This year's Interspeech will be held in Graz, Austria, whose famed clock tower was built in the mid-1500s
Photo courtesy of Getty Images

The 16 Alexa-related papers at this year’s Interspeech

At next week’s Interspeech, the largest conference on the science and technology of spoken-language processing, Alexa researchers have 16 papers, which span the five core areas of Alexa functionality: device activation, or recognizing speech intended for Alexa and other audio events that require processing; automatic speech recognition (ASR), or converting the speech signal into text; natural-language understanding, or determining the meaning of customer utterances; dialogue management, or handling multiturn conversational exchanges; and text-to-speech, or generating natural-sounding synthetic speech to convey Alexa’s responses. Two of the papers are also more-general explorations of topics in machine learning.

Device Activation

Model Compression on Acoustic Event Detection with Quantized Distillation
Bowen Shi, Ming Sun, Chieh-Chi Kao, Viktor Rozgic, Spyros Matsoukas, Chao Wang

The researchers combine two techniques to shrink neural networks trained to detect sounds by 88%, with no loss in accuracy. One technique, distillation, involves using a large, powerful model to train a leaner, more-efficient one. The other technique, quantization, involves using a fixed number of values to approximate a larger range of values.

Sub-band Convolutional Neural Networks for Small-footprint Spoken Term Classification
Chieh-Chi Kao, Ming Sun, Yixin Gao, Shiv Vitaladevuni, Chao Wang

Convolutional neural nets (CNNs) were originally designed to look for the same patterns in every block of pixels in a digital image. But they can also be applied to acoustic signals, which can be represented as two-dimensional mappings of time against frequency-based “features”. By restricting an audio-processing CNN’s search only to the feature ranges where a particular pattern is likely to occur, the researchers make it much more computationally efficient. This could make audio processing more practical for power-constrained devices.

A Study for Improving Device-Directed Speech Detection toward Frictionless Human-Machine Interaction
Che-Wei Huang, Roland Maas, Sri Harish Mallidi, Björn Hoffmeister

This paper is an update of prior work on detecting device-directed speech, or identifying utterances intended for Alexa. The researchers find that labeling dialogue turns (distinguishing initial utterances from subsequent utterances) and using signal representations based on Fourier transforms rather than mel-frequencies improve accuracy. They also find that, among the features extracted from speech recognizers that the system considers, confusion networks, which represent word probabilities at successive sentence positions, have the most predictive power.

Automatic Speech Recognition (ASR)

Acoustic Model Bootstrapping Using Semi-Supervised Learning
Langzhou Chen, Volker Leutnant

The researchers propose a method for selecting machine-labeled utterances for semi-supervised training of an acoustic model, the component of an ASR system that takes an acoustic signal as input. First, for each training sample, the system uses the existing acoustic model to identify the two most probable word-level interpretations of the signal at each position in the sentence. Then it finds examples in the training data that either support or contradict those probability estimates, which it uses to adjust the uncertainty of the ASR output. Samples that yield significant reductions in uncertainty are preferentially selected for training.

Improving ASR Confidence Scores for Alexa Using Acoustic and Hypothesis Embeddings
Prakhar Swarup, Roland Maas, Sri Garimella, Sri Harish Mallidi, Björn Hoffmeister

Speech recognizers assign probabilities to different interpretations of acoustic signals, and these probabilities can serve as inputs to a machine learning model that assesses the recognizer’s confidence in its classifications. The resulting confidence scores can be useful to other applications, such as systems that select machine-labeled training data for semi-supervised learning. The researchers append embeddings — fixed-length vector representations — of both the raw acoustic input and the speech recognizer’s best estimate of the word sequence to the inputs to a confidence-scoring network. The result: a 6.5% reduction in equal-error rate (the error rate that results when the false-negative and false-positive rates are set as equal).

Multi-Dialect Acoustic Modeling Using Phone Mapping and Online I-Vectors
Harish Arsikere, Ashtosh Sapru, Sri Garimella

Multi-dialect acoustic models, which help convert multi-dialect speech signals to words, are typically neural networks trained on pooled multi-dialect data, with separate output layers for each dialect. The researchers show that mapping the phones — the smallest phonetic units of speech — of each dialect to those of the others offers comparable results with shorter training times and better parameter sharing. They also show that recognition accuracy can be improved by adapting multi-dialect acoustic models, on the fly, to a target speaker.

Neural Machine Translation for Multilingual Grapheme-to-Phoneme Conversion
Alex Sokolov, Tracy Rohlin, Ariya Rastrow

Grapheme-to-phoneme models, which translate written words into their phonetic equivalents (“echo” to “E k oU”), enable speech recognizers to handle words they haven’t seen before. The researchers train a single neural model to handle grapheme-to-phoneme conversion in 18 languages. The results are comparable to those of state-of-the-art single-language models for languages with abundant training data and better for languages with sparse data. Multilingual models are more flexible and easier to maintain in production environments.

Scalable Multi Corpora Neural Language Models for ASR
Anirudh Raju, Denis Filimonov, Gautam Tiwari, Guitang Lan, Ariya Rastrow

Language models, which compute the probability of a given sequence of words, help distinguish between different interpretations of speech signals. Neural language models promise greater accuracy than existing models, but they’re difficult to incorporate into real-time speech recognition systems. The researchers describe several techniques to make neural language models practical, from a technique for weighting training samples from out-of-domain data sets to noise contrastive estimation, which turns the calculation of massive probability distributions into simple binary decisions.

Natural-Language Understanding

Neural Named Entity Recognition from Subword Units
Abdalghani Abujabal, Judith Gaspers

Named-entity recognition is crucial to voice-controlled systems — as when you tell Alexa “Play ‘Spirit’ by Beyoncé”. A neural network that recognizes named entities typically has dedicated input channels for every word in its vocabulary. This has two drawbacks: (1) the network grows extremely large, which makes it slower and more memory intensive, and (2) it has trouble handling unfamiliar words. The researchers trained a named-entity recognizer that instead takes subword units — characters, phonemes, and bytes — as inputs. It offers comparable performance with a vocabulary of only 332 subwords, versus 74,000-odd words.

Dialogue Management

HyST: A Hybrid Approach for Flexible and Accurate Dialogue State Tracking
Rahul Goel, Shachi Paul, Dilek Hakkani-Tür

Dialogue-based computer systems need to track “slots” — types of entities mentioned in conversation, such as movie names — and their values — such as Avengers: Endgame. Training a machine learning system to decide whether to pull candidate slot values from prior conversation or compute a distribution over all possible slot values improves slot-tracking accuracy by 24% over the best-performing previous system.

Towards Universal Dialogue Act Tagging for Task-Oriented Dialogues
Shachi Paul, Rahul Goel, Dilek Hakkani-Tür

Dialogue-based computer systems typically classify utterances by “dialogue act” — such as requesting, informing, and denying — as a way of gauging progress toward a conversational goal. As a first step in developing a system that will automatically label dialogue acts in human-human conversations (to, in turn, train a dialogue-act classifier), the researchers create a “universal tagging scheme” for dialogue acts. They use this scheme to reconcile the disparate tags used in different data sets.

Topical-Chat: Towards Knowledge-Grounded Open-Domain Conversations
Karthik Gopalakrishnan, Behnam Hedayatnia, Qinlang Chen, Anna Gottardi, Sanjeev Kwatra, Anu Venkatesh, Raefer Gabriel, Dilek Hakkani-Tür

The researchers report a new data set, which grew out of the Alexa Prize competition and is intended to advance research on AI agents that engage in social conversations. Pairs of workers recruited through Mechanical Turk were given information on topics that arose frequently during Alexa Prize interactions and asked to converse about them, documenting the sources of their factual assertions. The researchers used the resulting data set to train a knowledge-grounded response generation network, and they report automated and human evaluation results as state-of-the-art baselines.


Towards Achieving Robust Universal Neural Vocoding
Jaime Lorenzo Trueba, Thomas Drugman, Javier Latorre, Thomas Merritt, Bartosz Putrycz, Roberto Barra-Chicote, Alexis Moinet, Vatsal Aggarwal

A vocoder is the component of a speech synthesizer that takes the frequency-spectrum snapshots generated by other components and fills in the information necessary to convert them to audio. The researchers trained a neural-network-based vocoder using data from 74 speakers of both genders in 17 languages. The resulting “universal vocoder” outperformed speaker-specific vocoders, even on speakers and languages it had never encountered before and unusual tasks such as synthesized singing.

Fine-Grained Robust Prosody Transfer for Single-Speaker Neural Text-to-Speech
Viacheslav Klimkov, Srikanth Ronanki, Jonas Rohnke, Thomas Drugman

The researchers present a new technique for transferring prosody (intonation, stress, and rhythm) from a recording to a synthesized voice, enabling the user to choose whose voice will read recorded content, with inflections preserved. Where earlier prosody transfer systems used spectrograms — frequency spectrum snapshots — as inputs, the researchers’ system uses easily normalized prosodic features extracted from the raw audio.

Machine Learning

Two Tiered Distributed Training Algorithm for Acoustic Modeling
Pranav Ladkat, Oleg Rybakov, Radhika Arava, Sree Hari Krishnan Parthasarathi,I-Fan Chen, Nikko Strom

When neural networks are trained on large data sets, the training needs to be distributed, or broken up across multiple processors. A novel combination of two state-of-the-art distributed-learning algorithms — GTC and BMUF — achieves both higher accuracy and more-efficient training then either, when learning is distributed to 128 parallel processors.

The researchers' new method splits distributed processors into groups, and within each group, the processors use the highly accurate GTC method to synchronize their models. At regular intervals, designated representatives from all the groups use a different method — BMUF — to share their models and update them accordingly. Finally, each representative broadcasts its updated model to the rest of its group.
Animation by Nick Little

One-vs-All Models for Asynchronous Training: An Empirical Analysis
Rahul Gupta, Aman Alok, Shankar Ananthakrishnan

A neural network can be trained to perform multiple classifications at once: it might recognize multiple objects in an image, or assign multiple topic categories to a single news article. An alternative is to train a separate “one-versus-all” (OVA) classifier for each category, which classifies data as either in the category or out of it. The advantage of this approach is that each OVA classifier can be re-trained separately as new data becomes available. The researchers present a new metric that enables comparison of multiclass and OVA strategies, to help data scientists determine which is more useful for a given application.

About the Author
Larry Hardesty is a science writer at Amazon. Previously, he was managing editor of the Boston Book Review, a senior editor at MIT Technology Review, and the computer science writer at the MIT News Office.

Work with us

See More Jobs
US, WA, Seattle
Prime Video is shaping the future of digital video entertainment. Our mission is to build the widest selection of digital video content and make it easy for customers to stream wherever and whenever they want. The Video Personalization team’s mission is to give customers individual Video recommendations, to surface the titles that they will most enjoy watching.As a Data Scientist in the Video Personalization team, you will play a vital role in understanding how to optimize the customer experience, and how to measure and recognize long term customer value. Your insights will drive what the Personalization product engineering team invests in, and help set and measure annual goals.In this role, you will own measurement strategy, weblab A/B training and office hours, performance tracking, whilst developing new insightful metrics and scaling the existing customer ML models that help define success within online A/B tests. You will work closely with the director, senior managers, product managers, engineers and ML specialists in the Personalization team, as well as people across Video and Amazon in finance, merchandising and economist teams.The successful candidate for this position will understand customers and business problems, and being a skilled modeler will be able to go complex or simple as appropriate. The role will include:- Using R, Python, Sagemaker/Eider, Redshift to analysis and model huge quantities of data- Work with product leaders to understand and reshape their customer problems and deliver innovative solutions- Own a problem space end to end and deliver to agreed timelines- Deep dive experiments to better shape the next set of treatments- Create new metrics that will help guide and evaluation specific customer experience.- Understand how best to cohort customers to track and optimize customer journeys.
US, CA, Sunnyvale
Are you inspired by invention? Is problem solving through teamwork in your DNA? Do you like the idea of seeing how your work impacts the bigger picture? Answer yes to any of these and you’ll fit right in here at Amazon. We are a smart team of doers that work passionately to apply cutting edge advances in robotics and software to solve real-world challenges that will transform our customers’ experiences in ways we can’t even imagine yet.As a Manager, Applied Science, you will wear many different hats and work on multiple components of the entire system, in a highly collaborative environment that’s more startup than big company. You will tackle problems that span a variety of domains: computer vision, image recognition, machine learning, real-time and distributed systems.You will also help in defining the creative vision, product requirements, and user experience for this robotics program. Working with a cross-disciplinary team of analysts, designers, hardware engineers, and researchers, you will define and drive the product from concept all the way through to production.You should be a self-starter with a bias towards independent problem solving. Clear communication and prioritization will be important as you plan, design, and deliver the best experience for millions of customers. Your passion for the potential of using technology to improve people’s lives, and your experience leading complex technology projects will help you make strong business judgments.If you’re entrepreneurial and want to build and own transformative technology-driven products, join us in making history.As a part of this role, you will:· Lead the design and implementation of a cutting-edge technology in the Robotics space, focusing on SLAM.· Build and manage a team of Scientists.· Foster career growth and a strong team culture.· Recruit, hire, mentor, and coach technical staff.· Interface with our internal customers to understand requirements, set priorities and communicate direction and progress.· Own all operational metrics and support for your team.· Manage the agile development process and methodology to deliver incremental value to customers.· Help develop long-term roadmaps and business technology strategies.
US, WA, Seattle
Ever wonder how you can keep the world’s largest selection also the world’s safest and legally compliant selection? Then come join a team with the charter to monitor and classify the billions of items in the Amazon catalog to ensure compliance with various legal regulations.Our mission is to enable business users to self-serve build, deploy, maintain and improve prod cut classifiers at Amazon catalog scale. To enable this mission, one needs an ecosystem that proactively interacts and assists its users in delivering a robust ML solution. A self-serve product classification system that proactively engages with its user in a near real-time fashion by providing intuitive “insights and biases” about a model’s learning and seeks user feedback enabling continuous re-learning and correction.In this role, you will have an opportunity to lead state of the art machine learning algorithms on large datasets. You will need to lead & build Amazon scale applications running on Amazon Cloud that both leverage and create new technologies to process large volumes of data that derive patterns and conclusions from the data.We are seeking an Applied Science Manager who has a solid background in applied Machine Learning and AI, deep passion for building data-driven products; ability to communicate data insights and scientific vision, and has a proven track record of leading both applied scientists and software engineers to execute complex projects and deliver business impacts.In this role, you will:· Lead a group applied scientists (predominantly) and software engineers to deliver machine-learning and AI solutions to production.· Advance team's engineering craftsmanship and drive continued scientific innovation as a thought leader and practitioner.· Develop science and engineering roadmap, run Sprint/quarter and annual planning, and foster cross-team collaboration to execute complex projects.· Perform hands-on data analysis, build machine-learning models, run regular A/B tests, and communicate the impact to senior management.· Hire and develop top talents, provide technical and career development guidance to both scientists and engineers in the organization.
US, WA, Seattle
Are you passionate about conducting measurement research and experiments to assess and evaluate talent? Would you like to see your research in products that will drive key talent management behaviors globally to ensure we are raising the bar on our talent? If so, you should consider joining the Global Talent Management (GTM) Science Team!Amazon GTM Science team is an innovative organization that exists to propel Amazon HR toward being the most scientific HR organization on earth. The GTM Science mission is to use Science to assist and measurably improve every talent decision made at Amazon. GTM Science does this by discovering signals in workforce data, infusing intelligence into Amazon’s talent products, and guiding the broader GTM team to pursue high-impact opportunities with tangible returns. This multi-disciplinary approach spans capabilities, including: data engineering, reporting and analytics, research and behavioral sciences, and applied sciences such as economics and machine learning.We are seeking Assessment and Measurement Scientists with expertise developing assessment and validating measures (assessments, performance evaluations, and surveys) to evaluate talent at Amazon. This person will possess knowledge of different assessment approaches to evaluate performance, a strong psychometrics background, scientific survey methodology, computing various content validity analyses, and experience developing legally defensible talent evaluation programs. In this role you will:· Lead the global research strategy developing performance evaluations both quantitative and qualitative· Execute a scalable global content development and research strategy Amazon-wide· Conduct psychometrics analyses to evaluate integrity and practical application of content· Identify research streams to evaluate how to mitigate or remove sources of measurement error· Partner closely and drive effective collaborations across multi-disciplinary research and product teams· Manage full life cycle of large scale research programs (Develop strategy, gather requirements, execute, and evaluate)
US, WA, Seattle
Do you want to join a brand new team building an AI system that would disrupt the industry? Do you enjoy dealing with ambiguity and working on hard problems in a fast-paced environment?Amazon Connect is a highly disruptive cloud-based contact center that enables businesses to deliver engaging, dynamic, and personal customer service experiences. With Amazon Connect, you can create your own cloud-based contact center and be taking calls in minutes. Amazon Connect leverages the power of Artificial Intelligence and the large ecosystem of AWS services such as Lex, Transcribe, Lambda, S3, and Kinesis to provide a truly frustration free and natural customer experience. With this technology, we are transforming an industry and the way customers interact with businesses and how agents service them.As an Applied Scientist on our team, you will analyze data from huge data sets, create ML forecasting and classification models from conception to deployment, and work closely with other senior technical leaders within the team and across AWS. You will demonstrate your deep Applied Science knowledge and experience at prototyping and building accurate and effective ML models using technology such as AWS Sagemaker, PyTorch, and SparkML. Our team is at an early stage, so you will have significant impact on our ML deliverables with no operational load from existing models/systems.We have a rapidly growing customer base and an exciting charter in front of us that includes solving highly complex engineering and algorithmic problems. We are looking for passionate, talented, and experienced people to join us to innovate on this new service that addresses customer needs to build modern contact centers in the cloud. The position represents a rare opportunity to be a part of a fast-growing business soon after launch, and help shape the technology and product as we grow. You will be playing a crucial role in developing the next generation contact center, and get the opportunity to design and deliver scalable, resilient systems while maintaining a constant customer focus.Learn more about Amazon Connect here:
US, WA, Bellevue
Amazon's Customer Delivery Excellence team is looking for a motivated Data Scientist with proven ability to develop, enhance, automate, and manage optimization and cutting edge prediction and learning/Artificial Intelligence models using strong quantitative skills. The successful candidate will have strong data mining, statistical modeling, machine learning skills and is comfortable facilitating ideation and working from concept through to execution.The position will partner with Engineering, Supply Chain teams, Finance and Technology teams to enhance short term and long term volume prediction and optimization models that use a range of data science methodologies to automate data analysis or to solve complex business problems for the NA Transportation network. Responsibilities include building automated tools and support structures needed to analyze data, design metrics for complex systems, dive deep to determine root cause of forecast errors & changes, create statistical definition of the outliers and methodologies to systematically identify and mitigate model variance drivers.A qualified candidate must have demonstrated ability to develop and manage medium to large-scale models and methodologies that are statistically grounded but also functional and practical. Must possess strong written and verbal communication skills, proven ability to engage and collaborate with customers to drive improvements. Possess high intellectual curiosity with ability to quickly learn new concepts/frameworks, algorithms and technology.Additional Responsibilities include:· Research machine learning algorithms and implement by tailoring to particular business needs and test on large datasets.· Manipulating/mining data from databases (Redshift, SQL Server, Oracle DW)· Create automated metrics using complex databases· Providing analytical network support to improve quality and standard work results· Root cause research to identify process breakdowns within departments and providing data through use of various skill sets to find solutions to breakdown· Collaborate with BI/Data Engineer teams and drive the collection of new data and the refinement of existing data sources to continually improve data quality· Foster culture of continuous engineering improvement through mentoring, feedback, and metricsAmazon is an Equal Opportunity-Affirmative Action Employer- Female/Minority/Disability/Vet
UK, Cambridge
We are looking for someone who is excited to apply cutting-edge techniques from deep learning or natural language processing (NLP) to the text-to-speech (TTS) technology behind Alexa and our AWS cloud speech service.As a Machine Learning Scientist you will be responsible for leading the development and launch of core product features. You will have significant influence on our overall strategy by helping define these product features, drive the system architecture, and spearhead the best practices that enable a quality product.We believe that, like a human speaker, a text-to-speech system can produce more natural speech if it has an improved understanding of the meaning and context of the text. If you agree...You will have the opportunity to solve hard problems with voice – we’ve spent years of invention on this. When we started working on this, the technology didn’t even exist – we had to invent it. Join us if you want to apply your deep learning skills in a dynamic field that is revolutionising the way people interact with devices and services.We believe that voice will fundamentally improve the way people interact with technology. It can make the complex simple—it’s the most natural and convenient user interface. Voice is going to be a big part of our future and we are inventing it here.RESPONSIBILITIES· Use your expertise in deep learning to research and implement novel approaches to make improvements to our text-to-speech technology.· Lead and Mentor junior engineers and scientists.· Participate in the design, development, evaluation, deployment and updating of data-driven models and analytical solutions for spoken language applications.
IN, TS, Hyderabad
Amazon’s location and predictive routing inputs play a crucial role in our vehicle routing problems and other critical applications for Amazon , Prime Now, and other Amazon businesses. As part of the Last Mile Science & Technology organization, you’ll partner closely with other scientists and engineers in a collegial environment with a clear path to business impact. We have an exciting problem area to augment the maps and routing inputs used in Last Mile planning and navigation by leveraging the latest computer vision and deep learning techniques, and are actively looking to hire senior scientists to lead this area. Successful candidates will have a deep knowledge of competing machine learning methods for large scale computer vision problems, the ability to map models into production-worthy code, the communication skills necessary to explain complex technical approaches to a variety of stakeholders and customers, and the excitement to take iterative approaches to tackle big, long term problems.
IN, KA, Bangalore
Machine Learning, Big Data and related quantitative sciences have been strategic to Amazon from the early years. Amazon has been a pioneer in areas such as recommendation engines, eCommerce fraud detection and large-scale optimization of fulfillment center operations.As Amazon has rapidly grown and diversified, the opportunity for applying machine learning has exploded. We have a very broad collection of practical problems where machine learning systems can dramatically improve the customer experience, reduce cost, and drive speed and automation. These include automated pricing and demand forecasting for hundreds of millions of products, predicting ad click probabilities, ranking product search results, matching products from multiple sources, classifying products into large taxonomies, information extraction and sentiment analysis for product reviews, voice recognition, natural language understanding, question answering and conversational systems.Your work will focus on retrieving, cleansing and preparing large scale datasets, training and evaluating models and deploying them to production, with continuous monitoring and evaluation. You will work on large engineering efforts that solve significantly complex problems facing global customers. You will be trusted to operate with independence and are often assigned to focus on areas with significant impact on audience satisfaction. You must be equally comfortable with digging in to customer requirements as you are drilling into design with development teams and developing production ready learning models. You consistently bring strong, data-driven business and technical judgment to decisions.You will work with internal and external stakeholders, cross-functional partners, and end-users around the world at all levels. Our team makes a big impact because nothing is more important to us than pleasing our customers, continually earning their trust, and thinking long term. You are empowered to bring new technologies and deep learning approaches to your solutions.
IN, TS, Hyderabad
Do you know Amazon Logistics has grown by 6 times since 2016 in established markets like North America? Do you know that it is handling more than 80% of Amazon shipments in developing countries like India? Curious to develop sustainable solutions that cater to such rapidly scaling business and countries with widely different customer or driver expectations? Interested in modelling and understanding customer behavior through machine learning, artificial intelligence, and data mining over TB scale data with huge business impact on millions of customers?The problems spread across Routing, Maps, Geo-coding, Address Normalization, fraud detection, defect attribution, labor planning spaces, etc. Some examples include determining optimal route for a driver to deliver a set of packages with many customer and on-ground constraints. Creating a mapping infrastructure ground-up right from scratch. Detecting and pre-empting anomalous behavior by delivery agents on the road in real time or predicting what will be the average no. of deliveries made per hour by each driver in upcoming week. If you are looking for an opportunity to build creative technology solutions that positively impact hundreds of millions of international customers, and relish large ownership and diverse technologies join our team today!You will collaborate with Data Scientist, engineers, executives, and designers, You will be instrumental in shaping the product direction and will be actively involved in defining key product features that impact the business. Also, will be responsible to set up and hold a high software quality bar besides providing technical direction to a highly technical team of Software Engineers. The ideal candidate will be a leader, quick learner and be able to work independently. He/she should be able to operate in a very fast paced environment where time to hit market is super critical. The candidate will need to balance technical leadership and savvy with strong business judgment to make the right decisions about technology choices.
IN, KA, Bengaluru
Interested in Amazon Echo? We’re building the speech and language solutions behind Amazon Echo and other Amazon products and services. Come join us!Amazon is looking for passionate, talented and inventive Sr Applied Scientists for our Bangalore center. You will help build industry-leading speech and language solutions that will enhance user interactions with our products and services in order to provide the best-possible experience for our customers. As a Sr Applied Scientist on our team, you will build and optimize cutting-edge natural language understanding systems and conduct core research aimed at advancing the state of the art. This involves:· Researching the latest modeling techniques. Understanding trade-offs between competing approaches, and identifying the ones that are likely to have real impact on our customers.· Implementing and/or improving modeling tools, training recipes and prototypes utilizing programming skills in Python and/or Java.· Conducting experiments to assess the quality of natural language processing modules and to study the effect of different modeling techniques.· Analyzing field data in order to identify areas of possible improvement or enhancement of the system.· Presenting and discussing ideas and results within the team and beyond.· Mentoring team members
IN, KA, Bengaluru
Amazon Alexa Speech and Language Technology team aims to create a delightful experience for Amazon’s customers by making Alexa understand any spoken request quickly and accurately. As Applied Science Manager you will lead a team of talented scientists in the invention, design, development of novel ML algorithms and modeling techniques for advancing speech and language technology, testing, and deployment of technology components, supporting a range of products. This involves: Conducting and coordinating research leading to improved accuracy of automatic speech recognition (ASR) and/or natural language understanding (NLU) systems.Your responsibilities will include· Direct management of scientists· Providing technical and scientific guidance to your team members· Strategic planning and project management for research projects within the team· Advancing state-of-the-art ASR technologyThe successful candidate will;· Leads effectively and prioritizes well· Exhibits excellent judgment· Hires and develops great people· Is able to dive deep and is never out of touch with the details of the technology· Enables and fosters creativity and innovation on her/his team.· Is results-orientated and has a consistent track record of delivery.· Thinks big, out-of-the-box and long-term.· Communicates effectively with senior management as well as with colleagues from science, engineering and business backgrounds
US, CA, Sunnyvale
MULTIPLE POSITIONS AVAILABLE in Sunnyvale, CA for Services, LLC:1. Participate in the design, development, evaluation, deployment and updating of data driven models and analytical solutions for machine learning and/or natural language applications.2. Develop and/or apply statistical modeling techniques (e.g. Bayesian models and deep neural networks), optimization methods, and other ML techniques to different applications in business and engineering.3. Routinely build and deploy ML models on available data.4. Research and implement novel ML and statistical approaches to add value to the Business.5. Mentor junior engineers and scientists.#0000
US, WA, Seattle
Amazon has built a reputation for excellence with recent examples of being named the #1 most trusted company for customers. To deliver on this reputation for trust the Marketplace Abuse team is tasked with identifying and preventing abuse for our customers and brand owners worldwide.Selling Partner Abuse is hiring a Senior Research Scientist who will research and develop innovative statistics based approaches to predict the near-term future. We are working on cutting edge research and we support Patent and publishing papers. The Scientist must understand deeply about Statistical methods and love extracting, nurturing and transforming high volume of data into actionable business information and making it readily accessible to stakeholders.
US, WA, Seattle
This Senior Research Scientist & Strategist role is on a central strategy team supporting Global Corporate Affairs. The individual selected for this position will create and execute numerous high-impact projects related to global strategy development. The work will challenge you to be resourceful, think big while staying connected to the details, translate data-driven insights into strategic recommendations, and embrace a high degree of change and ambiguity. The scope and scale of what we strive to achieve is immense, but it’s also meaningful and energizing.Role responsibilities:· Independently (and in partnership with internal stakeholders) develop, propose, and execute research projects that provide actionable insights to our many global stakeholders· Manage and expand our partnerships with internal research teams and external research agencies with whom we partner· Connect findings across research projects (primary and secondary) to generate overarching insights· Communicate data-driven insights and recommendations through presentations, papers, and other internal communication channels· Provide consultation and create trainings on research fundamentals/best practices, survey design, and developing appropriate KPIs for teams across the organization** Location is open to Virigina, DC, WA, NY, CA, MA**
US, WA, Seattle
The Alexa Shopping team is looking for Applied Scientists to build the next generation of shopping assistants.As an applied scientist in Alexa Shopping, you will be responsible for the research, design and development of new natural language, search, and machine learning technologies for voice shopping. You will be working with top scientists and engineers, as well as with product teams and other research partners, both locally and abroad. Your work will combine data mining, systems and software development, exploration of new technologies, as well as publications and presentations at top scientific conferences.The ideal candidates have deep expertise in one or several of the following fields: Information Retrieval, Web search, Web data mining, Machine Learning, Natural Language Processing, Artificial Intelligence. An ideal candidate shows a bias for action and has a strong understanding of empirical methods. The ability to write clearly and speak convincingly as evidenced through participation in academic conferences and service in the scientific community are a must.Amazon is an Equal Opportunity Employer – Minority / Women / Disability / Veteran / Gender Identity / Sexual Orientation / Age.
US, WA, Seattle
Do you want to influence, design, and build tools that solve Amazon-scale, world-class analytical problems? Are you passionate about Big Data, machine learning, and econometrics? The Economic Trends and Insights (YETI) team is building and scaling an Automated Insights tool to spread data science across the Finance organization and turn Big Data into big business insights.As a Data Scientist on the YETI team, you design and build algorithms and tools to solve the biggest problems across the Consumer Finance organization. From anomaly and trend break detection, to causal inference, to forecasting - you turn complex problems into simple solutions used by technical and non-technical users. You are customer obsessed and enjoy diving deep into data to uncover insights, then partner with software developers and economists to scale your insights across all of Amazon. You treat everything like an experiment, and are able to look around corners to test assumptions and find the non-intuitive answers lurking beneath the surface.As part of the YETI team, you will:· Design, test, and build algorithmic solutions into a production Automated Insight software tool· Collaborate with leading economists and scientists to create scalable solutions to problems including trend break detection, causal inference, and automated root cause analysis· Partner with internal customers to deep dive anomalous trends to generate high impact insights
US, NY, New York
Amazon’s Advertising Audience and Media Insights team (AMI) is looking for a motivated and experienced Applied Science Manager to own and lead a growing science team within our org. Our team owns the product, science, technology and deployment roadmap for advanced analytics and media planning products across our Insights and Performance Team. Analytics and media planning is core to Amazon’s growth, as it helps our suppliers drive awareness, consideration, and purchase of their products by hundreds of millions of consumers around the world, and generates revenue which helps us lower prices and invest in improvements to our customer experience. We are a highly motivated, collaborative and fun-loving team with an entrepreneurial spirit and bias for action. With a broad mandate to experiment and innovate, we are growing at an unprecedented rate with a seemingly endless range of new opportunities.You are the single threaded owner of the Media Insights Science program. You are customer obsessed and have sharp business acumen to build a product vision and roadmap that contributes to the success of our advertisers. You are a technical leader with track record of building and growing applied science and engineering teams. You experiment, develop proof of concept, and ship software. You own the day to day maintenance and evolution of systems that serve the needs of our large and growing base of advertisers. We are looking for a leader who can thrive in a startup environment, thinks big, can move fast, and wants to change the way customers use data to drive business success. If you excel at proposing and developing unique analytical insights and are passionate about creating the future, come join us as we work hard, have fun, and make history.You will:* Own technical vision and direction: You identify and mitigate risks, construct project schedules, and prepare status reports; you embrace performance metrics and measurement techniques because they help you assess how well system-related services are running.* Grow your team: You'll be accountable for growing a team featuring a mix of scientists and engineers that deliver results. This means you'll wear a lot of hats -- from developing scientific models to leading technical discussions, recruiting, establishing processes, and so on. You'll be an example of Amazon's leadership principles and work to grow more leaders within your group. You will manage a fast-growing team that spans multiple locations.* Collaborate on product direction: You’ll build and maintain strong relationships between your team and partner disciplines (Product, User Experience, QA) to ensure that we're focused on delivering the right product for customers.* Lead beyond your team: You will be a key leader within Amazon’s advertising organization. You will contribute to the overall growth of our product development organization. You'll share your best practices and technical acumen in order to drive technology decisions across our organization.You are:* Pragmatic and iterative in developing models and building software: you have an ability to simplify and get things done quickly with a demonstrated track record of building and delivering software and working effectively with external and internal teams.* Highly analytical and data-driven: You solve problems in ways that can be backed up with verifiable data. You focus on driving processes, tools, and statistical methods which support rational decision-making.* Technically fearless: You aren't satisfied by performing 'as expected' and push the limits past conventional boundaries.* Team obsessed: You find and cultivate talent to be capable of achieving outstanding results. You help develop the creative atmosphere to let engineers innovate, while holding them accountable for making smart decisions and delivering results.* Comfortable with ambiguity: You're able to explore new problem spaces with unique constraints and thus non-obvious solutions; you’re quick to identify any gaps in the team and the right person to fill them to best deliver value to customers.* Solutions; you’re quick to identify any gaps in the team and the right person to fill them to best deliver value to customers.
US, NY, New York
Amazon AI is looking for world class scientists and engineers to join its AI Labs working within speech recognition. This group is entrusted with developing core data mining, natural language processing, and machine learning algorithms for AWS services. At Amazon AI you will invent, implement, and deploy state of the art machine learning algorithms and systems. You will build prototypes and explore conceptually new speech recognition solutions. You will interact closely with our customers and with the academic community. You will be at the heart of a growing and exciting focus area for AWS and work with other acclaimed engineers and world famous scientists.
US, NY, New York
Amazon AI is looking for world class scientists and engineers to join its AI Labs working within speech recognition. This group is entrusted with developing core data mining, natural language processing, and machine learning algorithms for AWS services. At Amazon AI you will invent, implement, and deploy state of the art machine learning algorithms and systems. You will build prototypes and explore conceptually new speech recognition solutions. You will interact closely with our customers and with the academic community. You will be at the heart of a growing and exciting focus area for AWS and work with other acclaimed engineers and world famous scientists.