Amazon Alexa’s new wake word research at Interspeech

Work aims to improve accuracy of models both on- and off-device.

Every interaction with Alexa begins with the wake word: usually “Alexa”, but sometimes “Amazon”, “Echo”, or “Computer” — or, now, “Hey Samuel”. Only after positively identifying the wake word does an Alexa-enabled device send your request to the cloud for further processing.

Six years after the announcement of the first Amazon Echo, the Alexa science team continues to innovate new approaches to wake word recognition, improving Alexa’s responsiveness and accuracy.

At this year’s Interspeech, for instance, Alexa researchers presented five different papers about new techniques for wake word recognition. One of these — “Building a robust word-level wakeword verification network” — describes models that run in the cloud to confirm on-device wake word detections.

Wake word spectrogram
Because audio signals can be represented as two-dimensional mappings of frequency (y-axis) against time (x-axis), convolutional neural networks apply naturally to them.
From "Accurate detection of wake word start and end using a CNN"

Another paper, “Metadata-aware end-to-end keyword spotting”, describes a new system that uses metadata about the state of the Alexa-enabled device — such as the type of device and whether it’s playing music or sounding an alarm — to improve the accuracy of the on-device wake word detector.

The wake word detectors reported in both papers rely, at least in part, on convolutional neural networks. Originally developed for image processing, convolutional neural nets, or CNNs, repeatedly apply the same “filter” to small chunks of input data. For object recognition, for instance, a CNN might step through an image file in eight-by-eight blocks of pixels, inspecting each block for patterns associated with particular objects. 

Since audio signals can be represented as two-dimensional mappings of frequency against time, CNNs apply naturally to them as well. Each of the filters applied to a CNN’s inputs defines a channel through the first layer of the CNN, and usually, the number of channels increases with every layer.

Varying norms

Metadata-aware end-to-end keyword spotting” is motivated by the observation that if a device is emitting sound — music, synthesized speech, or an alarm sound — it causes a marked shift in the input signal’s log filter bank energies, or LFBEs. The log filter banks are a set of differently sized frequency bands chosen to emphasize the frequencies in which human hearing is most acute.

Graph showing average values of acoustic properties of wake word signals when a device is emitting sound and when it’s not.
Average values of acoustic properties — log filter-bank energies — of wake word signals as measured on-device when the device is emitting sound (orange) and when it’s not (blue).
From “Metadata-aware end-to-end keyword spotting”

To address this problem, applied scientists Hongyi Liu and Apurva Abhyankar and their colleagues include device metadata as an input to their wake word model. The model embeds the metadata, or represents it as points in a multidimensional space, such that location in the space conveys information useful to the model. The model uses the embeddings in two different ways.

One is as an additional input to the last few layers of the network, which decide whether the acoustic input signal includes the wake word. The final outputs of the convolutional layers are flattened, or strung together into a single long vector. The metadata embedding vector is fed into a fully connected layer — a layer all of whose processing nodes pass their outputs to all of the nodes of the next layer — and the output is concatenated to the flattened audio feature vector. 

This fused vector passes to a final fully connected layer, which issues a judgment about whether the network contains the wake word or not.

The other use of the metadata embedding is to modulate the outputs of the convolutional layers while they’re processing the input signal. The filters that a CNN applies to inputs are learned during training, and they can vary greatly in size. Consequently, the magnitude of the values passing through the network’s various channels can vary as well.

With CNNs, it’s common practice to normalize the channels’ outputs between layers, so that they’re all on a similar scale, and no one channel swamps the others. But Liu, Abhyankar, and their colleagues train their model to vary the normalization parameters depending on the metadata vector, which improves the network’s ability to generalize to heterogenous data sets. 

The researchers believe that this model better captures the characteristics of the input audio signal when the Alexa-enabled device is emitting sound. In their paper, they report experiments showing that, on average, a model trained with metadata information achieves a 14.6% improvement in false-reject rate relative to a baseline CNN model.

Paying attention

The metadata-aware wake word detector runs on-device, but the next two papers describe models that run in the cloud. On-device models must have small memory footprints, which means that they sacrifice some processing power. If an on-device model thinks it has detected a wake word, it sends a short snippet of audio to the cloud for confirmation by a larger, more-powerful model.

The on-device model tries to identify the start of the wake word, but sometimes it misses slightly. To ensure that the cloud-based model receives the whole wake word, the snippet sent by the device includes the half-second of audio preceding the device’s estimate of the wake word’s start.

Model depicting variations in alignment of wake word signals send to the cloud for verification.
Wake word signals sent to the cloud for verification vary in the quality of their alignment. Sometimes, in trying to identify the start of the wake word, the device misses by a fraction of a second, which can cause difficulty for cloud models trained on well-aligned data.
From “Building a robust word-level wakeword verification network”

When CNNs are trained on well-aligned data, convolutional-layer outputs that focus on particular regions of the input can become biased toward finding wake word features in those regions. This can result in weaker performance when the alignment is noisy.

In “Building a robust word-level wakeword verification network”, applied scientist Rajath Kumar and his colleagues address this problem by adding recurrent layers to their network, to process the outputs of the convolutional layers. Recurrent layers can capture information as time sequences. Instead of learning where the wake word occurs in the input, the recurrent layers learn how the sequence changes temporally when the wake word is present. 

This allows the researchers to train their network on well-aligned data without suffering much performance drop off on noisy data. To further improve performance, the researchers also use an attention layer to process and re-weight the sequential outputs of the recurrent layers, to emphasize the outputs required for wake word verification. The model is thus a convolutional-recurrent-attention (CRA) model.

Diagrams indicating differences between a conventional CNN architecture and a convolutional-recurrent-attention architecture.
These diagrams indicate the differences between a conventional CNN architecture (top) and a convolutional-recurrent-attention (CRA) architecture (bottom).
From “Building a robust word-level wakeword verification network”

To evaluate their CRA model, the researchers compared its performance to that of several CNN-only models. Each example in the training data included 195 input frames, or sequential snapshots of the frequency spectrum. Within that 195-frame span, two of the CNN models looked at sliding windows of 76 frames or 100 frames. A third CNN model, and the CRA model, looked at all 195 frames. The models’ performance was assessed relative to a baseline wake word detector that combines a deep neural network with a hidden Markov model, an architecture was the industry standard for some time. 

On accurately aligned inputs, the CRA model offers only a slight improvement over the 195-frame CNN model. Compared to the baseline, the CNN model reduced the false-acceptance rate by 53%, while the CRA reduced it by 55%. On the same task, the 100-frame CNN model achieved only a 35% reduction.

Table showing percentage of decrease in FAR in comparison to 2-stage DNN-HMM.

On noisily aligned inputs, the CRA model offered a much more dramatic improvement. Relative to baseline, it reduced the false-acceptance rate by 60%. The 195-frame CNN model managed only 31%, the 100-frame model 44%.

Research areas

Related content

GB, Cambridge
Amazon Devices is an inventive research and development company that designs and engineer high-profile devices like Echo, Fire Tablets, Fire TV, and other consumer devices. We are looking for exceptional scientists to join our Applied Science team to advance the state-of-the-art in developing efficient multimodal language models across our product portfolio. Through close hardware-software integration, we design and train models for resource efficiency across the hardware and software tech stack. The Silicon and Solutions Group Edge AI team is looking for a talented Sr. Applied Scientist who will lead our efforts on inventing evaluation methods for multimodal language models and agents for new devices, including audio and vision experiences. Key job responsibilities - Collaborate with cross-functional engineers and scientists to advance the state of the art in multimodal model evaluations for devices, including audio, images, and videos - Invent and validate reliability for novel automated evaluation methods for perception tasks, such as fine-tuned LLM-as-judge - Develop and extend our evaluation framework(s) to support expanding capabilities for multimodal language models - Analyze large offline and online datasets to understand model gaps, develop methods to interpret model failures, and collaborate with training teams to enhance model capabilities for product use cases - Work closely with scientists, compiler engineers, data collection, and product teams to advance evaluation methods - Mentor less experienced Applied Scientists A day in the life As a Scientist with the Silicon and Solutions Group Edge AI team, you'll contribute to innovative methods for evaluating new product experiences and discover ways to enhance our model capabilities and enrich our customer experiences. You'll research new methods for reliably assessing perception capabilities for audio-visual tasks in multimodal language models, design and implement new metrics, and develop our evaluation framework. You'll collaborate across teams of engineers and scientists to identify and root cause issues in models and their system integration to continuously enhance the end-to-end experience. About the team Our Edge AI science team brings together our unique skills and experiences to deliver state-of-the-art multimodal AI models that enable new experiences on Amazon devices. We work at the intersection of hardware, software, and science to build models designed for our custom silicon.
US, CA, Sunnyvale
The Artificial General Intelligence (AGI) team is looking for a passionate, talented, and inventive Applied Scientist with a strong deep learning background, to help build industry-leading technology with generative AI (GenAI) and multi-modal systems. Key job responsibilities As an Applied Scientist with the AGI team, you will work with talented peers to develop algorithms and modeling techniques to advance the state of the art with multi-modal systems. Your work will directly impact our customers in the form of products and services that make use of vision and language technology. You will leverage Amazon’s large-scale computing resources to accelerate development with multi-modal Large Language Models (LLMs) and GenAI in Computer Vision. About the team The AGI team has a mission to push the envelope with multimodal LLMs and GenAI in Computer Vision, in order to provide the best-possible experience for our customers.
US, MA, Boston
The Artificial General Intelligence (AGI) team is looking for a passionate, talented, and inventive Applied Scientist with a strong deep learning background, to build industry-leading Generative Artificial Intelligence (GenAI) technology with Large Language Models (LLMs) and multimodal systems. Key job responsibilities As an Applied Scientist with the AGI team, you will work with talented peers to lead the development of algorithms and modeling techniques, to advance the state of the art with LLMs. Your work will directly impact our customers in the form of products and services that make use of GenAI technology. You will leverage Amazon’s heterogeneous data sources and large-scale computing resources to accelerate advances in LLMs. About the team The AGI team has a mission to push the envelope in GenAI with LLMs and multimodal systems, in order to provide the best-possible experience for our customers.
US, WA, Seattle
Prime Video is a first-stop entertainment destination offering customers a vast collection of premium programming in one app available across thousands of devices. Prime members can customize their viewing experience and find their favorite movies, series, documentaries, and live sports – including Amazon MGM Studios-produced series and movies; licensed fan favorites; and programming from Prime Video add-on subscriptions such as Apple TV+, Max, Crunchyroll and MGM+. All customers, regardless of whether they have a Prime membership or not, can rent or buy titles via the Prime Video Store, and can enjoy even more content for free with ads. Are you interested in shaping the future of entertainment? Prime Video's technology teams are creating best-in-class digital video experience. As a Prime Video technologist, you’ll have end-to-end ownership of the product, user experience, design, and technology required to deliver state-of-the-art experiences for our customers. You’ll get to work on projects that are fast-paced, challenging, and varied. You’ll also be able to experiment with new possibilities, take risks, and collaborate with remarkable people. We’ll look for you to bring your diverse perspectives, ideas, and skill-sets to make Prime Video even better for our customers. With global opportunities for talented technologists, you can decide where a career Prime Video Tech takes you! Key job responsibilities Develop ML models for various recommendation & search systems using deep learning, online learning, and optimization methods Work closely with other scientists, engineers and product managers to expand the depth of our product insights with data, create a variety of experiments to determine the high impact projects to include in planning roadmaps Stay up-to-date with advancements and the latest modeling techniques in the field Publish your research findings in top conferences and journals A day in the life We're using advanced approaches such as foundation models to connect information about our videos and customers from a variety of information sources, acquiring and processing data sets on a scale that only a few companies in the world can match. This will enable us to recommend titles effectively, even when we don't have a large behavioral signal (to tackle the cold-start title problem). It will also allow us to find our customer's niche interests, helping them discover groups of titles that they didn't even know existed. We are looking for creative & customer obsessed machine learning scientists who can apply the latest research, state of the art algorithms and ML to build highly scalable page personalization solutions. You'll be a research leader in the space and a hands-on ML practitioner, guiding and collaborating with talented teams of engineers and scientists and senior leaders in the Prime Video organization. You will also have the opportunity to publish your research at internal and external conferences.
US, CA, Palo Alto
About Sponsored Products and Brands The Sponsored Products and Brands (SPB) team at Amazon Ads is re-imagining the advertising landscape through generative AI technologies, revolutionizing how millions of customers discover products and engage with brands across Amazon.com and beyond. We are at the forefront of re-inventing advertising experiences, bridging human creativity with artificial intelligence to transform every aspect of the advertising lifecycle from ad creation and optimization to performance analysis and customer insights. We are a passionate group of innovators dedicated to developing responsible and intelligent AI technologies that balance the needs of advertisers, enhance the shopping experience, and strengthen the marketplace. If you're energized by solving complex challenges and pushing the boundaries of what's possible with AI, join us in shaping the future of advertising. About our team SPB Ad Response Prediction team is your choice, if you want to join a highly motivated, collaborative, and fun-loving team with a strong entrepreneurial spirit and bias for action. We are seeking an experienced and motivated Applied Scientist with machine learning engineering background who loves to innovate at the intersection of customer experience, deep learning, and high-scale machine learning systems. We are looking for a talented Applied Scientist with a strong background in machine learning engineering to join our team and help us grow the business. In this role, you will partner with a team of engineers and scientists to build advanced machine learning models and infrastructure, from training to inference, including emerging LLM-based systems, that deliver highly relevant ads to shoppers across all Amazon platforms and surfaces worldwide. Key job responsibilities As an Applied Scientist, you will: * Develop scalable and effective machine learning models and optimization strategies to solve business problems. * Conduct research on new machine learning modeling to optimize all aspects of Sponsored Products business. * Enhance the scalability, automation, and efficiency of large-scale training and real-time inference systems. * Pioneer the development of LLM inference infrastructure to support next-generation GenAI workloads at Amazon Ads scale.
US, WA, Seattle
The Economics Science team in the Amazon Manager Experience (AMX) organization builds science models supporting employee career-related experiences such as their evaluation, learning and development, onboarding, and promotion. Additionally, the team conducts experiments for a wide range of employee and talent-related product features, and measures the impact of product and program initiatives in enhancing our employees' career experiences at Amazon. The team is looking for an Economist who specializes in the field of macroeconomics and time series forecasting. This role combines traditional macroeconomic analysis with modern data science techniques to enhance understanding and forecasting of workforce dynamics at scale. Key job responsibilities The economists within ALX focus on enhancing causal evaluation, measurement, and experimentation tasks to ensure various science integrations and interventions achieve their goals in building more rewarding careers for our employees. The economists develop and implement complex randomization designs that address the nuances of experimentation in complex settings where multiple populations interact. Additionally, they engage in building a range of econometric models that surface various proactive and reactive inspection signals, aiming toward better alignment in the implementation of talent processes. The economists closely collaborate with scientists from diverse backgrounds, as well as program and product leaders, to implement and assess science solutions in our products.
GB, London
Are you looking to work at the forefront of Machine Learning and AI? Would you be excited to apply Generative AI algorithms to solve real world problems with significant impact? The Generative AI Innovation Center helps AWS customers implement Generative AI solutions and realize transformational business opportunities. This is a team of strategists, scientists, engineers, and architects working step-by-step with customers to build bespoke solutions that harness the power of generative AI. The team helps customers imagine and scope the use cases that will create the greatest value for their businesses, define paths to navigate technical or business challenges, develop proof-of-concepts, and make plans for launching solutions at scale. The GenAI Innovation Center team provides guidance on best practices for applying generative AI responsibly and cost efficiently. You will work directly with customers and innovate in a fast-paced organization that contributes to game-changing projects and technologies. You will design and run experiments, research new algorithms, and find new ways of optimizing risk, profitability, and customer experience. We’re looking for Data Scientists capable of using GenAI and other techniques to design, evangelize, and implement state-of-the-art solutions for never-before-solved problems. Key job responsibilities As a Data Scientist, you will • Collaborate with AI/ML scientists and architects to research, design, develop, and evaluate generative AI solutions to address real-world challenges • Interact with customers directly to understand their business problems, aid them in implementation of generative AI solutions, brief customers and guide them on adoption patterns and paths to production • Create and deliver best practice recommendations, tutorials, blog posts, sample code, and presentations adapted to technical, business, and executive stakeholder • Provide customer and market feedback to product and engineering teams to help define product direction About the team Diverse Experiences AWS values diverse experiences. Even if you do not meet all of the qualifications and skills listed in the job description, we encourage candidates to apply. If your career is just starting, hasn’t followed a traditional path, or includes alternative experiences, don’t let it stop you from applying. Why AWS? Amazon Web Services (AWS) is the world’s most comprehensive and broadly adopted cloud platform. We pioneered cloud computing and never stopped innovating — that’s why customers from the most successful startups to Global 500 companies trust our robust suite of products and services to power their businesses. Inclusive Team Culture Here at AWS, it’s in our nature to learn and be curious. Our employee-led affinity groups foster a culture of inclusion that empower us to be proud of our differences. Ongoing events and learning experiences, including our Conversations on Race and Ethnicity (CORE) and AmazeCon (gender diversity) conferences, inspire us to never stop embracing our uniqueness. Mentorship & Career Growth We’re continuously raising our performance bar as we strive to become Earth’s Best Employer. That’s why you’ll find endless knowledge-sharing, mentorship and other career-advancing resources here to help you develop into a better-rounded professional. Work/Life Balance We value work-life harmony. Achieving success at work should never come at the expense of sacrifices at home, which is why we strive for flexibility as part of our working culture. When we feel supported in the workplace and at home, there’s nothing we can’t achieve in the cloud.
IN, KA, Bengaluru
The Amazon Alexa AI team in India is seeking a talented, self-driven Applied Scientist to work on prototyping, optimizing, and deploying ML algorithms within the realm of Generative AI. The Applied Scientist will be in a team of exceptional scientists to develop novel algorithms and modeling techniques to advance the state of the art in Natural Language Processing (NLP) or Computer Vision (CV) related tasks. They will work in a hybrid, fast-paced organization where scientists, engineers, and product managers work together to build customer facing experiences. They will collaborate with and mentor other scientists to raise the bar of scientific research in Amazon. Their work will directly impact our customers in the form of products and services that make use of speech, language, and computer vision technologies. Key job responsibilities - Research, experiment and build Proof Of Concepts advancing the state of the art in AI & ML for GenAI. - Collaborate with cross-functional teams to architect and execute technically rigorous AI projects. - Thrive in dynamic environments, adapting quickly to evolving technical requirements and deadlines. - Engage in effective technical communication (written & spoken) with coordination across teams. - Conduct thorough documentation of algorithms, methodologies, and findings for transparency and reproducibility. - Publish research papers in internal and external venues of repute - Support on-call activities for critical issues A day in the life You will be working with a group of talented scientists on researching algorithm and running experiments to test scientific proposal/solutions to improve solutions powering customer experience on Alexa+. This will involve collaboration with partner teams including engineering, PMs, data annotators, and other scientists to discuss data quality, policy, and model development. You will mentor other scientists, review and guide their work, help develop roadmaps for the team. You work closely with partner teams across Alexa to deliver platform features that require cross-team leadership.
US, CA, Sunnyvale
As a Principal Scientist within the Artificial General Intelligence (AGI) organization, you are a trusted part of the technical leadership. You bring business and industry context to science and technology decisions. You set the standard for scientific excellence and make decisions that affect the way we build and integrate algorithms. You solicit differing views across the organization and are willing to change your mind as you learn more. Your artifacts are exemplary and often used as reference across organization. You are a hands-on scientific leader. Your solutions are exemplary in terms of algorithm design, clarity, model structure, efficiency, and extensibility. You tackle intrinsically hard problems, acquiring expertise as needed. You decompose complex problems into straightforward solutions. You amplify your impact by leading scientific reviews within your organization or at your location. You scrutinize and review experimental design, modeling, verification and other research procedures. You probe assumptions, illuminate pitfalls, and foster shared understanding. You align teams toward coherent strategies. You educate, keeping the scientific community up to date on advanced techniques, state of the art approaches, the latest technologies, and trends. You help managers guide the career growth of other scientists by mentoring and play a significant role in hiring and developing scientists and leads. You will play a critical role in driving the development of Generative AI (GenAI) technologies that can handle Amazon-scale use cases and have a significant impact on our customers' experiences. Key job responsibilities You will be responsible for defining key research directions, adopting or inventing new machine learning techniques, conducting rigorous experiments, publishing results, and ensuring that research is translated into practice. You will develop long-term strategies, persuade teams to adopt those strategies, propose goals and deliver on them. You will also participate in organizational planning, hiring, mentorship and leadership development. You will be technically fearless and with a passion for building scalable science and engineering solutions. You will serve as a key scientific resource in full-cycle development (conception, design, implementation, testing to documentation, delivery, and maintenance).
US, CA, Mountain View
Amazon launched the Generative AI Innovation Center (GAIIC) in Jun 2023 to help AWS customers accelerate the use of Generative AI to solve business and operational problems and promote innovation in their organization (https://press.aboutamazon.com/2023/6/aws-announces-generative-ai-innovation-center). GAIIC provides opportunities to innovate in a fast-paced organization that contributes to game-changing projects and technologies that get deployed on devices and in the cloud. As an Applied Science Manager in GAIIC, you'll partner with technology and business teams to build new GenAI solutions that delight our customers. You will be responsible for directing a team of data/research/applied scientists, deep learning architects, and ML engineers to build generative AI models and pipelines, and deliver state-of-the-art solutions to customer’s business and mission problems. Your team will be working with terabytes of text, images, and other types of data to address real-world problems. The successful candidate will possess both technical and customer-facing skills that will allow you to be the technical “face” of AWS within our solution providers’ ecosystem/environment as well as directly to end customers. You will be able to drive discussions with senior technical and management personnel within customers and partners, as well as the technical background that enables them to interact with and give guidance to data/research/applied scientists and software developers. The ideal candidate will also have a demonstrated ability to think strategically about business, product, and technical issues. Finally, and of critical importance, the candidate will be an excellent technical team manager, someone who knows how to hire, develop, and retain high quality technical talent. About the team About the team Diverse Experiences AWS values diverse experiences. Even if you do not meet all of the preferred qualifications and skills listed in the job description, we encourage candidates to apply. If your career is just starting, hasn’t followed a traditional path, or includes alternative experiences, don’t let it stop you from applying. Why AWS? Amazon Web Services (AWS) is the world’s most comprehensive and broadly adopted cloud platform. We pioneered cloud computing and never stopped innovating — that’s why customers from the most successful startups to Global 500 companies trust our robust suite of products and services to power their businesses. Inclusive Team Culture AWS values curiosity and connection. Our employee-led and company-sponsored affinity groups promote inclusion and empower our people to take pride in what makes us unique. Our inclusion events foster stronger, more collaborative teams. Our continual innovation is fueled by the bold ideas, fresh perspectives, and passionate voices our teams bring to everything we do. Mentorship & Career Growth We’re continuously raising our performance bar as we strive to become Earth’s Best Employer. That’s why you’ll find endless knowledge-sharing, mentorship and other career-advancing resources here to help you develop into a better-rounded professional. Work/Life Balance We value work-life harmony. Achieving success at work should never come at the expense of sacrifices at home, which is why we strive for flexibility as part of our working culture. When we feel supported in the workplace and at home, there’s nothing we can’t achieve.