Unpaired Image-to-Speech Synthesis With Multimodal Information Bottleneck

By Shuang Ma, Daniel J. McDuff, Yale Song. 2019

29 citations recorded by the source corpus.

Deep generative models have led to significant advances in cross-modal generation such as text-to-image synthesis. Training these models typically requires paired data with direct correspondence between modalities. We introduce the novel problem of translating instances from one modality to another without paired data by leveraging an intermediate modality shared by the two other modalities. To demonstrate this, we take the problem of translating images to speech. In this case, one could leverage disjoint datasets with one shared modality, e.g., image-text pairs and text-speech pairs, with text as the shared modality. We call this problem ``skip-modal generation'' because the shared modality is skipped during the generation process. We propose a multimodal information bottleneck approach that learns the correspondence between modalities from unpaired data (image and speech) by leveraging the shared modality (text). We address fundamental challenges of skip-modal generation: 1) learning multimodal representations using a single model, 2) bridging the domain gap between two unrelated datasets, and 3) learning the correspondence between modalities from unpaired data. We show qualitative results on image-to-speech synthesis; this is the first time such results have been reported in the literature. We also show that our approach improves performance on traditional cross-modal generation, suggesting that it improves data efficiency in solving individual tasks.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Generative Models, Multimodal Learning, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. David F. Harwath
2. James R. Glass
3. Wei-Ning Hsu
4. Jean-Baptiste Alayrac
5. Andrew Zisserman
6. Christopher Song
7. Relja Arandjelovi'c
8. Lucas Smaira

In Computer Vision and Pattern Recognition

1. James R. Glass
2. Andrew Zisserman
3. David F. Harwath
4. Luc Van Gool
5. Antonio Torralba
6. Cordelia Schmid
7. Kristen Grauman
8. Andrea Vedaldi

In Generative Models

1. Wei-Ning Hsu
2. David F. Harwath
3. James R. Glass
4. Christopher Song
5. Jeffrey De Fauw
6. Jason Ramapuram
7. Andrew Zisserman
8. Jean-Baptiste Alayrac

In Multimodal Learning

1. David F. Harwath
2. James R. Glass
3. Wei-Ning Hsu
4. Andrew Zisserman
5. Adrià Recasens
6. Christopher Song
7. Relja Arandjelovi'c
8. Jean-Baptiste Alayrac

In Signal Processing

1. James R. Glass
2. David F. Harwath
3. Wei-Ning Hsu
4. Andrew Zisserman
5. Christopher Song
6. Adrià Recasens
7. Jean-Baptiste Alayrac
8. Jeffrey De Fauw

Citations in the corpus

• Self-Supervised MultiModal Versatile Networks (2020, background): “Using audio, vision and language to learn representations has also been explored in past work [11, 31, 38, 50, 83].”
• Text-Free Image-to-Speech Synthesis Using Learned Segmental Units (2020, methodology): “[72] Shuang Ma, Daniel McDuff, and Yale Song.”