequential Monte Carlo variants,
(b) Models that integrate top-down and bottom-up model representations to perform amortized inference: variational autoencoders, deep latent Gaussian models, restricted Boltzmann machines, neural network based proposals in MCMC.
(c) Model specification languages that use black box techniques (probabilistic programming languages, neural network libraries such as Torch, Theano, Caffe),
(d) Applications to vision, speech, reinforcement learning, motor control, language learning.
Submitted abstracts should be 2–4 pages in NIPS format, sent to nips2015blackbox at gmail.com before October 2, 2015 (midnight PDT). Submissions need not be anonymized. Additional space beyond the fourth page is permitted for references, if required. Authors will be notified of acceptance and presentation type (poster, spotlight, talk) by November 2, 2015. Final versions of submitted abstracts are due December 2, 2015 and will be distributed on the workshop website.
2. Key Dates
Abstract submission: October 2, 2015
Acceptance notification: November 2, 2015
Final abstract submission: December 2, 2015
Workshop: December 12, 2015
3. Workshop Organizers
Jan-Willem van de Meent (Oxford)
Tejas Kulkarni (MIT)
Ali Eslami (Google DeepMind)
Brooks Paige (Oxford)
Joshua Tenenbaum (MIT)
Frank Wood (Oxford)
Zoubin Ghahramani (Cambridge)