Neural Information Processing Systems http://nips.cc/

We learn rich natural sound representations by capitalizing on large amounts of unlabeled sound data collected in the wild. We leverage the natural synchronization between vision and sound to learn an acoustic representation using two-million unlabeled videos. Unlabeled video has the advantage that it can be economically acquired at massive scales, yet contains useful signals about natural sound. We propose a student-teacher training procedure which transfers discriminative visual knowledge from well established visual recognition models into the sound modality using unlabeled video as a bridge. Our sound representation yields significant performance improvements over the state-of-the-art results on standard benchmarks for acoustic scene/object classification. Visualizations suggest some high-level semantics automatically emerge in the sound network, even though it is trained without ground truth labels.

This is a simple but interesting idea, and the results are promising. There are several points that could be improved: 1) What is the motivation for the particular architecture that was chosen? No other architecture is investigated, and this choice seems rather ad hoc. There exist many other neural networks that are being used in audio, whether for automatic speech recognition or for source separation, and I have yet to see such a network architecture. Why not simply use a DNN or LSTM on a frame-wise representation of the signal, e.g., STFT magnitude or log-magnitude?

We learn rich natural sound representations by capitalizing on large amounts of unlabeled sound data collected in the wild. We leverage the natural synchronization between vision and sound to learn an acoustic representation using two-million unlabeled videos. Unlabeled video has the advantage that it can be economically acquired at massive scales, yet contains useful signals about natural sound. We propose a student-teacher training procedure which transfers discriminative visual knowledge from well established visual recognition models into the sound modality using unlabeled video as a bridge. Our sound representation yields significant performance improvements over the state-of-the-art results on standard benchmarks for acoustic scene/object classification. Visualizations suggest some high-level semantics automatically emerge in the sound network, even though it is trained without ground truth labels.

We learn rich natural sound representations by capitalizing on large amounts of unlabeled sound data collected in the wild. We leverage the natural synchronization between vision and sound to learn an acoustic representation using two-million unlabeled videos. Unlabeled video has the advantage that it can be economically acquired at massive scales, yet contains useful signals about natural sound. We propose a student-teacher training procedure which transfers discriminative visual knowledge from well established visual recognition models into the sound modality using unlabeled video as a bridge. Our sound representation yields significant performance improvements over the state-of-the-art results on standard benchmarks for acoustic scene/object classification.

In this work, we train fully convolutional networks to detect anger in speech. Since training these deep architectures requires large amounts of data and the size of emotion datasets is relatively small, we use transfer learning. However, unlike previous approaches that use speech or emotion-based tasks for the source model, we instead use SoundNet, a fully convolutional neural network trained multimodally on a massive video dataset to classify audio, with ground-truth labels provided by vision-based classifiers. As a result of transfer learning from SoundNet, our trained anger detection model improves performance and generalizes well on a variety of acted, elicited, and natural emotional speech datasets. We also test the cross-lingual effectiveness of our model by evaluating our English-trained model on Mandarin Chinese speech emotion data. Furthermore, our proposed system has low latency suitable for real-time applications, only requiring 1.2 seconds of audio to make a reliable classification.

These interests in the soundscape and fantasizing AI led to my latest project, "Imaginary Soundscape". As I wrote, one can imagine scenes from a sound. Conversely, by taking a glance at a photo, we can imagine sounds we might hear if we were there. Can an AI system do the same? If so, what if we apply the method to images of Google Street View, so that we can walk around with the generated soundscape?

We learn rich natural sound representations by capitalizing on large amounts of unlabeled sound data collected in the wild. We leverage the natural synchronization between vision and sound to learn an acoustic representation using two-million unlabeled videos. Unlabeled video has the advantage that it can be economically acquired at massive scales, yet contains useful signals about natural sound. We propose a student-teacher training procedure which transfers discriminative visual knowledge from well established visual recognition models into the sound modality using unlabeled video as a bridge. Our sound representation yields significant performance improvements over the state-of-the-art results on standard benchmarks for acoustic scene/object classification. Visualizations suggest some high-level semantics automatically emerge in the sound network, even though it is trained without ground truth labels.