Home

Deep Discussion Group

Here we will post the list of papers discussed during each session.

Future Sessions

Session 14. 9 gen 2017

[Equilibrium Propagation: Bridging the Gap Between Energy-Based Models and Backpropagation](Equilibrium Propagation: Bridging the Gap Between Energy-Based Models and Backpropagation)

Session 12. 21 dec 2016

• Finding Tiny Faces [ @priba ]

Session 11. 19 dec 2016

• Tracking the World State with Recurrent Entity Networks [ @prlz77 ]

Past Sessions

Session 10. 15 dec 2016

• Feature Pyramid Networks for Object Detection [ @prlz77 ]

Session 9. 27 nov 2016

• CliqueCNN: Deep Unsupervised Exemplar Learning [ @edgarriba ]

Session 8. 25 nov 2016

• UNSUPERVISED REPRESENTATION LEARNING WITH DEEP CONVOLUTIONAL GENERATIVE ADVERSARIAL NETWORKS [ @prlz77 ]

Session 7. 9 nov 2016

• Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer [ @edgarriba ]

Session 6. 7 nov 2016

• Semantic Image Segmentation with Deep Convolutional Nets and Fully conected CRFs [ @gitabcworld ]

Session 5. 3 nov 2016

• Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering [@priba]
• Fully Convolutional Networks for Semantic Segmentation [@edgarriba]

Session 4. 28 oct 2016

• Fully Convolutional Networks for Semantic Segmentation [@edgarriba]

Session 3. 24 oct 2016

• R-FCN [@prlz77]

Session 2. 19 oct 2016

• FAIR team (pos 6 detection, pos 4 segmentation) A MultiPath Network for Object Detection [ @gitabcworld ]
• Slides ECCV MSRA Slides ECCV MSRA [ @prlz77 ]
• Keypoints ECCV Multi-Person Pose Estimation with Part Affinity Field [ @priba ]

Session 1. 17 oct 2016

MSCOCO 2016 Results, Leaderboard

• Base of most solutions Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks [ @edgarriba ]

Possible Papers

• G-RMI Object Detection [ @priba ]
• Conditional Random Fields as Recurrend Neural Networks

Especial de nadal?

Towards biologically plausible deep learning