Baidu USA LLC

D2s, Inc.

Gauda, Inc.

Erich Elsen

Bryan Catanzaro

Awni Hannun

Shubhabrata Sengupta

Ryan Prenger

Sanjeev Satheesh

Jesse Engel

Dario Amodei

Ilhami H Torunoglu

Xu Han

Gregory Diamos

Ahmet Karakas

Jun Zhan

Zhenyao Zhu

Chong Wang

Christopher Fougner

Mike Chrzanowski

Jingdong Chen

Dani Yogatama

Adam Coates

Andrew Ng

Jared Casper

Carl Case

Sengupta Shubhabrata

Systems and methods for a multi-core optimized Recurrent Neural Network (RNN) architecture are disclosed. The various architectures affect communication and synchronization operations according to the Multi-Bulk-Synchronous-Parallel (MBSP) model for a given processor. The resulting family of network architectures, referred to as MBSP-RNNs, perform similarly to a conventional RNNs having the same number of parameters, but are substantially more efficient when mapped onto a modern general purpose processor. Due to the large gain in computational efficiency, for a fixed computational budget, MBSP-RNNs outperform RNNs at applications such as end-to-end speech recognition.

Systems and methods for a multi-core optimized recurrent neural network

Presented herein are embodiments of state-of-the-art speech recognition systems developed using end-to-end deep learning. In embodiments, the model architecture is significantly simpler than traditional speech systems, which rely on laboriously engineered processing pipelines; these traditional systems also tend to perform poorly when used in noisy environments. In contrast, embodiments of the system do not need hand-designed components to model background noise, reverberation, or speaker variation, but instead directly learn a function that is robust to such effects. A phoneme dictionary, nor even the concept of a 'phoneme,' is needed. Embodiments include a well-optimized recurrent neural network (RNN) training system that can use multiple GPUs, as well as a set of novel data synthesis techniques that allows for a large amount of varied data for training to be efficiently obtained. Embodiments of the system can also handle challenging noisy environments better than widely used, state-of-the-art commercial speech systems.

Systems and methods for speech transcription

Embodiments of end-to-end deep learning systems and methods are disclosed to recognize speech of vastly different languages, such as English or Mandarin Chinese. In embodiments, the entire pipelines of hand-engineered components are replaced with neural networks, and the end-to-end learning allows handling a diverse variety of speech including noisy environments, accents, and different languages. Using a trained embodiment and an embodiment of a batch dispatch technique with GPUs in a data center, an end-to-end deep learning system can be inexpensively deployed in an online setting, delivering low latency when serving users at scale.

End-to-end speech recognition

Deployed end-to-end speech recognition

Computationally intensive electronic design automation operations are accelerated with algorithms utilizing one or more graphics processing units. The optical proximity correction (OPC) process calculates, improves, and optimizes one or more features on an exposure mask (used in semiconductor or other processing) so that a resulting structure realized on an integrated circuit or chip meets desired design and performance requirements. When a chip has billions of transistors or more, each with many fine structures, the computational requirements for OPC can be very large. This processing can be accelerated using one or more graphics processing units.

Techniques of optical proximity correction using GPU

Growing community of inventors

Mountain View, CA, United States of America

Erich Elsen