Michael Auli

Principal Research Scientist/Director
Facebook AI Research
Menlo Park, CA
1st.last@gmail.com
  Google Scholar
 LinkedIn profile
 Twitter

I am a Principal Research Scientist/Director at Facebook AI Research in Menlo Park where I work on speech processing and NLP which resulted in projects such as wav2vec, the fairseq toolkit, the first modern convolutional seq2seq models outperforming RNNs, as well as top ranked submissions at the WMT news translation task in 2018 and 2019. Before that I was at Microsoft Research, where I did early work on neural machine translation and neural dialogue models. I earned my Ph.D. at the University of Edinburgh where I was advised by Adam Lopez and Philipp Koehn.

News

• We released data2vec, a single self-supervised learning algorithm which achieves high peformance for vision, speech and language.
• XLS-R provides self-supervised speech representations for 128 languages and sets a new state of the art for speech translation, language identification and several ASR benchmarks.
• wav2vec-U enables speech recognition performance competitive to the best systems trained on 960h of labeled data from only two years ago and was accepted for oral presentation at NeurIPS 2021.
• I recently gave a talk about wav2vec at MIT, CMU and the University of Edinburgh.
• wav2vec 2.0 enables speech recognition systems using just 10 minutes of transcribed data. Applying it to cross-lingual training gives nice results on CommonVoice and BABEL.

Selected Papers (See Google Scholar for full list)

Efficient Self-supervised Learning with Contextualized Target Representations for Vision, Speech and Language
Alexei Baevski, Arun Babu, Wei-Ning Hsu, Michael Auli. In arXiv, 2022.
Abstract

×

Efficient Self-supervised Learning with Contextualized Target Representations for Vision, Speech and Language

Current self-supervised learning algorithms are often modality-specific and require large amounts of computational resources. To address these issues, we increase the training efficiency of data2vec, a learning objective that generalizes across several modalities. We do not encode masked tokens, use a fast convolutional decoder and amortize the effort to build teacher representations. data2vec 2.0 benefits from the rich contextualized target representations introduced in data2vec which enable a fast self-supervised learner. Experiments on ImageNet-1K image classification show that data2vec 2.0 matches the accuracy of Masked Autoencoders in 16.4x lower pre-training time, on Librispeech speech recognition it performs as well as wav2vec 2.0 in 10.6x less time, and on GLUE natural language understanding it matches a retrained RoBERTa model in half the time. Trading some speed for accuracy results in ImageNet-1K top-1 accuracy of 86.8\% with a ViT-L model trained for 150 epochs.

Blog Code

data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language
Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli. In arXiv, 2022.
Abstract

×

data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language

While the general idea of self-supervised learning is identical across modalities, the actual algorithms and objectives differ widely because they were developed with a single modality in mind. To get us closer to general self-supervised learning, we present data2vec, a framework that uses the same learning method for either speech,NLP or computer vision. The core idea is to predict latent representations of the full input data based on a masked view of the input in a selfdistillation setup using a standard Transformer architecture. Instead of predicting modality-specific targets such as words, visual tokens or units of human speech which are local in nature, data2vec predicts contextualized latent representations that contain information from the entire input. Experiments on the major benchmarks of speech recognition, image classification, and natural language understanding demonstrate a new state of the art or competitive performance to predominant approaches.

Blog Code

XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale
Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli. In arXiv, 2021.
Abstract

×

XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale

This paper presents XLS-R, a large-scale model for cross-lingual speech representation learning based on wav2vec 2.0. We train models with up to 2B parameters on nearly half a million hours of publicly available speech audio in 128 languages, an order of magnitude more public data than the largest known prior work. Our evaluation covers a wide range of tasks, domains, data regimes and languages, both high and low-resource. On the CoVoST-2 speech translation benchmark, we improve the previous state of the art by an average of 7.4 BLEU over 21 translation directions into English. For speech recognition, XLS-R improves over the best known prior work on BABEL, MLS, CommonVoice as well as VoxPopuli, lowering error rates by 14-34% relative on average. XLS-R also sets a new state of the art on VoxLingua107 language identification. Moreover, we show that with sufficient model size, cross-lingual pretraining can outperform English-only pretraining when translating English speech into other languages, a setting which favors monolingual pretraining. We hope XLS-R can help to improve speech processing tasks for many more languages of the world.

Blog Code

Unsupervised Speech Recognition
Alexei Baevski, Wei-Ning Hsu, Alexis Conneau, Michael Auli. In Proc. of NeurIPS, 2021.
Abstract

×

Unsupervised Speech Recognition

Despite rapid progress in the recent past, current speech recognition systems still require labeled training data which limits this technology to a small fraction of the languages spoken around the globe. This paper describes wav2vec-U, short for wav2vec Unsupervised, a method to train speech recognition models without any labeled data. We leverage self-supervised speech representations to segment unlabeled audio and learn a mapping from these representations to phonemes via adversarial training. The right representations are key to the success of our method. Compared to the best previous unsupervised work, wav2vec-U reduces the phoneme error rate on the TIMIT benchmark from 26.1 to 11.3. On the larger English Librispeech benchmark, wav2vec-U achieves a word error rate of 5.9 on test-other, rivaling some of the best published systems trained on 960 hours of labeled data from only two years ago. We also experiment on nine other languages, including low-resource languages such as Kyrgyz, Swahili and Tatar.

Blog Code

Beyond english-centric multilingual machine translation
Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli*, Armand Joulin*. In JMLR, 2020.
Abstract

×

Beyond english-centric multilingual machine translation

Existing work in translation demonstrated the potential of massively multilingual machine translation by training a single model able to translate between any pair of languages. However, much of this work is English-Centric by training only on data which was translated from or to English. While this is supported by large sources of training data, it does not reflect translation needs worldwide. In this work, we create a true Many-to-Many multilingual translation model that can translate directly between any pair of 100 languages. We build and open source a training dataset that covers thousands of language directions with supervised data, created through large-scale mining. Then, we explore how to effectively increase model capacity through a combination of dense scaling and language-specific sparse parameters to create high quality models. Our focus on non-English-Centric models brings gains of more than 10 BLEU when directly translating between non-English directions while performing competitively to the best single systems of WMT. We open-source our scripts so that others may reproduce the data, evaluation, and final M2M-100 model.

Blog Code

wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations
Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli. In NeurIPS, 2020.
Abstract

×

wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations

We show for the first time that learning powerful representations from speech audio alone followed by fine-tuning on transcribed speech can outperform the best semi-supervised methods while being conceptually simpler. wav2vec 2.0 masks the speech input in the latent space and solves a contrastive task defined over a quantization of the latent representations which are jointly learned. We set a new state of the art on both the 100 hour subset of Librispeech as well as on TIMIT phoneme recognition. When lowering the amount of labeled data to one hour, our model outperforms the previous state of the art on the 100 hour subset while using 100 times less labeled data. Using just ten minutes of labeled data and pre-training on 53k hours of unlabeled data still achieves 5.7/10.1 WER on the noisy/clean test sets of Librispeech. This demonstrates the feasibility of speech recognition with limited amounts of labeled data. Fine-tuning on all of Librispeech achieves 1.9/3.5 WER using a simple baseline model architecture. We will release code and models.

Blog Code

wav2vec: Unsupervised Pre-training for Speech Recognition
Steffen Schneider, Alexei Baevski, Ronan Collobert, Michael Auli. In Proc. of Interspeech, 2019.
Abstract

×

wav2vec: Unsupervised Pre-training for Speech Recognition

We explore unsupervised pre-training for speech recognition by learning representations of raw audio. wav2vec is trained on large amounts of unlabeled audio data and the resulting representations are then used to improve acoustic model training. We pre-train a simple multi-layer convolutional neural network optimized via a noise contrastive binary classification task. Our experiments on WSJ reduce WER of a strong character-based log-mel filterbank baseline by up to 32% when only a few hours of transcribed data is available. Our approach achieves 2.43% WER on the nov92 test set. This outperforms Deep Speech 2, the best reported character-based system in the literature while using three orders of magnitude less labeled training data.

Blog Code

fairseq: A fast, extensible toolkit for sequence modeling
Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli. In Proc. of NAACL, Demonstrations, 2019.
Abstract

×

fairseq: A fast, extensible toolkit for sequence modeling

fairseq is an open-source sequence modeling toolkit that allows researchers and developers to train custom models for translation, summarization, language modeling, and other text generation tasks. The toolkit is based on PyTorch and supports distributed training across multiple GPUs and machines. We also support fast mixed-precision training and inference on modern GPUs. A demo video can be found here: https://www.youtube.com/watch?v=OtgDdWtHvto

Code

Pay Less Attention with Lightweight and Dynamic Convolutions
Felix Wu, Angela Fan, Alexei Baevski, Yann N Dauphin, Michael Auli. In Proc. of ICLR, 2019.
Abstract

×

Pay Less Attention with Lightweight and Dynamic Convolutions

Self-attention is a useful mechanism to build generative models for language and images. It determines the importance of context elements by comparing each element to the current time step. In this paper, we show that a very lightweight convolution can perform competitively to the best reported self-attention results. Next, we introduce dynamic convolutions which are simpler and more efficient than self-attention. We predict separate convolution kernels based solely on the current time-step in order to determine the importance of context elements. The number of operations required by this approach scales linearly in the input length, whereas self-attention is quadratic. Experiments on large-scale machine translation, language modeling and abstractive summarization show that dynamic convolutions improve over strong self-attention models. On the WMT'14 English-German test set dynamic convolutions achieve a new state of the art of 29.7 BLEU.

Code

Understanding Back-Translation at Scale
Sergey Edunov, Myle Ott, David Grangier, Michael Auli. In Proc. of EMNLP, 2018.
Abstract

×

Understanding Back-Translation at Scale

An effective method to improve neural machine translation with monolingual data is to augment the parallel training corpus with back-translations of target language sentences. This work broadens the understanding of back-translation and investigates a number of methods to generate synthetic source sentences. We find that in all but resource poor settings back-translations obtained via sampling or noised beam outputs are most effective. Our analysis shows that sampling or noisy synthetic data gives a much stronger training signal than data generated by beam or greedy search. We also compare how synthetic data compares to genuine bitext and study various domain effects. Finally, we scale to hundreds of millions of monolingual sentences and achieve a new state of the art of 35 BLEU on the WMT'14 English-German test set.

Code

Convolutional Sequence to Sequence Learning
Jonas Gehring, Michael Auli, David Grangier, Denis Yarats, Yann N. Dauphin. In Proc. of ICML, 2017.
Abstract

×

Convolutional Sequence to Sequence Learning

The prevalent approach to sequence to sequence learning maps an input sequence to a variable length output sequence via recurrent neural networks. We introduce an architecture based entirely on convolutional neural networks. Compared to recurrent models, computations over all elements can be fully parallelized during training and optimization is easier since the number of non-linearities is fixed and independent of the input length. Our use of gated linear units eases gradient propagation and we equip each decoder layer with a separate attention module. We outperform the accuracy of the deep LSTM setup of Wu et al. (2016) on both WMT'14 English-German and WMT'14 English-French translation at an order of magnitude faster speed, both on GPU and CPU.

Blog Code

Sequence Level Training with Recurrent Neural Networks
Marc'Aurelio Ranzato, Sumit Chopra, Michael Auli, and Wojciech Zaremba. In Proc. of ICLR, 2016.
Abstract

×

Sequence Level Training with Recurrent Neural Networks

Many natural language processing applications use language models to generate text. These models are typically trained to predict the next word in a sequence, given the previous words and some context such as an image. However, at test time the model is expected to generate the entire sequence from scratch. This discrepancy makes generation brittle, as errors may accumulate along the way. We address this issue by proposing a novel sequence level training algorithm that directly optimizes the metric used at test time, such as BLEU or ROUGE. On three different tasks, our approach outperforms several strong baselines for greedy generation. The method is also competitive when these baselines employ beam search, while being several times faster.

Code

Press

Talks

Unified self-supervised learning for speech, vision and NLP
Talk at ECCV Workshop on Perception

wav2vec: Self-supervised learning of speech representations
Talk at MIT, CMU, U of Edinburgh, Spring 2021.

Efficient Sequence Modeling
Talk at WNGT'19, Stanford, Berkeley, Nov 2019.

Sequence to Sequence Learning: Fast Training and Inference with Gated Convolutions
Talk at Johns Hopkins University, Oct 2017.

Learning to translate with neural networks
Talk at Facebook, Google, Amazon and the University of Washington, 2014.

Integrated Parsing and Tagging
Talk at Carnegie Mellon University, Johns Hopkins University, BBN Technologies, IBM Research and Microsoft Research, 2011.