Deep neural networks have boosted the convergence of multimedia data analytics in a unified framework shared by practitioners in natural language, vision and speech. Image captioning, lip reading or video sonorization are some of the first applications of a new and exciting field of research exploiting the generalization properties of deep neural representation. This tutorial will firstly review the basic neural architectures to encode and decode vision, text and audio, to later review the those models that have successfully translated information across modalities. The contents of this tutorial are available at: https://telecombcn-dl.github.io/2019-mmm-tutorial/.

1. Multimodal Deep Learning
#MMM2019
Xavier Giro-i-Nieto
xavier.giro@upc.edu
Associate Professor
Intelligent Data Science and Artificial
Intelligence Center (IDEAI)
Universitat Politecnica de Catalunya (UPC)
Barcelona Supercomputing Center (BSC)
TUTORIAL
Thessaloniki, Greece
8 January 2019

2. bit.ly/MMM2019
@DocXavi
2
Acknowledgments

3. bit.ly/MMM2019
@DocXavi
3
Densely linked slides

4. bit.ly/MMM2019
@DocXavi
4
Outline
1. Motivation
2. Deep Neural Topologies
3. Multimedia Encoding and Decoding
4. Multimodal Architectures
a. Cross-modal
b. Self-supervised Learning
c. Multimodal (input)

5. bit.ly/MMM2019
@DocXavi
Text
Audio
5
Speech
Vision

6. bit.ly/MMM2019
@DocXavi
Text
Audio
6
Speech
Vision

7. bit.ly/MMM2019
@DocXavi
Text
Audio
7
Speech
Vision

8. bit.ly/MMM2019
@DocXavi
8

9. bit.ly/MMM2019
@DocXavi
9
Encoder
0
1
0
Cat
A Krizhevsky, I Sutskever, GE Hinton “Imagenet classification with deep convolutional neural networks” NIPS 2012

10. bit.ly/MMM2019
@DocXavi
10Slide concept: Perronin, F., Tutorial on LSVR @ CVPR’14, Output embedding for LSVR
One-hot Representation
[1,0,0]
[0,1,0]
[0,0,1]

11. bit.ly/MMM2019
@DocXavi
11
Encoder
Representation

12. bit.ly/MMM2019
@DocXavi
12
Encoder
Representation

13. bit.ly/MMM2019
@DocXavi
13
Decoder
Radford, Alec, Luke Metz, and Soumith Chintala. "Unsupervised representation learning with deep convolutional generative
adversarial networks." ICLR 2016. #DCGAN
0
1
0
Cat
Fig: Xudong Mao #DCGAN

14. bit.ly/MMM2019
@DocXavi
14
Encoder Decoder
Representation

15. bit.ly/MMM2019
@DocXavi
15
Encoder Decoder
Representation

16. bit.ly/MMM2019
@DocXavi
16
Outline
1. Motivation
2. Deep Neural Topologies
3. Multimedia Encoding and Decoding
4. Multimodal Architectures
a. Cross-modal
b. Self-supervised Learning
c. Multimodal (input)

17. bit.ly/MMM2019
@DocXavi
17
One Perceptron

18. bit.ly/MMM2019
@DocXavi
18
One Perceptron
Multiple options as activation functions f(·):

19. bit.ly/MMM2019
@DocXavi
19
A Layer of N Perceptrons

20. bit.ly/MMM2019
@DocXavi
20
Neural Network (single hidden layer)

21. bit.ly/MMM2019
@DocXavi
21
Multi-Layer Perceptron (MLP)
INPUT(x)
OUTPUT(y)
FeedForward
Hidden
States
h1
& h2
Feed-forward
Weights (Wi
)
Figure: Hugo Larochelle

22. bit.ly/MMM2019
@DocXavi
22
MLPs for Multimedia Data
Limitation #1:
Very large amount of input data samples (xi
),
which requires a gigantic amount of model
parameters.
Figure: Ranzatto
Limitation #2:
Does not naturally handle input data of
variable dimension
(eg. audio/video/word sequences).
#1
#2

23. bit.ly/MMM2019
@DocXavi
23
Limitation #1:
Very large amount of input data samples (xi
), which requires a gigantic amount of
model parameters.
For a 200x200 image, we have 4x104
neurons each one with 4x104
inputs,
that is 16x108
parameters (*), only for
one layer!!!
Figure Credit: Ranzatto
16x108
(*) biases not counted
MLPs for Multimedia Data #1

24. bit.ly/MMM2019
@DocXavi
24
MLPs for Multimedia Data
Locally connected network:
For a 200x200 image, we have 4x104
neurons each one with 10x10 “local
connections” (also called receptive
field) inputs, that is 4x106
What else can we do to reduce the
number of parameters?
Figure Credit: Ranzatto
4x106
#1

25. bit.ly/MMM2019
@DocXavi
25
CNNs for Multimedia Data (2D)
#1
104
Ex: With 100 different filters (or
feature extractors) of size 10x10,
the number of parameters is 104a
Convolutional Neural Networks
(ConvNets, CNNs):
Translation invariance: we can use same
parameters to capture a specific “feature” in
any area of the image. We can use different
sets of parameters to capture different
features.
These operations are equivalent to perform
convolutions with different filters.

26. bit.ly/MMM2019
@DocXavi
26
CNNs for Multimedia Data (2D)
LeCun, Yann, Léon Bottou, Yoshua Bengio, and Patrick Haffner. "Gradient-based learning applied to document recognition." Proceedings of the IEEE
86, no. 11 (1998): 2278-2324.
LeNet-5

27. bit.ly/MMM2019
@DocXavi
27
CNNs for Multimedia Data (1D)
hi there how are you doing ?
100 dim
Let’s say we have a sequence of 100 dimensional vectors describing words (text).
sequence length = 8
arrangein 2D
100
dim
sequence length = 8
,
Slide: Santiago Pascual (UPC DLAI 2018)

28. bit.ly/MMM2019
@DocXavi
28
CNNs for Multimedia Data (1D)
Slide: Santiago Pascual (UPC DLAI 2018)
We can apply a 1D convolutional activation over the 2D matrix: for an arbitrary kernel of width=3
100
dim
sequence length = 8
K1
Each 1D convolutional kernel is a
2D matrix of size (3, 100)100
dim

29. bit.ly/MMM2019
@DocXavi
29
CNNs for Multimedia Data (1D)
Slide: Santiago Pascual (UPC DLAI 2018)
(Keep in mind we are working with depth=100 although here we depict just depth=1 for simplicity)
1 2 3 4 5 6 7 8
w1 w2 w3
w1 w2 w3
w1 w2 w3
w1 w2 w3
w1 w2 w3
1 2 3 4 5 6
w1 w2 w3
The length result of the convolution is
known to be:
seq_length - filter_width + 1 = 8 - 3 + 1 = 6
So the output matrix will be (6, 100).

30. bit.ly/MMM2019
@DocXavi
30
CNNs for Multimedia Data (1D)
Slide: Santiago Pascual (UPC DLAI 2018)
When can add zero padding on both sides of the sequence:
1 2 3 4 5 6 7 8
w1 w2 w3
w1 w2 w3
w1 w2 w3
w1 w2 w3
w1 w2 w3
1 2 3 4 5 6 7 8
w1 w2 w3
The length result of the convolution is well
known to be:
seq_length - filter_width + 1 = 10 - 3 + 1 = 8
So the output matrix will be (8, 100).
0 0
w1 w2 w3
w1 w2 w3

31. bit.ly/MMM2019
@DocXavi
31
CNNs for Multimedia Data (1D)
Slide: Santiago Pascual (UPC DLAI 2018)
When can add zero padding on just one sides of the sequence:
1 2 3 4 5 6 7 8
w1 w2 w3
w1 w2 w3
w1 w2 w3
w1 w2 w3
w1 w2 w3
1 2 3 4 5 6 7 8
w1 w2 w3
The length result of the convolution is well
known to be:
seq_length - filter_width + 1 = 10 - 3 + 1 = 8
So the output matrix will be (8, 100) because
we had padding
HOWEVER: now every time-step t depends
on the two previous inputs as well as the
current time-step → every output is causal
Roughly: We make a causal convolution by
padding left the sequence with
(filter_width - 1) zeros
0
w1 w2 w3
w1 w2 w3
0

32. bit.ly/MMM2019
@DocXavi
32
MLPs for Sequences
Limitation #2:
Dimensionality of input data is variable (eg. audio/video/word sequences).
If we have a sequence of samples...
predict sample x[t+1] knowing previous values {x[t], x[t-1], x[t-2], …, x[t-τ]}
Slide: Santiago Pascual (UPC 2017)
#2

33. bit.ly/MMM2019
@DocXavi
33
MLPs for Sequences
Slide: Santiago Pascual (UPC 2017)
Feed Forward approach:
● static window of size L
● slide the window time-step wise
...
...
...
x[t+1]
x[t-L], …, x[t-1], x[t]
x[t+1]
L
#2

34. bit.ly/MMM2019
@DocXavi
34
MLPs for Sequences
Slide: Santiago Pascual (UPC 2017)
Feed Forward approach:
● static window of size L
● slide the window time-step wise
...
...
...
x[t+2]
x[t-L+1], …, x[t], x[t+1]
...
...
...
x[t+1]
x[t-L], …, x[t-1], x[t]
x[t+2]
L
#2

35. bit.ly/MMM2019
@DocXavi
35
MLPs for Sequences
Slide: Santiago Pascual (UPC 2017)
35
Feed Forward approach:
● static window of size L
● slide the window time-step wise
x[t+3]
L
...
...
...
x[t+3]
x[t-L+2], …, x[t+1], x[t+2]
...
...
...
x[t+2]
x[t-L+1], …, x[t], x[t+1]
...
...
...
x[t+1]
x[t-L], …, x[t-1], x[t]
#2

36. bit.ly/MMM2019
@DocXavi
36
Slide: Santiago Pascual (UPC 2017)
36
...
...
...
x1, x2, …, xL
Problems for the feed forward + static window approach:
● What’s the matter increasing L? → Fast growth of num of parameters!
● Decisions are independent between time-steps!
○ The network doesn’t care about what happened at previous time-step, only present window
matters → doesn’t look good
x1, x2, …, xL, …, x2L
...
...
x1, x2, …, xL, …, x2L, …, x3L
...
...
...
...
MLPs for Sequences #2

37. bit.ly/MMM2019
@DocXavi
37
RNNs for Sequences
#2
Solution A: Build specific connections capturing the temporal
evolution
→ Shared weights in time

38. bit.ly/MMM2019
@DocXavi
38
RNNs for Sequences
#2
Recurrent
Weights (U)
Feed-forward
Weights (W)

39. bit.ly/MMM2019
@DocXavi
39
RNNs for Sequences
#2
Updated
state
Previous
state
INPUT(x)

40. bit.ly/MMM2019
@DocXavi
40
RNNs for Sequences
#2
time
time
Unfold
(Rotation 90o
)
Unfold
(Rotation 90o
)

41. bit.ly/MMM2019
@DocXavi
41
RNNs for Sequences
#2
Raimi Karim, “Animated RNN, LSTM and GRU” (Toward Data Science 2018)

42. bit.ly/MMM2019
@DocXavi
42
LSTMs for Sequences
#2
Hochreiter, Sepp, and Jürgen Schmidhuber. "Long short-term memory." Neural computation 9, no. 8 (1997): 1735-1780.

43. bit.ly/MMM2019
@DocXavi
43
LSTMs for Sequences
#2
Raimi Karim, “Animated RNN, LSTM and GRU” (Toward Data Science 2018)

44. bit.ly/MMM2019
@DocXavi
44
GRUs for Sequences
#2
Cho, Kyunghyun, Bart Van Merriënboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua
Bengio. "Learning phrase representations using RNN encoder-decoder for statistical machine translation." AMNLP 2014.
GRU obtains a similar performance as LSTM with one gate less.

45. bit.ly/MMM2019
@DocXavi
45
GRUs for Sequences
#2
Raimi Karim, “Animated RNN, LSTM and GRU” (Toward Data Science 2018)

46. bit.ly/MMM2019
@DocXavi
46
Efficient RNN for Sequences
Used
Unused
Victor Campos, Brendan Jou, Xavier Giro-i-Nieto, Jordi Torres, and Shih-Fu Chang. “Skip RNN: Learning to Skip State
Updates in Recurrent Neural Networks”, ICLR 2018. #SkipRNN

47. bit.ly/MMM2019
@DocXavi
47
Efficient RNN for Sequences
Victor Campos, Brendan Jou, Xavier Giro-i-Nieto, Jordi Torres, and Shih-Fu Chang. “Skip RNN: Learning to Skip State
Updates in Recurrent Neural Networks”, ICLR 2018. #SkipRNN
Used Unused
CNN CNN CNN...
RNN RNN RNN...

48. bit.ly/MMM2019
@DocXavi
48
Attention
Chis Olah & Shan Cate, “Attention and Augmented Recurrent Neural Networks” (Google Bain 2016)

49. bit.ly/MMM2019
@DocXavi
49
Attention
Chis Olah & Shan Cate, “Attention and Augmented Recurrent Neural Networks” (Google Brain 2016)

50. bit.ly/MMM2019
@DocXavi
50
Self-Attention
Jay Alammar, “The Illustrated Transformer”
Self-attention refers to attending to other elements from the SAME sequence.

51. bit.ly/MMM2019
@DocXavi
51
Auto-regressive FF for Sequences
#2
van den Oord, Aaron, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew
Senior, and Koray Kavukcuoglu. "WaveNet: A Generative Model for Raw Audio." arXiv preprint arXiv:1609.03499 (2016).

52. bit.ly/MMM2019
@DocXavi
52
Auto-regression for Sequences
Van Oord, Aaron, Nal Kalchbrenner, and Koray Kavukcuoglu. "Pixel Recurrent Neural Networks." ICML 2016.
#2

53. bit.ly/MMM2019
@DocXavi
53
Outline
1. Motivation
2. Deep Neural Topologies
3. Multimedia Encoding and Decoding
4. Multimodal Architectures
a. Cross-modal
b. Self-supervised Learning
c. Multimodal (input)
d. Multi-task (output)

54. bit.ly/MMM2019
@DocXavi
54

55. bit.ly/MMM2019
@DocXavi
55
Encoder
Representation

56. bit.ly/MMM2019
@DocXavi
56
Image Encoding
A Krizhevsky, I Sutskever, GE Hinton “Imagenet classification with deep convolutional neural networks” NIPS 2012
Cat
CNN FC

57. bit.ly/MMM2019
@DocXavi
57
Video Encoding
Slide: Víctor Campos (UPC 2018)
CNN CNN CNN...
Combination method
Combination is commonly
implemented as a small NN on
top of a pooling operation
(e.g. max, sum, average).
Drawback: pooling is not
aware of the temporal order!
Ng et al., Beyond short snippets: Deep networks for video classification, CVPR 2015

58. bit.ly/MMM2019
@DocXavi
58
Video Encoding
Slide: Víctor Campos (UPC 2018)
Recurrent Neural Networks are
well suited for processing
sequences.
Drawback: RNNs are sequential
and cannot be parallelized.
Donahue et al., Long-term Recurrent Convolutional Networks for Visual Recognition and Description, CVPR 2015
CNN CNN CNN...
RNN RNN RNN...

59. bit.ly/MMM2019
@DocXavi
59
Decoder
Representation

60. bit.ly/MMM2019
@DocXavi
60
Image Decoding
CNN
Radford, Alec, Luke Metz, and Soumith Chintala. "Unsupervised representation learning with deep convolutional generative
adversarial networks." ICLR 2016. #DCGAN

61. bit.ly/MMM2019
@DocXavi
61
Encoder Decoder
Representation

62. bit.ly/MMM2019
@DocXavi
62
Image Encoding and Decoding
Noh et al. Learning Deconvolution Network for Semantic Segmentation. ICCV 2015
“Regular” VGG “Upside down” VGG

63. bit.ly/MMM2019
@DocXavi
63
Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. "U-net: Convolutional networks for biomedical image segmentation."
MICCAI 2015.
Isola, Phillip, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. "Image-to-image translation with conditional adversarial
networks." CVPR 2017.

64. bit.ly/MMM2019
@DocXavi
64
Encoder
Representation

65. bit.ly/MMM2019
@DocXavi
65
Encoder
Representation

66. 66
?
How to encode text ?

67. 67
Example: letters. |V| = 30
‘a’: x = 1
‘b’: x = 2
‘c’: x = 3
.
.
.
‘.’: x = 30
We impose fake range ordering
How to encode text ?

68. 68
One hot encoding
Example: letters. |V| = 30
‘a’: xT
= [1,0,0, ..., 0]
‘b’: xT
= [0,1,0, ..., 0]
‘c’: xT
= [0,0,1, ..., 0]
.
.
.
‘.’: xT
= [0,0,0, ..., 1]

69. 69
One hot encoding
Number of words, |V| ?
B2: 5K
C2: 18K
LVSR: 50-100K
Wikipedia (1.6B): 400K
Crawl data (42B): 2M
cat: xT
= [1,0,0, ..., 0]
dog: xT
= [0,1,0, ..., 0]
.
.
house: xT
= [0,0,0, …,0,1,0,...,0]
.
.
.

70. cat: xT
= [1,0,0, ..., 0]
dog: xT
= [0,1,0, ..., 0]
.
.
house: xT
= [0,0,0, …,0,1,0,...,0]
.
.
.
70
One hot encoding
● Large dimensionality
● Sparse representation (mostly zeros)
● Blind representation
○ Only operators: ‘!=’ and ‘==’

71. 71
Text projection to word embeddings
The one-hot is linearly projected to
a embedded space of lower
dimension with a MLP.
FC
Representation

72. 72
Embed high dimensional data points
(i.e. feature codes) so that pairwise
distances are preserved in local
neighborhoods.
Maaten & Hinton. Visualizing High-Dimensional Data using t-SNE. Journal of Machine Learning Research (2008) #tsne.
t-SNE
Figure:
Christopher Olah, Visualizing Representations
Text projection to word embeddings

73. 73Pennington, Jeffrey, Richard Socher, and Christopher Manning. "Glove: Global vectors for word representation." EMNLP
2014
Woman-Man
Text projection to word embeddings

74. 74
● Represent words using vectors of reduced dimension d (~100 - 500)
● Meaningful (semantic, syntactic) distances
● Good embeddings are useful for many other tasks
Text projection to word embeddings

75. 75
Training Word Embeddings
Figure:
TensorFlow tutorial
Bengio, Yoshua, Réjean Ducharme, Pascal Vincent, and Christian Jauvin. "A neural probabilistic language
model." Journal of machine learning research 3, no. Feb (2003): 1137-1155.
Self-supervised
learning

76. 76Mikolov, Tomas, Ilya Sutskever, Kai Chen, Greg S. Corrado, and Jeff Dean. "Distributed representations of
words and phrases and their compositionality." NIPS 2013 #word2vec #continuousbow
the cat climbed a tree
Given context:
a, cat, the, tree
Estimate prob. of
climbed
Self-supervised
learning
Training Word Embeddings

77. 77Mikolov, Tomas, Ilya Sutskever, Kai Chen, Greg S. Corrado, and Jeff Dean. "Distributed representations of
words and phrases and their compositionality." NIPS 2013 #word2vec #skipgram
Self-supervised
learning
the cat climbed a tree
Given word:
climbed
Estimate prob. of context words:
a, cat, the, tree
Training Word Embeddings

78. bit.ly/MMM2019
@DocXavi
78
Fig: Kyunghyun Cho, “Introduction to Neural Machine Translation with GPUs” (2015)
Cho, Kyunghyun, Bart Van Merriënboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua
Bengio. "Learning phrase representations using RNN encoder-decoder for statistical machine translation." EMNLP 2014.
(2)
(3)
Text Encoding
RNN
FC
Representation

79. bit.ly/MMM2019
@DocXavi
79
Gehring, Jonas, Michael Auli, David Grangier, Denis Yarats, and Yann N. Dauphin. "Convolutional sequence to sequence
learning." ICML 2017.
Text Encoding
CNN

80. bit.ly/MMM2019
@DocXavi
80
Decoder
Representation

81. bit.ly/MMM2019
@DocXavi
81
Text Decoding
Kyunghyun Cho, “Introduction to Neural Machine Translation with GPUs” (2015)
RNN
Representation

82. bit.ly/MMM2019
@DocXavi
82
Encoder Decoder
Representation

83. bit.ly/MMM2019
@DocXavi
83
Neural Machine Translation (NMT)
Kyunghyun Cho, “Introduction to Neural Machine Translation with GPUs” (2015)

84. 84
Neural Machine Translation (NMT)

85. 85
Kyunghyun Cho, “Introduction to Neural Machine Translation with GPUs” (2015)
Representation or
Embedding
Neural Machine Translation (NMT)

86. 86
Neural Machine Translation (NMT)
Sutskever, Ilya, Oriol Vinyals, and Quoc V. Le. "Sequence to sequence learning with neural networks." NIPS 2014.
The Seq2Seq variation:
● trigger the output generation with an input <go> symbol.
● the predicted word at timestep t, becomes the input at t+1.

87. bit.ly/MMM2019
@DocXavi
87
NMT with Attention
Slide: Marta R. Costa-jussà (UPC DLAI 2018)
encoder
decoder
+
Attention allows to use
multiple vectors, based on
the length of the input.

88. bit.ly/MMM2019
@DocXavi
88
NMT with Attention
Attention
RNN

89. bit.ly/MMM2019
@DocXavi
89
NMT with Attention
Chis Olah & Shan Cate, “Attention and Augmented Recurrent Neural Networks” (Google Bain 2016)

90. 90
Neural Machine Translation (NMT)
CNN
Gehring, Jonas, Michael Auli, David Grangier, Denis Yarats, and Yann N. Dauphin. "Convolutional sequence to sequence
learning." ICML 2017.
Attention

91. 91
Neural Machine Translation (NMT)
Self-Attention
Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia
Polosukhin. "Attention is all you need." NIPS 2017.

92. bit.ly/MMM2019
@DocXavi
92
Encoder Decoder
Representation

93. bit.ly/MMM2019
@DocXavi
93
Encoder
Representation

94. bit.ly/MMM2019
@DocXavi
94
Chan, William, Navdeep Jaitly, Quoc Le, and Oriol Vinyals. "Listen, attend and spell: A neural network for large vocabulary
conversational speech recognition." ICASSP 2016.
Speech Encoding
RNN
CNN
Oord, Aaron van den, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew
Senior, and Koray Kavukcuoglu. "Wavenet: A generative model for raw audio." arXiv preprint arXiv:1609.03499 (2016).

95. bit.ly/MMM2019
@DocXavi
95
Decoder
Representation

96. bit.ly/MMM2019
@DocXavi
96
Audio Decoding
Mehri, Soroush, Kundan Kumar, Ishaan Gulrajani, Rithesh Kumar, Shubham Jain, Jose Sotelo, Aaron Courville, and Yoshua
Bengio. "SampleRNN: An unconditional end-to-end neural audio generation model." ICLR 2017.
RNN
CNN
Oord, Aaron van den, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew
Senior, and Koray Kavukcuoglu. "Wavenet: A generative model for raw audio." arXiv preprint arXiv:1609.03499 (2016).

97. bit.ly/MMM2019
@DocXavi
97
Encoder Decoder
Representation

98. bit.ly/MMM2019
@DocXavi
98
Encoder Decoder
Representation
Raw
MFCC
Mel spectrum
Raw
MFCC
Mel spectrum

99. bit.ly/MMM2019
@DocXavi
99
Speech Enhancement
Pascual, Santiago, Antonio Bonafonte, and Joan Serra. "SEGAN: Speech enhancement generative adversarial network."
Interspeech 2017.

100. bit.ly/MMM2019
@DocXavi
100
Encoder Decoder
Representation

101. bit.ly/MMM2019
@DocXavi
101
Outline
1. Motivation
2. Deep Neural Topologies
3. Multimedia Encoding and Decoding
4. Multimodal Architectures
a. Cross-modal
b. Self-supervised Learning
c. Multimodal (input)
d. Multi-task (output)

102. bit.ly/MMM2019
@DocXavi
102
Encoder Decoder
Representation

103. bit.ly/MMM2019
@DocXavi
103
Encoder Decoder
Representation

104. bit.ly/MMM2019
@DocXavi
104
Automatic Speech Recognition (ASR)
Slide: Hannun, Awni. "Sequence Modeling with CTC." Distill 2.11 (2017): e8.

105. bit.ly/MMM2019
@DocXavi
105
Automatic Speech Recognition (ASR)
Graves et al. Connectionist Temporal Classification: Labelling Unsegmented Sequence Data with Recurrent
Neural Networks. ICML 2006
The Connectionist Temporal Classification loss (CTC) allows training RNNs with
no need of exact alignement.
Figure: Hannun, Awni. "Sequence Modeling with CTC." Distill 2.11 (2017): e8.
● Avoiding the need for
alignment between input and
output sequence by
predicting an additional “_”
blank word
● Before computing the loss,
repeated words and blank
tokens are removed

106. bit.ly/MMM2019
@DocXavi
106
Automatic Speech Recognition (ASR)
Sequence of spectrograms

107. bit.ly/MMM2019
@DocXavi
107
Automatic Speech Recognition (ASR)
Baidu Research – 34 authors- , “Deep Speech 2: End-to-end Speech Recognition in English and Mandarin”, arXiv:1512.02595
(Dec 2015) [Demo]

108. bit.ly/MMM2019
@DocXavi
108
Automatic Speech Speller w/Attention
Chan, William, Navdeep Jaitly, Quoc Le, and Oriol Vinyals. "Listen, attend and spell: A neural network for large vocabulary
conversational speech recognition." ICASSP 2016. #LAS
Listener (encoder)
Speller (decoder)

109. bit.ly/MMM2019
@DocXavi
109Chis Olah & Shan Cate, “Attention and Augmented Recurrent Neural Networks” (Google Bain 2016)
Automatic Speech Speller w/ Attention

110. bit.ly/MMM2019
@DocXavi
110
Encoder Decoder
Representation

111. bit.ly/MMM2019
@DocXavi
111
Speech Synthesis
Oord, Aaron van den, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew
Senior, and Koray Kavukcuoglu. "Wavenet: A generative model for raw audio." arXiv preprint arXiv:1609.03499 (2016).

112. bit.ly/MMM2019
@DocXavi
112
Speech Synthesis
Prenger, Ryan, Rafael Valle, and Bryan Catanzaro. "WaveGlow: A Flow-based Generative Network for Speech Synthesis." arXiv
preprint arXiv:1811.00002 (2018).

113. bit.ly/MMM2019
@DocXavi
113
Encoder Decoder
Representation

114. bit.ly/MMM2019
@DocXavi
114
Vinyals, Oriol, Alexander Toshev, Samy Bengio, and Dumitru Erhan. "Show and tell: A neural image caption generator."
CVPR 2015.
Image Captioning

115. bit.ly/MMM2019
@DocXavi
115
Image Captioning
(Slides by Marc Bolaños): Karpathy, Andrej, and Li Fei-Fei. "Deep visual-semantic alignments for generating image
descriptions." CVPR 2015 #DeepImageSent

116. bit.ly/MMM2019
@DocXavi
116
Captioning: Show, Attend & Tell
Xu, Kelvin, Jimmy Ba, Ryan Kiros, Kyunghyun Cho, Aaron C. Courville, Ruslan Salakhutdinov, Richard S. Zemel, and Yoshua
Bengio. "Show, Attend and Tell: Neural Image Caption Generation with Visual Attention." ICML 2015

117. bit.ly/MMM2019
@DocXavi
117
Captioning: Show, Attend & Tell
Xu, Kelvin, Jimmy Ba, Ryan Kiros, Kyunghyun Cho, Aaron C. Courville, Ruslan Salakhutdinov, Richard S. Zemel, and Yoshua
Bengio. "Show, Attend and Tell: Neural Image Caption Generation with Visual Attention." ICML 2015

118. bit.ly/MMM2019
@DocXavi
118
Johnson, Justin, Andrej Karpathy, and Li Fei-Fei. "Densecap: Fully convolutional localization networks for dense
captioning." CVPR 2016
Captioning (+ Detection): DenseCap

119. bit.ly/MMM2019
@DocXavi
119
Captioning (+ Detection): DenseCap
XAVI: “man has
short hair”, “man
with short hair”
AMAIA:”a woman
wearing a black
shirt”, “
BOTH: “two men
wearing black
glasses”
Johnson, Justin, Andrej Karpathy, and Li Fei-Fei. "Densecap: Fully convolutional localization networks for dense
captioning." CVPR 2016

120. bit.ly/MMM2019
@DocXavi
120
Johnson, Justin, Andrej Karpathy, and Li Fei-Fei. "Densecap: Fully convolutional localization networks for dense
captioning." CVPR 2016
Captioning (+ Detection): DenseCap

121. bit.ly/MMM2019
@DocXavi
121
Jeffrey Donahue, Lisa Anne Hendricks, Sergio Guadarrama, Marcus Rohrbach, Subhashini Venugopalan, Kate Saenko, Trevor
Darrel. Long-term Recurrent Convolutional Networks for Visual Recognition and Description, CVPR 2015. code
Captioning: Video

122. bit.ly/MMM2019
@DocXavi
122
(Slides by Marc Bolaños) Pingbo Pan, Zhongwen Xu, Yi Yang,Fei Wu,Yueting Zhuang Hierarchical Recurrent Neural
Encoder for Video Representation with Application to Captioning, CVPR 2016.
LSTM unit
(2nd layer)
Time
Image
t = 1 t = T
hidden state
at t = T
first chunk
of data
Captioning: Video

123. bit.ly/MMM2019
@DocXavi
123
Sign Language Translation
Camgoz, Necati Cihan, et al. Neural Sign Language Translation. CVPR 2018.

124. bit.ly/MMM2019
@DocXavi
124
Assael, Yannis M., Brendan Shillingford, Shimon Whiteson, and Nando de Freitas. "LipNet: End-to-End Sentence-level Lipreading."
(2016).

125. bit.ly/MMM2019
@DocXavi
125
Lip Reading
Assael, Yannis M., Brendan Shillingford, Shimon Whiteson, and Nando de Freitas. "LipNet: End-to-End Sentence-level
Lipreading." (2016).

126. bit.ly/MMM2019
@DocXavi
126
Chung, Joon Son, Andrew Senior, Oriol Vinyals, and Andrew Zisserman. "Lip reading sentences in the wild."
CVPR 2017

127. bit.ly/MMM2019
@DocXavi
127
Lipreading: Watch, Listen, Attend & Spell
Audio
features
Image
features
Chung, Joon Son, Andrew Senior, Oriol Vinyals, and Andrew Zisserman. "Lip reading sentences in the wild." CVPR 2017

128. bit.ly/MMM2019
@DocXavi
128
Lipreading: Watch, Listen, Attend & Spell
Chung, Joon Son, Andrew Senior, Oriol Vinyals, and Andrew Zisserman. "Lip reading sentences in the wild." CVPR 2017
Attention over output
states from audio and
video is computed at
each timestep

129. bit.ly/MMM2019
@DocXavi
129
Encoder Decoder
Representation

130. bit.ly/MMM2019
@DocXavi
130
Reed, Scott, Zeynep Akata, Xinchen Yan, Lajanugen Logeswaran, Bernt Schiele, and Honglak Lee. "Generative adversarial
text to image synthesis." ICML 2016.
Text-to-Image

131. bit.ly/MMM2019
@DocXavi
131
Reed, Scott, Zeynep Akata, Xinchen Yan, Lajanugen Logeswaran, Bernt Schiele, and Honglak Lee. "Generative adversarial
text to image synthesis." ICML 2016.
Text-to-Image

132. bit.ly/MMM2019
@DocXavi
132
Text-to-Image
Salvador, Amaia, Michal Drozdzal, Xavier Giro-i-Nieto, and Adriana Romero. "Inverse Cooking: Recipe
Generation from Food Images." arXiv preprint arXiv:1812.06164 (2018).

133. bit.ly/MMM2019
@DocXavi
133
Encoder Encoder
Representation

134. bit.ly/MMM2019
@DocXavi
134
Self-supervised Feature Learning
Owens, Andrew, Jiajun Wu, Josh H. McDermott, William T. Freeman, and Antonio Torralba. "Ambient sound
provides supervision for visual learning." ECCV 2016
Based on the assumption that ambient sound in video is related to the visual
semantics.

135. bit.ly/MMM2019
@DocXavi
135
Self-supervised Feature Learning
Owens, Andrew, Jiajun Wu, Josh H. McDermott, William T. Freeman, and Antonio Torralba. "Ambient sound
provides supervision for visual learning." ECCV 2016
Use videos to train a CNN that predicts the audio statistics of a frame.

136. bit.ly/MMM2019
@DocXavi
136
Self-supervised Feature Learning
Owens, Andrew, Jiajun Wu, Josh H. McDermott, William T. Freeman, and Antonio Torralba. "Ambient sound
provides supervision for visual learning." ECCV 2016
Task: Use the predicted audio stats to clusters images. Audio clusters built with
K-means overthe training set
Cluster assignments at test time (one row=one cluster)

137. bit.ly/MMM2019
@DocXavi
137
Self-supervised Feature Learning
Owens, Andrew, Jiajun Wu, Josh H. McDermott, William T. Freeman, and Antonio Torralba. "Ambient sound
provides supervision for visual learning." ECCV 2016
Although the CNN was not trained with class labels, local units with semantic
meaning emerge.

138. bit.ly/MMM2019
@DocXavi
138
Video Sonorization
Owens, Andrew, Phillip Isola, Josh McDermott, Antonio Torralba, Edward H. Adelson, and William T. Freeman.
"Visually indicated sounds." CVPR 2016.
Retrieve matching sounds for videos of people hitting objects with a drumstick.

139. bit.ly/MMM2019
@DocXavi
139
Video Sonorization
Owens, Andrew, Phillip Isola, Josh McDermott, Antonio Torralba, Edward H. Adelson, and William T. Freeman.
"Visually indicated sounds." CVPR 2016.
The Greatest Hits Dataset

140. bit.ly/MMM2019
@DocXavi
140
Video Sonorization
Owens, Andrew, Phillip Isola, Josh McDermott, Antonio Torralba, Edward H. Adelson, and William T. Freeman.
"Visually indicated sounds." CVPR 2016.
Audio Clip
Retrieval
Not end-to-end

141. bit.ly/MMM2019
@DocXavi
141
Owens, Andrew, Phillip Isola, Josh McDermott, Antonio Torralba, Edward H. Adelson, and William T.
Freeman. "Visually indicated sounds." CVPR 2016.

142. bit.ly/MMM2019
@DocXavi
142
DecoderEncoder
Representation

143. bit.ly/MMM2019
@DocXavi
143
Speech Reconstruction
Ephrat, Ariel, and Shmuel Peleg. "Vid2speech: speech reconstruction from silent video." ICASSP 2017.
CNN
(VGG)
Frame from a
silent video
Audio feature
Post-hoc
synthesis

144. bit.ly/MMM2019
@DocXavi
144
Speech Reconstruction
Ephrat, Ariel, Tavi Halperin, and Shmuel Peleg. "Improved speech reconstruction from silent video." In ICCV 2017
Workshop on Computer Vision for Audio-Visual Media. 2017.

145. bit.ly/DLCV2018
#DLUPC
145
Ephrat, Ariel, Tavi Halperin, and Shmuel Peleg. "Improved speech reconstruction from silent video." In ICCV Workshop on
Computer Vision for Audio-Visual Media. 2017.

146. bit.ly/MMM2019
@DocXavi
146
Encoder Decoder
Representation

147. bit.ly/MMM2019
@DocXavi
147
Speech to Pixels
Amanda Duarte, Francisco Roldan, Miquel Tubau, Janna Escur, Santiago Pascual, Amaia Salvador, Eva Mohedano et al.
“Wav2Pix: Speech-conditioned Face Generation using Generative Adversarial Networks” (under progress)

148. bit.ly/MMM2019
@DocXavi
148
Speech to Pixels
Amanda Duarte, Francisco Roldan, Miquel Tubau, Janna Escur, Santiago Pascual, Amaia Salvador, Eva Mohedano et al.
“Wav2Pix: Speech-conditioned Face Generation using Generative Adversarial Networks” (under progress)
Generated faces from known identities.

149. bit.ly/MMM2019
@DocXavi
149
Speech to Pixels
Amanda Duarte, Francisco Roldan, Miquel Tubau, Janna Escur, Santiago Pascual, Amaia Salvador, Eva Mohedano et al.
“Wav2Pix: Speech-conditioned Face Generation using Generative Adversarial Networks” (under progress)
Faces from average speeches.

150. bit.ly/MMM2019
@DocXavi
150
Speech to Pixels
Amanda Duarte, Francisco Roldan, Miquel Tubau, Janna Escur, Santiago Pascual, Amaia Salvador, Eva Mohedano et al.
“Wav2Pix: Speech-conditioned Face Generation using Generative Adversarial Networks” (under progress)
Interpolated faces from interpolated speeches.

151. bit.ly/MMM2019
@DocXavi
151
Outline
1. Motivation
2. Deep Neural Topologies
3. Multimedia Encoding and Decoding
4. Multimodal Architectures
a. Cross-modal
b. Joint Representations (embeddings)
c. Multimodal (input)
d. Multi-task (output)

152. bit.ly/MMM2019
@DocXavi
152
Encoder Encoder
Representation

153. bit.ly/MMM2019
@DocXavi
153
Encoder Encoder
Representation

154. bit.ly/MMM2019
@DocXavi
154
Joint Representations (Embeddings)
Frome, Andrea, Greg S. Corrado, Jon Shlens, Samy Bengio, Jeff Dean, and Tomas Mikolov. "Devise: A deep
visual-semantic embedding model." NIPS 2013

155. bit.ly/MMM2019
@DocXavi
155
Zero-shot learning
Socher, R., Ganjoo, M., Manning, C. D., & Ng, A., Zero-shot learning through cross-modal transfer. NIPS 2013 [slides] [code]
No images from “cat” in
the training set...
...but they can still be
recognised as “cats”
thanks to the
representations learned
from text .

156. bit.ly/MMM2019
@DocXavi
156
Multimodal Retrieval
Aytar, Yusuf, Lluis Castrejon, Carl Vondrick, Hamed Pirsiavash, and Antonio Torralba. "Cross-Modal Scene Networks."
CVPR 2016.

157. bit.ly/MMM2019
@DocXavi
157
Multimodal Retrieval
Aytar, Yusuf, Lluis Castrejon, Carl Vondrick, Hamed Pirsiavash, and Antonio Torralba. "Cross-Modal Scene Networks."
CVPR 2016.

158. bit.ly/MMM2019
@DocXavi
158
Multimodal Retrieval
Amaia Salvador, Nicholas Haynes, Yusuf Aytar, Javier Marín, Ferda Ofli, Ingmar Weber, Antonio Torralba,
“Learning Cross-modal Embeddings for Cooking Recipes and Food Images”. CVPR 2017 #pic2recipe

159. bit.ly/MMM2019
@DocXavi
159
Encoder Encoder
Representation

160. bit.ly/MMM2019
@DocXavi
160
Multimodal Retrieval
Amanda Duarte, Dídac Surís, Amaia Salvador, Jordi Torres, and Xavier Giró-i-Nieto. "Cross-modal Embeddings for
Video and Audio Retrieval." ECCV Women in Computer Vision Workshop 2018.
Best
match
Audio feature

161. bit.ly/MMM2019
@DocXavi
161
Multimodal Retrieval
Best
match
Visual feature Audio feature
Amanda Duarte, Dídac Surís, Amaia Salvador, Jordi Torres, and Xavier Giró-i-Nieto. "Cross-modal Embeddings for
Video and Audio Retrieval." ECCV Women in Computer Vision Workshop 2018.

162. bit.ly/MMM2019
@DocXavi
162
Feature Learning by Label Transfer
Aytar, Yusuf, Carl Vondrick, and Antonio Torralba. "Soundnet: Learning sound representations from unlabeled video."
NIPS 2016.
Teacher network: Visual Recognition (object & scenes)

163. bit.ly/MMM2019
@DocXavi
163
Aytar, Yusuf, Carl Vondrick, and Antonio Torralba. "Soundnet: Learning sound representations from unlabeled video." NIPS
2016.

164. bit.ly/MMM2019
@DocXavi
164
Aytar, Yusuf, Carl Vondrick, and Antonio Torralba. "Soundnet: Learning sound representations from unlabeled video."
NIPS 2016.
Learned audio features are good for environmental sound recognition.
Feature Learning by Label Transfer

165. bit.ly/MMM2019
@DocXavi
165
Aytar, Yusuf, Carl Vondrick, and Antonio Torralba. "Soundnet: Learning sound representations from unlabeled video."
NIPS 2016.
Learned audio features are good for environmental sound recognition.
Feature Learning by Label Transfer

166. bit.ly/MMM2019
@DocXavi
166
Aytar, Yusuf, Carl Vondrick, and Antonio Torralba. "Soundnet: Learning sound representations from unlabeled video."
NIPS 2016.
Visualization of the 1D filters over raw audio in conv1.
Feature Learning by Label Transfer

167. bit.ly/MMM2019
@DocXavi
167
Aytar, Yusuf, Carl Vondrick, and Antonio Torralba. "Soundnet: Learning sound representations from unlabeled video."
NIPS 2016.
Visualization of the 1D filters over raw audio in conv1.
Feature Learning by Label Transfer

168. bit.ly/MMM2019
@DocXavi
168
Aytar, Yusuf, Carl Vondrick, and Antonio Torralba. "Soundnet: Learning sound representations from unlabeled video."
NIPS 2016.
Visualize video frames that mostly activate a neuron in a late layer (conv7)
Feature Learning by Label Transfer

169. bit.ly/MMM2019
@DocXavi
169
Aytar, Yusuf, Carl Vondrick, and Antonio Torralba. "Soundnet: Learning sound representations from unlabeled video."
NIPS 2016.
Visualize video frames that mostly activate a neuron in a late layer (conv7)
Feature Learning by Label Transfer

170. bit.ly/MMM2019
@DocXavi
170
Cross-modal Label Transfer
S Albanie, A Nagrani, A Vedaldi, A Zisserman, “Emotion Recognition in Speech using Cross-Modal Transfer in the Wild”
ACM Multimedia 2018.
Teacher network: Facial Emotion Recognition (visual)

171. bit.ly/MMM2019
@DocXavi
171
Joint Feature Learning
Korbar, Bruno, Du Tran, and Lorenzo Torresani. "Cooperative Learning of Audio and Video Models from Self-Supervised
Synchronization." NIPS 2018. #AVTS #selfsupervision

172. bit.ly/MMM2019
@DocXavi
172
Joint Feature Learning
Korbar, Bruno, Du Tran, and Lorenzo Torresani. "Cooperative Learning of Audio and Video Models from Self-Supervised
Synchronization." NIPS 2018. #AVTS #selfsupervision

173. bit.ly/MMM2019
@DocXavi
173
Joint Feature Learning
Korbar, Bruno, Du Tran, and Lorenzo Torresani. "Cooperative Learning of Audio and Video Models from Self-Supervised
Synchronization." NIPS 2018. #AVTS #selfsupervision

174. bit.ly/MMM2019
@DocXavi
174Arandjelović, Relja, and Andrew Zisserman. "Look, Listen and Learn." ICCV 2017. #selfsupervision
Joint Feature Learning

175. bit.ly/MMM2019
@DocXavi
175Arandjelović, Relja, and Andrew Zisserman. "Look, Listen and Learn." ICCV 2017.
Joint Feature Learning
Most activated unit in pool4 layer of the visual network.

176. bit.ly/MMM2019
@DocXavi
176Arandjelović, Relja, and Andrew Zisserman. "Look, Listen and Learn." ICCV 2017.
Joint Feature Learning
Visual features used to train a linear classifier on ImageNet.

177. bit.ly/MMM2019
@DocXavi
177Arandjelović, Relja, and Andrew Zisserman. "Look, Listen and Learn." ICCV 2017.
Joint Feature Learning
Most activated unit in pool4 layer of the audio network

178. bit.ly/MMM2019
@DocXavi
178Arandjelović, Relja, and Andrew Zisserman. "Look, Listen and Learn." ICCV 2017.
Joint Feature Learning
Audio features achieve state of the art performance.

179. bit.ly/MMM2019
@DocXavi
179
Sound Source Localization
Arandjelović, Relja, and Andrew Zisserman. "Objects that Sound." ECCV 2018. #selfsupervision

180. bit.ly/MMM2019
@DocXavi
180Arandjelović, Relja, and Andrew Zisserman. "Objects that Sound." ECCV 2018.

181. bit.ly/MMM2019
@DocXavi
181
Sound Source Localization
Senocak, Arda, Tae-Hyun Oh, Junsik Kim, Ming-Hsuan Yang, and In So Kweon. "Learning to Localize Sound Source in Visual
Scenes." CVPR 2018.

182. bit.ly/MMM2019
@DocXavi
182
Senocak, Arda, Tae-Hyun Oh, Junsik Kim, Ming-Hsuan Yang, and In So Kweon. "Learning to Localize Sound Source in Visual
Scenes." CVPR 2018.

183. bit.ly/MMM2019
@DocXavi
183
Encoder Encoder
Representation

184. bit.ly/MMM2019
@DocXavi
184
Speech Grounding (temporal)
Harwath, David, Antonio Torralba, and James Glass. "Unsupervised learning of spoken language with visual context." NIPS
2016. [talk]
Train a visual & speech networks with pairs of (non-)corresponding images & speech.

185. bit.ly/MMM2019
@DocXavi
185
Harwath, David, Antonio Torralba, and James Glass. "Unsupervised learning of spoken language with visual context." NIPS
2016. [talk]
Similarity curve show which regions of the spectrogram are relevant for the image.
Important: no text transcriptions used during the training !!
Speech Grounding (temporal)

186. bit.ly/MMM2019
@DocXavi
186
Speech Grounding (spatiotemporal)
Harwath, David, Adrià Recasens, Dídac Surís, Galen Chuang, Antonio Torralba, and James Glass. "Jointly Discovering Visual
Objects and Spoken Words from Raw Sensory Input." ECCV 2018.

187. bit.ly/MMM2019
@DocXavi
187
Harwath, David, Adrià Recasens, Dídac Surís, Galen Chuang, Antonio Torralba, and James Glass. "Jointly Discovering Visual Objects
and Spoken Words from Raw Sensory Input." ECCV 2018.
Regions matching the spoken word “WOMAN”:
Speech Grounding (spatiotemporal)

188. bit.ly/MMM2019
@DocXavi
188
Harwath, David, Adrià Recasens, Dídac Surís, Galen Chuang, Antonio Torralba, and James Glass. "Jointly Discovering Visual Objects
and Spoken Words from Raw Sensory Input." ECCV 2018

189. bit.ly/MMM2019
@DocXavi
189
Outline
1. Motivation
2. Deep Neural Topologies
3. Multimedia Encoding and Decoding
4. Multimodal Architectures
a. Cross-modal
b. Joint Representations (Embeddings)
c. Multimodal inputs

190. bit.ly/MMM2019
@DocXavi
190
Encoder
Decoder
Representation
Encoder
Representation

191. bit.ly/MMM2019
@DocXavi
191
Encoder
Decoder
Representation
Encoder
Representation

192. bit.ly/MMM2019
@DocXavi
192
Visual Question Answering
Antol, Stanislaw, Aishwarya Agrawal, Jiasen Lu, Margaret Mitchell, Dhruv Batra, C. Lawrence Zitnick, and Devi Parikh. "VQA:
Visual question answering." CVPR 2015.

193. bit.ly/MMM2019
@DocXavi
193
Visual Question Answering (VQA)
[z1
, z2
, … zN
] [y1
, y2
, … yM
]
“Is economic growth decreasing ?”
“Yes”
Encode
Encode
Decode

194. bit.ly/MMM2019
@DocXavi
194
Extract visual
features
Embedding
Predict answerMerge
Question
What object is flying?
Answer
Kite
Slide credit: Issey Masuda
Visual Question Answering (VQA)

195. bit.ly/MMM2019
@DocXavi
195
Visual Question Answering (VQA)
Masuda, Issey, Santiago Pascual de la Puente, and Xavier Giro-i-Nieto. "Open-Ended Visual
Question-Answering." ETSETB UPC TelecomBCN (2016).
Image
Question
Answer

196. bit.ly/MMM2019
@DocXavi
196
Visual Question Answering (VQA)
Francisco Roldán, Issey Masuda, Santiago Pascual de la Puente, and Xavier Giro-i-Nieto.
"Visual Question-Answering 2.0." ETSETB UPC TelecomBCN (2017).

197. bit.ly/MMM2019
@DocXavi
197
Noh, H., Seo, P. H., & Han, B. Image question answering using convolutional neural network with
dynamic parameter prediction. CVPR 2016
Dynamic Parameter Prediction Network (DPPnet)
Visual Question Answering (VQA)

198. bit.ly/MMM2019
@DocXavi
198
Visual Question Answering: Dynamic
(Slides and Slidecast by Santi Pascual): Xiong, Caiming, Stephen Merity, and Richard Socher. "Dynamic
Memory Networks for Visual and Textual Question Answering." ICML 2016

199. bit.ly/MMM2019
@DocXavi
199
Visual Question Answering: Grounded
(Slides and Screencast by Issey Masuda): Zhu, Yuke, Oliver Groth, Michael Bernstein, and Li Fei-Fei."Visual7W: Grounded
Question Answering in Images." CVPR 2016.

200. bit.ly/MMM2019
@DocXavi
200
Visual Reasoning
Johnson, Justin, Bharath Hariharan, Laurens van der Maaten, Li Fei-Fei, C. Lawrence Zitnick, and Ross Girshick. "CLEVR: A
Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning." CVPR 2017

201. bit.ly/MMM2019
@DocXavi
201
Visual Reasoning
(Slides by Fran Roldan) Justin Johnson, Bharath Hariharan, Laurens van der Maaten, Judy Hoffman, Fei-Fei Li, Larry
Zitnick, Ross Girshick , “Inferring and Executing Programs for Visual Reasoning”. ICCV 2017
Program Generator Execution Engine

202. bit.ly/MMM2019
@DocXavi
202
Visual Reasoning
Santoro, Adam, David Raposo, David G. Barrett, Mateusz Malinowski, Razvan Pascanu, Peter Battaglia, and Timothy
Lillicrap. "A simple neural network module for relational reasoning." NIPS 2017.
Relation Networks concatenate all possible pairs of objects with the an encoded question to later find the
answer with a MLP.

203. bit.ly/MMM2019
@DocXavi
203
Multimodal Machine Translation
Challenge on Multimodal Image Translation:
http://www.statmt.org/wmt17/multimodal-task.html#task1

204. bit.ly/MMM2019
@DocXavi
204
Encoder
Decoder
Representation
Encoder Representation

205. bit.ly/MMM2019
@DocXavi
205
Speech Separation with Vision (lips)
Afouras, Triantafyllos, Joon Son Chung, and Andrew Zisserman. "The Conversation: Deep Audio-Visual Speech Enhancement."
Interspeech 2018.

206. bit.ly/DLCV2018
#DLUPC
206
Afouras, Triantafyllos, Joon Son Chung, and Andrew Zisserman. "The Conversation: Deep Audio-Visual Speech
Enhancement." Interspeech 2018..

207. bit.ly/MMM2019
@DocXavi
207
Encoder
Decoder
Representation
Encoder Representation

208. bit.ly/MMM2019
@DocXavi
208
Visual Re-dubbing (pixels)
Chung, Joon Son, Amir Jamaludin, and Andrew Zisserman. "You said that?." BMVC 2017. #speech2vid

209. bit.ly/DLCV2018
#DLUPC
209Chung, Joon Son, Amir Jamaludin, and Andrew Zisserman. "You said that?." BMVC 2017. #speech2vid

210. bit.ly/MMM2019
@DocXavi
210
Chen, Lele, Zhiheng Li, Ross K. Maddox, Zhiyao Duan, and Chenliang Xu. "Lip Movements Generation at a Glance." ECCV
2018.
Visual Re-dubbing (pixels)

211. bit.ly/DLCV2018
#DLUPC
211Chen, Lele, Zhiheng Li, Ross K. Maddox, Zhiyao Duan, and Chenliang Xu. "Lip Movements Generation at a Glance." ECCV 2018.

212. bit.ly/MMM2019
@DocXavi
212
Visual Re-dubbing (pixels)
Vougioukas, Konstantinos, Stavros Petridis, and Maja Pantic. "End-to-End Speech-Driven Facial Animation with Temporal
GANs." arXiv preprint arXiv:1805.09313 (2018).
Adversarial losses at frame (spatial) & sequence (temporal) scales.

213. bit.ly/DLCV2018
#DLUPC
213
Vougioukas, Konstantinos, Stavros Petridis, and Maja Pantic. "End-to-End Speech-Driven Facial Animation with Temporal
GANs." arXiv preprint arXiv:1805.09313 (2018).

214. bit.ly/MMM2019
@DocXavi
214
Visual Re-dubbing (lip keypoints)
Suwajanakorn, Supasorn, Steven M. Seitz, and Ira Kemelmacher-Shlizerman. "Synthesizing Obama: learning lip sync from
audio." SIGGRAPH 2017.

215. bit.ly/DLCV2018
#DLUPC
215
Karras, Tero, Timo Aila, Samuli Laine, Antti Herva, and Jaakko Lehtinen. "Audio-driven facial animation by
joint end-to-end learning of pose and emotion." SIGGRAPH 2017

216. bit.ly/MMM2019
@DocXavi
216
Visual Re-dubbing (3D meshes)
Karras, Tero, Timo Aila, Samuli Laine, Antti Herva, and Jaakko Lehtinen. "Audio-driven facial animation by joint end-to-end
learning of pose and emotion." SIGGRAPH 2017

217. bit.ly/DLCV2018
#DLUPC
217
Karras, Tero, Timo Aila, Samuli Laine, Antti Herva, and Jaakko Lehtinen. "Audio-driven facial animation by
joint end-to-end learning of pose and emotion." SIGGRAPH 2017

218. bit.ly/MMM2019
@DocXavi
218
Outline
1. Motivation
2. Deep Neural Topologies
3. Multimedia Encoding and Decoding
4. Multimodal Architectures
a. Cross-modal
b. Joint Representations (Embeddings)
c. Multimodal inputs

219. bit.ly/MMM2019
@DocXavi
Text
Audio
219
Speech
Vision

220. bit.ly/MMM2019
@DocXavi
220
Deep Learning online courses by UPC:
● MSc course [2017] [2018]
● BSc course [2018] [2019]
● 1st edition (2016)
● 2nd edition (2017)
● 3rd edition (2018)
● 4th edition (2019)
● 1st edition (2017)
● 2nd edition (2018)
BSc course 22 to 29 January 2019
Registrations open for Spring 2019
(Speech) / Summer 2019 (NLP)Registration open for 2019

221. bit.ly/MMM2019
@DocXavi
221
Training for professionals
Sign up here. Course starts on February 2019.

222. bit.ly/MMM2019
@DocXavi
222
Deep Learning & AI in Barcelona
deeplearning.barcelona bcn.ai

223. bit.ly/MMM2019
@DocXavi
223
Phd Opening @ BSC on Multimodal
[Phd in Multimodal Deep Reinforcement Learning]

224. bit.ly/MMM2019
@DocXavi
224
Our team at UPC-BSC Barcelona
Victor
Campos
Amaia
Salvador
Amanda
Duarte
Dèlia
Fernández
Eduard
Ramon
Andreu
Girbau
Dani
Fojo
Oscar
Mañas
Santi
Pascual
Xavi
Giró
Miriam
Bellver
Janna
Escur
Carles
Ventura
Miquel
Tubau
Paula
Gómez
Benet
Oriol
Mariona
Carós
Jordi
Torres

225. bit.ly/MMM2019
@DocXavi
@DocXavi
Xavier Giro-i-Nieto
Slides available in 24 hours at:
http://bit.ly/mmm2019-xavigiro
xavier.giro@upc.edu
#MMM2019
Suggestions for improvement
this tutorial (refs, study cases…) ?