Amazon paper uses metric differential privacy to guard privacy in text analysesamazon.science·6 pts·georgecarlyle76·0
Alexa AI's head scientist on Alexa's 5th b-day: how we got here, where we'll godeveloper.amazon.com·5 pts·georgecarlyle76·0
Algorithms for selecting training data improve cross-lingual transfer learningdeveloper.amazon.com·5 pts·georgecarlyle76·0
The Fever Data Set: What Doesn’t Kill It Will Make It Strongerdeveloper.amazon.com·1 pts·georgecarlyle76·0
Synthetic-Data-Generating Tools Helped Bootstrap Alexa’s New-Language Releasesdeveloper.amazon.com·2 pts·georgecarlyle76·0
Amazon releases public data set to aid research on speech separationdeveloper.amazon.com·4 pts·georgecarlyle76·1
How to Construct the Optimal Neural Architecture for Your Machine Learning Taskdeveloper.amazon.com·2 pts·georgecarlyle76·0
Amazon Releases Data Set of Annotated Dialogues to Aid Development of Socialbotsdeveloper.amazon.com·4 pts·georgecarlyle76·0
Turning Dialogue Tracking into a Reading Comprehension Problemdeveloper.amazon.com·5 pts·georgecarlyle76·0
Summaries of the 16 Alexa-related papers at this year's Interspeechdeveloper.amazon.com·4 pts·georgecarlyle76·0
How to Make Neural Language Models Practical for Speech Recognitiondeveloper.amazon.com·1 pts·georgecarlyle76·0
Amazon transfers inflection, rhythm from recorded speaker to synthesized voicedeveloper.amazon.com·6 pts·georgecarlyle76·0
Papers show how to make Alexa's neural models efficient enough to run on-devicedeveloper.amazon.com·6 pts·georgecarlyle76·0
ML architecture enables training of related Alexa skills on pooled datadeveloper.amazon.com·6 pts·georgecarlyle76·0
Alexa team wins best-paper award for work on tracking referring terms in dialogsdeveloper.amazon.com·5 pts·georgecarlyle76·0
For complex QA, system stitches search results into ad hoc knowledge graphsdeveloper.amazon.com·4 pts·georgecarlyle76·0
Alexa researcher pairs headlines with associated texts to train neural searchdeveloper.amazon.com·4 pts·georgecarlyle76·0
Alexa scientist: How to do fast, accurate multi-category classificationdeveloper.amazon.com·3 pts·georgecarlyle76·0
Alexa's take on “active learning”, or automatically selecting ML training datadeveloper.amazon.com·7 pts·georgecarlyle76·0
Alexa's new approach to tracking referring terms during dialoguesdeveloper.amazon.com·8 pts·georgecarlyle76·0
Amazon Unveils New Alexa Dialog Modeling for Natural, Cross-Skill Conversationsdeveloper.amazon.com·103 pts·georgecarlyle76·31
An Alexa scientist on how to detect speakers' emotions from speech signalsdeveloper.amazon.com·5 pts·georgecarlyle76·0
Should Alexa read “2/3” as “two-thirds” or “February 3”?: How to normalize textdeveloper.amazon.com·5 pts·georgecarlyle76·0
Training a machine learning model in English improves performance in Japanesedeveloper.amazon.com·4 pts·georgecarlyle76·0