Deep learning

Domain consensus clustering for universal domain adaptation

Guangrui Li, Guoliang Kang, Yi Zhu, Yunchao Wei, Yi Yang

CVPR 2021

2021

In this paper, we investigate Universal Domain Adaptation (UniDA) problem, which aims to transfer the knowledge from source to target under unaligned label space. The main challenge of UniDA lies in how to separate common classes (i.e., classes shared across domains), from private classes (i.e., classes only exist in one domain). Previous works treat the private samples in the target as one generic class

Computer vision

Multi-task self-supervised pre-training for music classification

Ho-Hsiang Wu, Chieh-Chi Kao, Qingming Tang, Ming Sun, Brian McFee, Juan Pablo Bello, Chao Wang

ICASSP 2021

2021

Deep learning is very data hungry, and supervised learning especially requires massive labeled data to work well. Machine listening research often suffers from limited labeled data problem, as human annotations are costly to acquire, and annotations for audio are time consuming and less intuitive. Besides, models learned from labeled dataset often embed biases specific to that particular dataset. Therefore

Machine learning

Bipartite dynamic representations for abuse detection

Andrew Z Wang, Rex Ying, Pan Li, Nikhil Rao, Karthik Subbian, Jure Leskovec

KDD 2021

2021

Abusive behavior in online retail websites and communities threatens the experience of regular community members. Such behavior often takes place within a complex, dynamic, and large-scale network of users interacting with items. Detecting abuse is challenging due to the scarcity of labeled abuse instances and complexity of combining temporal and network patterns while operating at a massive scale. Previous

Security, privacy, and abuse prevention

Dive into Deep Learning adds attention mechanism chapter

Douglas Gantenbein

March 02, 2021

The newest chapter addresses a problem that often bedevils nonparametric machine learning models.

Machine learning

Listen with intent: Improving speech recognition with audio-to-intent front-end

Swayambhu Nath Ray, Minhua Wu, Anirudh Raju, Pegah Ghahremani, Raghavendra Bilgi, Milind Rao, Harish Arsikere, Ariya Rastrow, Andreas Stolcke, Jasha Droppo

Interspeech 2021

2021

Comprehending the overall intent of an utterance helps a listener recognize the individual words spoken. Inspired by this fact, we perform a novel study of the impact of explicitly incorporating intent representations as additional information to improve a recurrent neural network-transducer (RNN-T) based automatic speech recognition (ASR) system. An audio-to-intent (A2I) model encodes the intent of the

Machine learning

Factorization-aware training of transformers for natural language understanding on the edge

Hamidreza Saghir, Samridhi Choudhary, Sepehr Eghbali, Clement Chung

Interspeech 2021

2021

Fine-tuning transformer-based models have shown to outperform other methods for many Natural Language Understanding (NLU) tasks. Recent studies to reduce the size of transformer models have achieved reductions of > 80%, making on-device inference on powerful devices possible. However, other resource-constrained devices, like those enabling voice assistants (VAs), require much further reductions. In this

Conversational AI

Improving the expressiveness of neural vocoding with non-affine normalizing flows

Adam Gabrys, Yunlong Jiao, Daniel Korzekwa, Roberto Barra-Chicote

Interspeech 2021

2021

This paper proposes a general enhancement to the Normalizing Flows (NF) used in neural vocoding. As a case study, we improve expressive speech vocoding with a revamped Parallel Wavenet (PW). Specifically, we propose to extend the affine transformation of PW to the more expressive invertible nonaffine function. The greater expressiveness of the improved PW leads to better-perceived signal quality and naturalness

Conversational AI

Multi-channel transformer transducer for speech recognition

Feng-Ju (Claire) Chang, Martin Radfar, Thanasis Mouchtaris, Maurizio Omologo

Interspeech 2021

2021

Multi-channel inputs offer several advantages over singlechannel, to improve the robustness of on-device speech recognition systems. Recent work on multi-channel transformer, has proposed a way to incorporate such inputs into end-to-end ASR for improved accuracy. However, this approach is characterized by a high computational complexity, which prevents it from being deployed in on-device systems. In this

Conversational AI

End-to-end spoken language understanding for generalized voice assistants

Michael Saxon, Samridhi Choudhary, Joseph McKenna, Thanasis Mouchtaris

Interspeech 2021

2021

End-to-end (E2E) spoken language understanding (SLU) systems predict utterance semantics directly from speech using a single model. Previous work in this area has focused on targeted tasks in fixed domains, where the output semantic structure is assumed a priori and the input speech is of limited complexity. In this work we present our approach to developing an E2E model for generalized SLU in commercial

Conversational AI

Weakly-supervised word-level pronunciation error detection in non-native English speech

Daniel Korzekwa, Jaime Lorenzo Trueba, Thomas Drugman, Shira Calamaro, Bozena Kostek

Interspeech 2021

2021

We propose a weakly-supervised model for word-level mispronunciation detection in non-native (L2) English speech. To train this model, phonetically transcribed L2 speech is not required and we only need to mark mispronounced words. The lack of phonetic transcriptions for L2 speech means that the model has to learn only from a weak signal of word-level mispronunciations. Because of that and due to the limited

Machine learning

Deep learning

Work with us