NeurIPS 2015

NeurIPS 2015
Name	NeurIPS 2015
Venue	International Convention Center
Location	Montreal, Canada
Date	December 4–9, 2015
Organizers	Neural Information Processing Systems Foundation
Attendance	≈6,000
Proceedings	Advances in Neural Information Processing Systems 28

NeurIPS 2015

NeurIPS 2015 was the twenty-eighth annual meeting of the community around Neural Information Processing Systems Foundation, held in Montreal, Canada, drawing researchers from institutions such as Google DeepMind, Facebook AI Research, Microsoft Research, Stanford University, and Massachusetts Institute of Technology. The conference combined peer-reviewed presentations, keynote lectures, workshops, tutorials, poster sessions, and competitions, featuring work from teams at University of Toronto, Carnegie Mellon University, University of California, Berkeley, Princeton University, and industry labs like Amazon Research and IBM Research. The proceedings, published as Advances in Neural Information Processing Systems 28, captured advances in deep learning, Bayesian methods, reinforcement learning, and structured prediction.

Overview

The conference agenda reflected ongoing dialogues among communities represented by Geoffrey Hinton, Yoshua Bengio, Yann LeCun, Andrew Ng, and contributors from Google Brain, DeepMind, and academic groups at University of Montreal and École Polytechnique. Themes included connections between optimization research from Yann LeCun's lineage, probabilistic modeling influenced by David MacKay's school, and reinforcement learning developments tied to Richard Sutton and Andrew Barto's frameworks. Industry participation from Apple Inc., Intel Corporation, NVIDIA, and Uber AI Labs signaled shifts in resource support and deployment focus.

Conference Organization and Program

Organizers from the Neural Information Processing Systems Foundation coordinated program committees drawing on reviewers from NeurIPS Program Committee 2015 members representing University of Oxford, University of Cambridge, Columbia University, and Yale University. Sessions included oral presentations, poster sessions, and spotlights, with tracks covering supervised learning, unsupervised learning, graphical models, and reinforcement learning pioneered by groups at DeepMind and University College London. Tutorials led by faculty from University of Toronto and researchers from Google DeepMind and Facebook AI Research provided hands-on material aligned with software releases from Theano, TensorFlow, and Torch. Local arrangements involved partnerships with Montreal Institute for Learning Algorithms and municipal agencies in Montreal.

Keynote and Invited Speakers

Keynote and invited speaker slots featured prominent figures, including talks by researchers with affiliations to Google DeepMind, Facebook AI Research, Microsoft Research, and laboratories at Princeton University and MIT. Speakers drew intellectual lineages from Geoffrey Hinton, Yoshua Bengio, Yann LeCun, and recipients of awards from institutions such as the Association for Computing Machinery and the Royal Society. Panels juxtaposed perspectives from leaders at DeepMind and academics from University of Toronto and Carnegie Mellon University, and invited lectures referenced classical work by Judea Pearl and contemporary advances from Ian Goodfellow's research on generative adversarial networks.

Notable Papers and Workshops

The oral program included influential papers on deep architectures and optimization that built on prior results from Geoffrey Hinton's group and methods developed at Google Brain and Stanford University. Workshops covered topics such as generative modeling connected to Ian Goodfellow's GAN research, Bayesian deep learning related to Yarin Gal's contributions, and reinforcement learning extending frameworks from Richard Sutton and David Silver. Specialized workshops brought together experts from DeepMind, University College London, Oxford University, and Carnegie Mellon University to explore hierarchical reinforcement learning, probabilistic programming influenced by Frank Wood and Noah Goodman, and scalability issues aligned with work at NVIDIA and Intel Research.

Competitions and Demonstrations

Competitions at the conference showcased benchmarks and challenge results from teams at Stanford University, University of California, Berkeley, ETH Zurich, and industry labs including Google and Facebook. Demonstrations highlighted software and hardware ecosystems including frameworks such as Theano, TensorFlow, and Torch, and accelerator technologies from NVIDIA and Intel Nervana Systems. Challenge tasks emphasized image recognition inspired by datasets used in competitions associated with ImageNet and reinforcement learning tasks reflecting environments from OpenAI and simulators used by DeepMind.

Attendance, Diversity, and Community Impact

Attendance numbered in the thousands, with delegates from universities like University of Washington, Harvard University, Brown University, and research institutions including Lawrence Berkeley National Laboratory and Sandia National Laboratories. Community discussions addressed representation concerns involving scholars from University of Toronto, McGill University, and international institutions such as Tsinghua University and Peking University. Outreach events connected student groups from École Polytechnique and local meetup communities, while industry recruitment drew participation from Google, Facebook, Microsoft, and startups in the Montreal ecosystem.

Legacy and Influence on Machine Learning Research

NeurIPS 2015 contributed to the acceleration of research trajectories that favored deep learning paradigms, influencing follow-up work at Google DeepMind, OpenAI, Facebook AI Research, and academic groups at Stanford University and MIT. Insights presented at the conference informed subsequent advances in generative modeling, reinforcement learning algorithms, and scalable training techniques adopted by NVIDIA and cloud services from Amazon Web Services and Google Cloud Platform. The proceedings served as a reference point for researchers in the machine learning community, shaping curricula at institutions such as Carnegie Mellon University and University of California, Berkeley and guiding industrial roadmaps at Intel Corporation and Apple Inc..

Category:NeurIPS conferences