Aplicaciones del deep learning

Procesamiento del lenguaje natural

Traducción automática

• Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation, September 2016, https://arxiv.org/abs/1609.08144
• Microsoft Presentation Translator, Microsoft, 2017, https://translator.microsoft.com/help/presentation-translator/
• Artificial intelligence goes bilingual—without a dictionary, Science, November 2017, http://www.sciencemag.org/news/2017/11/artificial-intelligence-goes-bilingual-without-dictionary
• Speak easy: How neural networks are transforming the world of translation, IT Pro, August 2017, http://www.itpro.co.uk/desktop-software/29155/speak-easy-how-neural-networks-are-transforming-the-world-of-translation

Reconocimiento de voz

• Lip Reading Sentences in the Wild, arXiv, January 2017 (v2), https://arxiv.org/abs/1611.05358
• Deep clustering: Discriminative embeddings for segmentation and separation, arXiv, 2015, https://arxiv.org/abs/1508.04306 (An AI has learned how to pick a single voice out of a crowd, New Scientist, October 2017, https://www.newscientist.com/article/2151268-an-ai-has-learned-how-to-pick-a-single-voice-out-of-a-crowd/)

Bots

• Automated Crowdturfing Attacks and Defenses in Online Review Systems, ACM Conference on Computer and Communications Security, CCS'2017, https://acmccs.github.io/papers/p1143-yaoA.pdf
• DeepProbe: Information Directed Sequence Understanding and Chatbot Design via Recurrent Neural Networks, KDD'2017, http://www.kdd.org/kdd2017/papers/view/deepprobe-information-directed-sequence-understanding-and-chatbot-design-vi

Detección de emociones

• DeepMoji: https://deepmoji.mit.edu/ (An Algorithm Trained on Emoji Knows When You’re Being Sarcastic on Twitter, MIT Technology Review, August 2017, https://www.technologyreview.com/s/608387/an-algorithm-trained-on-emoji-knows-when-youre-being-sarcastic-on-twitter/)
• Emotion Recognition in Sound, Neuroinformatics 2017, https://link.springer.com/chapter/10.1007/978-3-319-66604-4_18
• DeepMood: Modeling Mobile Phone Typing Dynamics for Mood Detection, KDD'2017, http://www.kdd.org/kdd2017/papers/view/deepmood-modeling-mobile-phone-typing-dynamics-for-mood-detection

Vehículos autónomos

• NVIDIA DRIVE PX Scalable Supercomputer for Autonomous Driving, http://www.nvidia.com/object/drive-px.html

Visión artificial

• SharkSpotter: "Shark hunting AI-powered algorithm to begin flying over Aussie beaches", news.com.au, August 2017, http://www.news.com.au/technology/innovation/inventions/shark-hunting-aipowered-algorithm-to-begin-flying-over-aussie-beaches/news-story/10723640634388c2c760f6256bceb2d0
• Convolutional neural networks at the interface of physical and digital data, AIPR 2016, http://ieeexplore.ieee.org/abstract/document/8010606/ (pyCBIR: Machine learning and deep learning programs provide a helping hand to scientists analyzing images, Phys.org, November 2017, https://phys.org/news/2017-11-machine-deep-scientists-images.html)
• Automatic learning of gait signatures for people identification, arXiv, June 2016, https://arxiv.org/abs/1603.01006
• Algorithm reconstructs processes from individual images, Phys.org, October 2017, https://phys.org/news/2017-09-algorithm-reconstructs-individual-images.html

Problemas de seguridad

• Accessorize to a Crime: Real and Stealthy Attacks on State-of-the-Art Face Recognition, CCS'2016, https://dl.acm.org/citation.cfm?id=2978392
• Robust Physical-World Attacks on Deep Learning Models, arXiv, July 2017, https://arxiv.org/abs/1707.08945

Visión 3D

• Generative Query Networks: Neural scene representation and rendering, Science, June 2018. http://science.sciencemag.org/cgi/doi/10.1126/science.aar6170
• SurfNet: Generating 3D shape surfaces using deep residual networks, arXiv, March 2017. https://arxiv.org/abs/1703.04079
• Face Normals “in-the-wild” using Fully Convolutional Networks, CVPR'2017, http://openaccess.thecvf.com/content_cvpr_2017/html/Trigeorgis_Face_Normals_In-The-Wild_CVPR_2017_paper.html

Síntesis de imágenes

• Kernel-Predicting Convolutional Networks for Denoising Monte Carlo Renderings, SIGGRAPH 2017, http://cvc.ucsb.edu/graphics/Papers/SIGGRAPH2017_KPCN/ (cf. http://www.news.ucsb.edu/2017/018150/intelligent-animation)
• iDeepColor, Real-Time User-Guided Image Colorization with Learned Deep Priors, SIGGRAPH 2017, https://richzhang.github.io/ideepcolor/
• Colorful Image Colorization, ECCV 2016, http://richzhang.github.io/colorization/

Medicina

Diagnóstico de enfermedades

• AI vs. Doctors, IEEE Spectrum, September 2017. https://spectrum.ieee.org/static/ai-vs-doctors
• Self-taught artificial intelligence beats doctors at predicting heart attacks, Science News, April 2017. http://www.sciencemag.org/news/2017/04/self-taught-artificial-intelligence-beats-doctors-predicting-heart-attacks
• Deep Learning Parkinson's from Smartphone Data (CloudUPDRS), New Scientist, March 2017. https://www.newscientist.com/article/2123459-super-fast-parkinsons-app-will-track-symptoms-more-closely/
• Assisting Pathologists in Detecting Cancer with Deep Learning, Google Research, March 2017. https://research.googleblog.com/2017/03/assisting-pathologists-in-detecting.html
• Detecting diabetic eye disease with machine learning, Google, November 2016, https://www.blog.google/topics/machine-learning/detecting-diabetic-eye-disease-machine-learning/
• Stanford algorithm can diagnose pneumonia better than radiologists, Stanford News, November 2017, https://news.stanford.edu/2017/11/15/algorithm-outperforms-radiologists-diagnosing-pneumonia/ (CheXNet, https://arxiv.org/abs/1711.05225)
• PupilScreen aims to allow parents, coaches, medics to detect concussion, brain injuries with a smartphone, UW News, September 2017, http://www.washington.edu/news/2017/09/06/pupilscreen-aims-to-allow-parents-coaches-medics-to-detect-concussion-brain-injuries-with-a-smartphone/ (PupilScreen, https://ubicomplab.cs.washington.edu/publications/pupilscreen/)
• Stanford's AI Predicts Death for Better End-of-Life Care, IEEE Spectrum, January 2018, https://spectrum.ieee.org/the-human-os/biomedical/diagnostics/stanfords-ai-predicts-death-for-better-end-of-life-care
• DeepGestalt: "Artificial intelligence could diagnose rare disorders using just a photo of a face," Science, January 2019. https://www.sciencemag.org/news/2019/01/artificial-intelligence-could-diagnose-rare-disorders-using-just-photo-face

Genética

• DeepSEA, http://deepsea.princeton.edu/job/analysis/create/ & Predicting effects of noncoding variants with deep learning–based sequence model, Nature Methods, August 2015, http://www.nature.com/nmeth/journal/v12/n10/full/nmeth.3547.html
• Genome-wide prediction and functional characterization of the genetic basis of autism spectrum disorder, Nature Neuroscience, August 2016 http://www.nature.com/neuro/journal/v19/n11/full/nn.4353.html

Neurociencia

• Artificial intelligence is learning to read your mind—and display what it sees, Science, October 2017, http://www.sciencemag.org/news/2017/10/artificial-intelligence-learning-read-your-mind-and-display-what-it-sees
• What artificial brains can teach us about how our real brains learn, Science, September 2017, http://www.sciencemag.org/news/2017/09/what-artificial-brains-can-teach-us-about-how-our-real-brains-learn

Otras aplicaciones médicas

• Using machine learning to improve patient care, MIT News, August 2017, http://news.mit.edu/2017/using-machine-learning-improve-patient-care-0821
• A deep convolutional neural network for classification of red blood cells in sickle cell anemia. PLoS Comput Biol 13(10): e1005746. doi.org/10.1371/journal.pcbi.1005746 (New machine learning system can automatically identify shapes of red blood cells, Phys.org, October 2017, https://phys.org/news/2017-10-machine-automatically-red-blood-cells.html)
• Prospective identification of hematopoietic lineage choice by deep learning. Nature Methods, February 2017, DOI: 10.1038/nmeth.4182 (Deep learning predicts hematopoietic stem cell development, Phys.org, September 2017, https://phys.org/news/2017-02-deep-hematopoietic-stem-cell.html)
• Reconstructing cell cycle and disease progression using deep learning, Nature Communications (2017). September 2017, DOI 10.1038/s41467-017-00623-3 (Algorithm reconstructs processes from individual images, Phys.org, October 2017, https://phys.org/news/2017-09-algorithm-reconstructs-individual-images.html)

Videojuegos

• Deep Q Network [DQN] @ Human-level control through deep reinforcement learning, Nature, February 2015. https://www.nature.com/nature/journal/v518/n7540/abs/nature14236.html
• Policy Distillation, arXiv:1511.06295, November 2015 (v1) - January 2016(v2). https://arxiv.org/abs/1511.06295
• Actor-mimic: Deep multitask and transfer reinforcement learning, arXiv:1511.06342, November 2015 (v1) - February 2016 (v4), ICLR'2016, https://arxiv.org/abs/1511.06342
• Elastic Weight Consolidation @ Overcoming catastrophic forgetting in neural networks, PNAS, March 2017. http://www.pnas.org/content/114/13/3521.full.pdf

Programación automática

• "DeepCoder: Learning to Write Programs", arXiv, 2016 (v1) - 2017 (v2), https://arxiv.org/abs/1611.01989
• DeepFix: Fixing Common C Language Errors by Deep Learning, AAAI'2017, https://aaai.org/ocs/index.php/AAAI/AAAI17/paper/view/14603
• Differentiable Programs with Neural Libraries, arXiv, 2016(v1) - 2017 (v2), https://arxiv.org/abs/1611.02109

Seguridad

• PassGAN: A Deep Learning Approach for Password Guessing, arXiv, September 2017, https://arxiv.org/abs/1709.00440 & Artificial intelligence just made guessing your password a whole lot easier, Science News, September 2017. http://www.sciencemag.org/news/2017/09/artificial-intelligence-just-made-guessing-your-password-whole-lot-easier
• Fast, Lean, and Accurate: Modeling Password Guessability Using Neural Networks, 25th USENIX Security Symposium, August 2016, https://www.usenix.org/conference/usenixsecurity16/technical-sessions/presentation/melicher
• Deep Models Under the GAN: Information Leakage from Collaborative Deep Learning, ACM Conference on Computer and Communications Security, CCS'2017, https://acmccs.github.io/papers/p603-hitajA.pdf
• Automated Crowdturfing Attacks and Defenses in Online Review Systems, ACM Conference on Computer and Communications Security, CCS'2017, https://acmccs.github.io/papers/p1143-yaoA.pdf
• DeepLog: Anomaly Detection and Diagnosis from System Logs through Deep Learning, ACM Conference on Computer and Communications Security, CCS'2017, https://acmccs.github.io/papers/p1285-duA.pdf
• Neural Network-based Graph Embedding for Cross-Platform Binary Code Similarity Detection, ACM Conference on Computer and Communications Security, CCS'2017, https://acmccs.github.io/papers/p363-xuAemb.pdf

Música

• NSynth: Neural Audio Synthesis, Google, 2017. https://magenta.tensorflow.org/nsynth
• Magenta: How Google is making music with artificial intelligence, Science News, August 2017. http://www.sciencemag.org/news/2017/08/how-google-making-music-artificial-intelligence

Más aplicaciones científicas

• "Found in Translation": Predicting Outcomes of Complex Organic Chemistry Reactions using Neural Sequence-to-Sequence Models, NIPS'2017. https://arxiv.org/abs/1711.04810
• DeepSD [Deep learning-based statistical downscaling]: Generating High Resolution Climate Change Projections through Single Image Super-Resolution, KDD'2017, http://www.kdd.org/kdd2017/papers/view/deepsd-generating-high-resolution-climate-change-projections-through-single
• RNNoise: Learning Noise Suppression, https://people.xiph.org/~jm/demo/rnnoise/ & arXiv, 2017 https://arxiv.org/pdf/1709.08243.pdf
• NSERC Science Use Cases, http://www.nersc.gov/users/data-analytics/data-analytics-2/deep-learning/
• A look at Deep Learning for Science, O'Reilly, April 2017, https://www.oreilly.com/ideas/a-look-at-deep-learning-for-science

Tutoriales y casos prácticos

• 6 Deep Learning Applications a beginner can build in minutes: https://www.analyticsvidhya.com/blog/2017/02/6-deep-learning-applications-beginner-python
  • Photo colorization (Algorithmia API)
  • ChatBot (Watson API)
  • News aggregator based on sentiment (Aylien API)
  • Grammatical corrector (TensorFlow)
  • Change of sex in photographs (TensorFlow)
  • Reinforcement learning to play FlappyBird (TensorFlow & PyGame)
• Neural Art (image style transfer)
  • Keras, https://www.analyticsvidhya.com/blog/2016/11/creating-an-artificial-artist-color-your-photos-using-neural-networks/
• LSTMs for Human Activity Recognition, https://github.com/guillaume-chevalier/LSTM-Human-Activity-Recognition