Approximate LSTM Computing for Energy-Efficient Speech Recognition
نویسندگان
چکیده
منابع مشابه
Approximate computing for energy-efficient error-resilient multimedia systems
The rapid advancement in scaled silicon technology has resulted in the influx of numerous consumer devices with a plethora of applications. Multimedia applications which use image and video processing, pattern or facial recognition, data mining and synthesis have seen a significant increase in user base. These applications not only demand complex signal processing of digital data to achieve qua...
متن کاملDeep LSTM for Large Vocabulary Continuous Speech Recognition
Recurrent neural networks (RNNs), especially long shortterm memory (LSTM) RNNs, are effective network for sequential task like speech recognition. Deeper LSTM models perform well on large vocabulary continuous speech recognition, because of their impressive learning ability. However, it is more difficult to train a deeper network. We introduce a training framework with layer-wise training and e...
متن کاملAuxiliary Multimodal LSTM for Audio-visual Speech Recognition and Lipreading
The Aduio-visual Speech Recognition (AVSR) which employs both the video and audio information to do Automatic Speech Recognition (ASR) is one of the application of multimodal leaning making ASR system more robust and accuracy. The traditional models usually treated AVSR as inference or projection but strict prior limits its ability. As the revival of deep learning, Deep Neural Networks (DNN) be...
متن کاملHighway-LSTM and Recurrent Highway Networks for Speech Recognition
Recently, very deep networks, with as many as hundreds of layers, have shown great success in image classification tasks. One key component that has enabled such deep models is the use of “skip connections”, including either residual or highway connections, to alleviate the vanishing and exploding gradient problems. While these connections have been explored for speech, they have mainly been ex...
متن کاملImprecise Minority-Based Full Adder for Approximate Computing Using CNFETs
Nowadays, the portable multimedia electronic devices, which employ signal-processing modules, require power aware structures more than ever. For the applications associating with human senses, approximate arithmetic circuits can be considered to improve performance and power efficiency. On the other hand, scaling has led to some limitations in performance of nanoscale circuits. According...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Electronics
سال: 2020
ISSN: 2079-9292
DOI: 10.3390/electronics9122004