.. image:: pystat.png
    :height: 20
    :alt: Statistique
    :target: http://www.xavierdupre.fr/app/ensae_teaching_cs/helpsphinx/td_2a_notions.html#pour-un-profil-plutot-data-scientist

Apprentissage sans labels
+++++++++++++++++++++++++

.. toctree::
    :maxdepth: 2

    ../specials/nolabel

*Notebooks*

*(à venir)*

*Lectures*

*Autoencoders - réduction de dimensionnalité*

* `Why Does Unsupervised Pre-training Help Deep Learning? <http://www.jmlr.org/papers/volume11/erhan10a/erhan10a.pdf>`_
* `Autoencoders <http://ufldl.stanford.edu/tutorial/unsupervised/Autoencoders/>`_
* `Autoencoders, Unsupervised Learning, and Deep Architectures <http://www.jmlr.org/proceedings/papers/v27/baldi12a/baldi12a.pdf>`_
* `Generative Models <https://openai.com/blog/generative-models/>`_,
  `Adversarial Autoencoders <http://hjweide.github.io/adversarial-autoencoders>`_
* `Tutorial on Variational Autoencoders <https://arxiv.org/abs/1606.05908>`_,
  `Denoising Autoencoders (dA) <http://deeplearning.net/tutorial/dA.html>`_
* `Generative Adversarial Networks <https://arxiv.org/abs/1406.2661>`_,
  `NIPS 2016 Tutorial: Generative Adversarial Networks <https://arxiv.org/abs/1701.00160>`_
* `Adversarial Autoencoders <https://arxiv.org/abs/1511.05644>`_
* `Adversarial Autoencoders (with Pytorch) <https://blog.paperspace.com/adversarial-autoencoders-with-pytorch/>`_
* `Marginalizing Stacked Linear Denoising Autoencoders <http://www.jmlr.org/papers/volume16/chen15c/chen15c.pdf>`_
* `What Regularized Auto-Encoders Learn from the Data-Generating Distribution <http://jmlr.csail.mit.edu/papers/volume15/alain14a/alain14a.pdf>`_
* `Compressed sensing and single-pixel cameras <https://terrytao.wordpress.com/2007/04/13/compressed-sensing-and-single-pixel-cameras/>`_
* `Multi-Label Prediction via Compressed Sensing <Multi-Label Prediction via Compressed Sensing>`_
* `Inference in generative models using the Wasserstein distance <https://arxiv.org/abs/1701.05146>`_,
  `Coupling of Particle Filters <https://arxiv.org/abs/1606.01156>`_
* `Auto-Encoding Variational Bayes <https://arxiv.org/pdf/1312.6114.pdf>`_

*No label, weak labels*

* `Unsupervised Supervised Learning I: Estimating Classification and Regression Errors without Labels <http://www.jmlr.org/papers/volume11/donmez10a/donmez10a.pdf>`_
* `Unsupervised Supervised Learning II: Margin-Based Classification without Labels <http://www.jmlr.org/proceedings/papers/v15/balasubramanian11a/balasubramanian11a.pdf>`_,
  `Unsupervised Supervised Learning II: Margin-Based Classification Without Labels <http://www.jmlr.org/papers/volume12/balasubramanian11a/balasubramanian11a.pdf>`_ (longer version)
* `Large-scale Multi-label Learning with Missing Labels <http://jmlr.org/proceedings/papers/v32/yu14.pdf>`_
* `Reducing Label Complexity by Learning From Bags <http://www.jmlr.org/proceedings/papers/v9/sabato10a/sabato10a.pdf>`_
* `Learning from Corrupted Binary Labels via Class-Probability Estimation <http://jmlr.org/proceedings/papers/v37/menon15.pdf>`_
* `Generalized Expectation Criteria for Semi-Supervised Learning with Weakly Labeled Data <http://www.jmlr.org/papers/volume11/mann10a/mann10a.pdf>`_
* `Multitask Learning without Label Correspondences <http://users.sussex.ac.uk/~nq28/pubs/Quaetal11.pdf>`_
* `Training Highly Multiclass Classifiers <http://jmlr.org/papers/volume15/gupta14a/gupta14a.pdf>`_

*Online training*

* `Online Incremental Feature Learning with Denoising Autoencoders <http://jmlr.csail.mit.edu/proceedings/papers/v22/zhou12b/zhou12b.pdf>`_
* `Fast Kernel Classifiers with Online and Active Learning <http://www.jmlr.org/papers/volume6/bordes05a/bordes05a.pdf>`_,
  `A Framework for Learning Predictive Structures from Multiple Tasks and Unlabeled Data <http://www.jmlr.org/papers/volume6/ando05a/ando05a.pdf>`_
* `Multi Kernel Learning with Online-Batch Optimization <http://www.jmlr.org/papers/volume13/orabona12a/orabona12a.pdf>`_

*Improving training set*

* `Data Programming: Creating Large Training Sets, Quickly <https://papers.nips.cc/paper/6523-data-programming-creating-large-training-sets-quickly.pdf>`_
* `Foolbox is a Python toolbox to create adversarial examples that fool neural networks. <https://foolbox.readthedocs.io/en/latest/>`_

*Adversarial Examples*

* `The Limitations of Deep Learning in Adversarial Settings <https://arxiv.org/pdf/1511.07528v1.pdf>`_ :
  l'article montre des limites de l'approche deep learning en construisant des exemples
  proches des exemples initiaux mais qui font dérailler le modèle.