How our scientists are making Alexa smarter
Amazon Alexa head scientist shares how his teams are working to improve customers' experiences with Echo devices. Read more
Research
We fundamentally believe that scientific innovation is essential to being the most customer-centric company in the world. It’s this ability to have an impact at scale that allows us to attract some of the brightest minds in artificial intelligence and related fields.

Our scientists continue to publish, teach, and engage with the academic community, but one of the unique aspects of research at Amazon is our working backwards method applied to science.

Each research project starts with a customer pain point and a description of how a scientific solution will improve the customer experience.

Expanding the Natural-Language Processing community

Amazon helps sponsor second annual Widening Natural-Language Processing workshop.

Where computer science and linguistics meet

Amazon researchers will co-present 12 papers at the North American Chapter of the Association for Computational Linguistics conference.
Publications

  • Natural language processing
    Transfer Learning for Neural Semantic Parsing
    Author: Xing Fan, Emilio Monti, Lambert Mathias, Markus Dreyer
    Conference: ACL 2017, Workshop "Representation Learning for NLP"
    Our research

  • Conversational AI
    On Evaluating and Comparing Conversational Agents
    Author: Anu Venkatesh, Chandra Khatri, Ashwin Ram, Fenfei Guo, Raefer Gabriel, Ashish Nagar, Rohit Prasad, Ming Cheng, Behnam Hedayatnia, Angeliki Metallinou, Rahul Goel, Shaohua Yang, Anirudh Raju
    Conference: NIPS 2017
    Our research

  • Conversational AI
    Scalable Distributed DNN Training Using Commodity GPU Cloud Computing
    Author: Nikko Strom
    Conference: Interspeech 2015
    Our research

  • Conversational AI
    Domain-Specific Utterance End-Point Detection for Speech Recognition
    Author: Roland Maas, Ariya Rastrow, Kyle Goehner, Gautam Tiwari, Shaun Joseph, Björn Hoffmeister
    Conference: Interspeech 2017, Stockholm
    Our research

  • Conversational AI
    Direct Modeling of Raw Audio with DNNS for Wake Word Detection
    Author: Kenichi Kumatani, Sankaran Panchapagesan, Minhua Wu, Minjae Kim, Nikko Strom, Gautam Tiwari, Arindam Mandal
    Conference: ASRU 2017
    Our research

  • Operations research
    Bayesian Intermittent Demand Forecasting for Large Inventories
    Author: Matthias Seeger, David Salinas, Valentin Flunkert
    Conference: NIPS 2016
    Our research

Teaching Alexa when not to respond

Real-time acoustic fingerprinting will prevent Alexa devices from waking when her name is called during our commercial in the big game.

Alexa's newest teacher is you

A team of Amazon linguists and data scientists makes it easy for you to teach your language or dialect to Alexa.

Making AI more accurate

Amazon and University of Sheffield researchers make fact extraction and verification dataset publicly available.

Amazon joins Partnership on AI

We’re in a golden age of Machine Learning and Artificial Intelligence (AI). As a scientific community, we are still a long way from being able to do things the way humans do things, but we’re solving unbelievably complex problems every day and making incredibly rapid progress.

Announcing the teams competing for the $3.5 million Alexa Prize

Teams will create socialbots that can converse with humans on a range of topics.

Amazon scientists present at conversational AI workshop

Workshop among the most popular at sold out NIPS Conference
How we pack a Prime plane
Each Amazon Prime Air plane can hold over one million
Echo Dots, 10,000 Instant Pots, or 5,000 teddy bears.
Check it out
AA Out for Delivery
Get the latest news and stories from Amazon, delivered right to your inbox.