3. DL TOPICS
Image recognition
Speech recognition
Video captioning
NLP word embedding, Translation
Biological model…
Neural Arts
Games
Self-Driving Cars
Deep Residual Learning
GPUs
4. NEURAL ART
A Neural Algorithm of Artistic Style
Leon A. Gatys, Alexander S. Ecker, Matthias Bethge
10. NEURAL ART – STYLE
GENERATION
Code by Mark Chang, National Taiwan University
https://github.com/ckmarkoh
11.
12. GAMES – DEEP Q-NETWORK (DQN)
https://deepmind.com/dqn.html
Deep Neural Network + Reinforcement Learning
how the game agents should act in an environment in order to
maximize future cumulative reward (e.g., a game score)
13. GAMES – DEEP Q-NETWORK (DQN)
Naveen Appiah, Sagar Vare, Stanford.
Playing FlappyBird with Deep Reinforcement Learning
Develop a CNN model to
learn features from
snapshots of the game
and train the agent to
take the right actions at
each game instance.
14. GAMES – DEEP Q-NETWORK (DQN)
Naveen Appiah, Sagar Vare, Stanford.
Playing FlappyBird with Deep Reinforcement Learning
Deep neural network to learn
game specific features from
raw pixels and decide on what
actions to take.
Instead, a reinforcement
learning setting tries to
evaluate the actions at a given
state based on the reward it
observes by executing it.
16. SELF-DRIVING CARS: NVIDIA
DRIVE PX
• Dual NVIDIA Tegra® X1 processors delivering a
combined 2.3 Teraflops
• Interfaces for up to 12 cameras, radar, lidar, and
ultrasonic sensors
• Rich middleware for graphics, computer vision, and
deep learning
23. DEEP RESIDUAL NETWORK
FOR IMAGE RECOGNITION
Microsoft Research:
• ImageNet computer vision challenge championship
3.57% error on the ImageNet test set.
• 152 layers of deep networks, 8 times deeper than VGG
• Image classification, detection and localization
Deep Residual Learning for Image Recognition http://arxiv.org/pdf/1512.03385v1.pdf
32. EASY LIFE
Everything is in Tensor
Softmax Function
https://www.tensorflow.org/versions/r0.7/tutorials/mnist/beginners/index.html#mnist-for-ml-beginners
x
b
W Matmul
Add
Softm
ax
34. A FEW TENSORFLOW COMMUNITY
EXAMPLES
● DQN: github.com/nivwusquorum/tensorflow-deepq
● NeuralArt: github.com/woodrush/neural-art-tf
● Char RNN: github.com/sherjilozair/char-rnn-tensorflow
● Keras ported to TensorFlow: github.com/fchollet/keras
● Show and Tell: github.com/jazzsaxmafia/show_and_tell.tensorflow
● Mandarin translation: github.com/jikexueyuanwiki/tensorflow-zh
40. TENSORFLOW
• GOOGLE OPEN TENSORFLOW IN LAST YEAR
• HTTP://TENSORFLOW.ORG/
• HTTPS://GITHUB.COM/TENSORFLOW/TENSORFLOW
• HTTP://DOWNLOAD.TENSORFLOW.ORG/PAPER/WHITEPAPER20
15.PDF
These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play.
This allows algorithms to accurately understand the full 360 degree environment around the car to produce a robust representation, including static and dynamic objects. Use of Deep Neural Networks (DNN) for the detection and classification of objects dramatically increases the accuracy of the resulting fused sensor data. - See more at: http://www.nvidia.com/object/drive-px.html#sthash.Vq3ZHH11.dpuf
This allows algorithms to accurately understand the full 360 degree environment around the car to produce a robust representation, including static and dynamic objects. Use of Deep Neural Networks (DNN) for the detection and classification of objects dramatically increases the accuracy of the resulting fused sensor data. - See more at: http://www.nvidia.com/object/drive-px.html#sthash.Vq3ZHH11.dpuf
Under circustanes
Under circustanes
150 Macbook Pro
The combination of these software capabilities running on Pascal-powered Tesla GPUs allows applications to run 12x faster than with previous GPU-accelerated solutions.