We are pleased to announce the ICLR-2021 Workshop: Energy Based Models: Current Perspectives, Challenges, and Opportunities, to be held virtually on May 8, 2021
Workshop page (please consult for up-to-date information): https://sites.google.com/view/ebm-workshop-iclr2021 ABOUT THE WORKSHOP: Energy-Based Models (EBMs) are a learning framework that assigns a quality score to any given input, its energy; contrary to probabilistic models, there is no a priori requirement that these scores be normalized (i.e. sum to one). Energies are typically computed through a neural network, and training an EBM corresponds to shaping the energy function such that data points nearby the underlying data manifold are associated with lower energies than data points that are far from it. Not imposing normalization affords a great power and flexibility to the modelling process, e.g. in terms of combining energies, on conditioning on certain variables, of computing global scores on complex structured objects, or on expressing prior knowledge. However, this freedom comes with significant technical challenges, in terms of learning and inference. A strong comeback of EBMs is currently underway. The goals of this Workshop are: § To provide a forum for discussing cutting-edge research on different EBM approaches (e.g for continuous vs. discrete problems, strategies to shape the energy function, etc.), and to position these in the broader ML landscape (MCMC, GANs, RL, Variational Inference, etc.) § To bring together different communities, not only from core ML, but also from application domains such as vision, nlp, biology, neuroscience, etc. § To learn and exchange ideas about current and potential applications of EBMs, and to provide insights which may inspire novel EBM developments. SUBMISSION GUIDELINES: We welcome submissions on (non exhaustive list): § Techniques for high-dimensional continuous EBMs (e.g. in Vision, Neuroscience) and for discrete EBMs (e.g. in NLP, Biology). § Probabilistic (e.g. advanced MCMC and variational inference) and non-probabilistic (e.g. contrastive learning) approaches for training EBMs. § Comparisons, connections and combinations with other ML approaches and frameworks (Autoregressive models, GANs, Variational Auto-Encoders, Reinforcement Learning, ... ). § Multidisciplinary applications of EBMs to different domains. § Applications of EBMs to unsupervised, self-supervised and latent models. § Emerging models of computation exploiting Energy characteristics in physical processes. § Opinion pieces in favor or against EBMs. Submissions will be either short papers (between 4 and 6 pages, excluding references) or extended abstracts (2 pages, excluding references). We welcome submissions that are work in progress, as well as submissions that are currently under review at other venues, as long as that fit the format above. Submissions should follow ICLR2021 submission format (LaTeX, overleaf). Submissions will be evaluated by a Programme Committee to be announced later. Accepted papers will be presented during poster sessions, with a few contributed talks. They will be posted on the workshop site, under a non-archival status. Submission A link for submitting will be provided soon on the workshop page. The review process will be double-blind. Authors should make sure to anonymise their submissions by removing names and affiliations. Available grants The workshop will provide ICLR registrations for authors of a few selected papers. TIMELINE: February 24, 2021 : Submission Deadline (11:59PM Anywhere on Earth) March 26, 2021: Notification of acceptance TBD: Camera-ready papers due TBD: Presentation recordings due May 8, 2021: Workshop INVITED SPEAKERS/PANELISTS: Benjamin Scellier, Mila, Université de Montréal Debora Marks, Harvard University Simon Osindero, DeepMind Stefano Ermon, Stanford Will Grathwohl, University of Toronto Yann Le Cun, Facebook AI Research Yingszhen Li, Microsoft Research ORGANIZERS: Adji Bousso Dieng, Google Brain Hady Elsahar, Naver Labs Europe Igor Mordatch, Google Brain Marc Dymetman, Naver Labs Europe Marc’Aurelio Ranzato, Facebook AI Research PROGRAMME COMMITTEE: To be announced
_______________________________________________ uai mailing list uai@engr.orst.edu https://it.engineering.oregonstate.edu/mailman/listinfo/uai
