Amazon ScienceConsumer division CEO, Jeff Wilke, discusses the history of Amazon's recommendation algorithm at re:MARS 2019, including collaborative filtering and beyond. Read more: https://www.amazon.science/the-history-of-amazons-recommendation-algorithm
The history of Amazons recommendation algorithm | Amazon ScienceAmazon Science2020-06-10 | Consumer division CEO, Jeff Wilke, discusses the history of Amazon's recommendation algorithm at re:MARS 2019, including collaborative filtering and beyond. Read more: https://www.amazon.science/the-history-of-amazons-recommendation-algorithm
Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScienceWhats it like to work in Supply Chain Optimization Technologies as a scientist?Amazon Science2023-04-05 | The Supply Chain Optimization Technologies (SCOT) team builds and manages the systems that power Amazon’s core consumer supply chain. The SCOT community spans across a variety of disciplines and job families to address the most complex challenges. Learn more about how SCOT scientists are working with their cross-functional partners to collaborate throughout the company and engage with the research community.
Read the latest from SCOT on Amazon Science: https://www.amazon.science/tag/supply-chain-optimization-technologies
#AmazonScience #SupplyChain #OperationsResearchSpliced binned-Pareto for the NFL | NeurIPS 2022 Demo | Amazon ScienceAmazon Science2022-12-08 | At NeurIPS 2022, Amazon scientists present a robust method, the spliced binned-Pareto (SBP), for modeling heavy-tailed problems. In this demo, they discuss how the SBP’s ability to provide multi-model, heavy-tailed distribution forecasts of an American football play’s yards gained outcome given the players’ temporal dynamics, allowed for the NFL’s QB Passing Score announced at the 2022 Super Bowl. They further demo the robustness and accuracy of their method against other state-of-the-art methods on a public benchmark data set.
Amazon scientists include: • François-Xavier Aubet, applied machine learning scientist • Elena Ehrlich, principal data science manager • Paul Budnarain, applied scientist • Brad Gross, senior data scientist
Read more about the method on Amazon Science: https://www.amazon.science/blog/the-science-behind-nfl-next-gen-stats-new-passing-metric
#AmazonScience #NeurIPS2022 #NeurIPSFrom Seeing to Doing: Understanding and Interacting with the Real World | Fei-Fei Li at AMLC 2022Amazon Science2022-11-30 | In October, during Amazon’s Machine Learning Conference (AMLC 2022) Fei-Fei Li, the inaugural Sequoia Professor in the Computer Science Department at Stanford University, and Co-Director of Stanford’s Human-Centered AI Institute, gave a keynote titled, "From Seeing to Doing: Understanding and Interacting with the Real World". Her talk discusses work in her lab that spans both perception and robotic learning, underscoring the importance of an "ecological approach to learning."
Li previously served as the Director of Stanford’s AI Lab from 2013 to 2018. And during her sabbatical from Stanford from January 2017 to September 2018, she was Vice President at Google and served as Chief Scientist of AI/ML at Google Cloud. Dr. Fei-Fei Li obtained her B.A. degree in physics from Princeton in 1999 with High Honors, and her PhD degree in electrical engineering from California Institute of Technology (Caltech) in 2005. She also holds a Doctorate Degree (Honorary) from Harvey Mudd College.
Li’s current research interests include cognitively inspired AI, machine learning, deep learning, computer vision and AI+healthcare especially ambient intelligent systems for healthcare delivery. In the past she has also worked on cognitive and computational neuroscience. Dr. Li has published more than 200 scientific articles in top-tier journals and conferences, including Nature, PNAS, Journal of Neuroscience, CVPR, ICCV, NIPS, ECCV, ICRA, IROS, RSS, IJCV, IEEE-PAMI, New England Journal of Medicine, Nature Digital Medicine, etc. Dr. Li is the inventor of ImageNet and the ImageNet Challenge, a critical large-scale dataset and benchmarking effort that has contributed to the latest developments in deep learning and AI. In addition to her technical contributions, she is a national leading voice for advocating diversity in STEM and AI. She is co-founder and chairperson of the national non-profit AI4ALL aimed at increasing inclusion and diversity in AI education.
#AmazonScience #MachineLearningLearning to See by Looking at Noise | Antonio Torralba at AMLC 2022Amazon Science2022-11-30 | In October, during Amazon’s Machine Learning Conference (AMLC 2022), Antonio Torralba, Delta Electronics Professor of Electrical Engineering and Computer Science at the Massachusetts Institute of Technology (MIT) and Head of the AI+D faculty in the EECS department, gave a keynote titled, "Learning to See by Looking at Noise".
Torralba received a degree in telecommunications engineering from Telecom BCN, Spain in 1994 and a Ph.D. degree in signal, image, and speech processing from the Institut National Polytechnique de Grenoble, France in 2000. From 2000 to 2005, he spent postdoctoral training at the Brain and Cognitive Sciences Department and the Computer Science and Artificial Intelligence Laboratory, MIT, where he is now a professor. Learn more about Torralba's work here: https://groups.csail.mit.edu/vision/torralbalab/
#AmazonScience #MachineLearningEpisodes of Experience and Generative Intelligence | Linda B. Smith at AMLC 2022Amazon Science2022-11-30 | In October, during Amazon’s Machine Learning Conference (AMLC 2022), Linda B. Smith, distinguished professor at Indiana University Bloomington, gave a keynote titled, "Episodes of Experience and Generative Intelligence".
Smith is an internationally recognized leader in cognitive science and cognitive development. Taking a complex systems perspective, she seeks to understand the interdependencies among perceptual, motor and cognitive developments during the first three years of post-natal life. Using wearable sensors, including head-mounted cameras, she studies how the young learner’s own behavior creates the statistical structure of the learning environments. Her work has led to novel insights currently being extended through collaborations to robotics and artificial intelligence. She received her PhD from the University of Pennsylvania in 1977 and immediately after joined the faculty at Indiana University.
Her talk, she addressed how humans, including toddlers, are adept at taking knowledge from past experiences and using it in compelling new ways. She highlights findings from studies of toddler’s natural egocentric experiences, honing in on the main point that everyday experiences occur in time-extended episodes and proposed that these statistics are the secret ingredient to innovative intelligence. Moreover, they provide novel insights into the internal processes that learn, generalize and innovate.
#AmazonScience #MachineLearningValue Based NLP | Pascale Fung at AMLC 2022Amazon Science2022-11-30 | In October, during Amazon’s Machine Learning Conference (AMLC 2022), Pascale Fung, a Chair Professor at the Department of Electronic & Computer Engineering at The Hong Kong University of Science & Technology (HKUST), and a visiting professor at the Central Academy of Fine Arts in Beijing, gave a keynote titled, "Value Based NLP".
Fung's research interest lies in building intelligent systems that can understand and empathize with humans. To achieve this goal, her specific areas of research are using statistical modelling and deep learning for natural language processing, spoken language systems, emotion and sentiment recognition, and other areas of AI.
Learn more about Fung's work here: https://pascale.home.ece.ust.hk/about.html
#AmazonScience #MachineLearningAlexa Prize Taskbot Challenge Finals | Amazon ScienceAmazon Science2022-08-12 | Launched in 2021, the Alexa Prize TaskBot Challenge is a competition for university students to develop bots that could assist customers in completing cooking or do-it-yourself home improvement tasks that required multiple steps and decisions. The teams’ goal: build taskbots that assist customers in multi-step tasks, and adapt those instructions based on the resources and tools available to the customer. Customers interacted with one of ten taskbots and rated the interaction—on a scale from 1 to 5—on how helpful that taskbot was with the task. Five teams were selected for the finals earlier this year, and GRILLBot, developed by a team of graduate students at the University of Glasgow, emerged as the winner of the challenge. Learn more: https://www.amazon.science/alexa-prize/three-top-performers-emerge-in-inaugural-alexa-prize-taskbot-challenge
#AmazonScience #AlexaPrizeLatinX in AI and Amazon Science panel discussionAmazon Science2022-07-22 | On April 27, five science leaders from Amazon held a panel discussion for members of the LatinX in AI community to hear about their career journey, the type of research they’re working on, and to answer pre-submitted questions from the audience about diversity and innovation. Learn more about the work Amazon is doing to address issues related to diversity, equity, and inclusion: https://www.amazon.science/tag/diversity-and-inclusion
Panelists: Juan Huerta – science manager, Search and AI Julian Pachon - director, Last Mile Science Darwin Villagomez – applied science manager Walterio Mayol – principal research scientist Francisca Gil Ureta – applied scientist
About LatinX in AI: LatinX in AI is a non-profit organization focused on furthering AI innovation and resources for LatinX individuals globally. The organization supports research, development, infrastructure, and mentoring programs to boost innovation and capabilities of LatinX professionals working in artificial intelligence, machine learning, and data science. Learn more: latinxinai.org
#AmazonScience #LatinXinAI #ArtificialIntelligence #MachineLearningAmazon Science celebrates Pi DayAmazon Science2022-03-14 | To mark Pi Day this year, Amazon Science is utilized a Times Square billboard, normally used by Amazon Music, to honor scientists, engineers, and mathematicians past, present, and future. Learn more: https://www.amazon.science/latest-news/pi-day-2022
#PiDay #AmazonScienceAmazon Robotics and Hampton University panel discussion | Amazon ScienceAmazon Science2022-03-03 | Jovonia Thibert, director of robotics strategy at Amazon, talks with Alissa Harrison, vice president for Information Technology at Hampton; Jean Muhammad, chair of the Hampton computer science department; and Demetris Geddis, assistant dean at Hampton University. Learn more: https://www.amazon.science/working-at-amazon/how-chance-encounters-sparked-a-career-in-engineering-and-robotics
#AmazonScience #AmazonRobotics #HamptonUniversity #Robotics #ArtificialIntelligence #AIConversational recommendations for Alexa | RecSys 2021 | Amazon ScienceAmazon Science2022-01-10 | During the ACM Recommender Systems (RecSys 2021) conference, the premier international forum for the presentation of new research results, systems and techniques in the broad field of recommender systems, Francois Mairesse, a senior applied scientist at Amazon Music, gave a session on his team's latest research around conversational recommendations for Alexa. Learn more about the session here: https://www.amazon.science/conferences-and-events/recsys-2021
#AmazonScience #AmazonMusic #Alexa #RecSys #ArtificialIntelligence #MachineLearningThe challenge and promise of quantum computing | Amazon ScienceAmazon Science2021-12-15 | During NeurIPS 2021, seven quantum computer scientists from Amazon came together to discuss the current state of quantum computing, some of the biggest challenges facing the field, and what the future might hold. Learn more: https://www.amazon.science/videos-webinars/amazon-quantum-scientists-and-academics-discuss-the-challenge-and-the-promise-of-quantum-computing
Panelists included: • Simone Severini, director of quantum computing • Antia Lamas-Linares, principal research scientist • Earl Campbell, senior research scientist • John Preskill, Amazon Scholar • Katharine Hyatt, applied scientist • James Whitfield, Amazon Visiting Academic • Helmut Katzgraber, senior practice manager
#AmazonScience #Quantum #QuantumComputing #QuantumTech #NeurIPSA conversation with AI luminaries | NeurIPS 2021 | Amazon ScienceAmazon Science2021-12-02 | Amazon Scholars Michael I. Jordan and Michael Kearns, and Amazon vice president and distinguished scientist Bernhard Schölkopf discuss the future of AI ahead of NeurIPS 2021. Watch the recorded event where these industry luminaries cover a range of topics including the history of ML in the past decade, its social impacts, the role of causal reasoning in ML, and whether or not autonomous, general-purpose intelligence should really be the aim of AI. Learn more: https://www.amazon.science/videos-webinars/neurips-luminaries-on-the-future-of-ai
#AmazonScience #NeurIPS #NeurIPS2021 #ArtificialIntelligence #MachineLearning #AI #MLInnovation and diversity panel discussion | INFORMS 2021 | Amazon ScienceAmazon Science2021-10-29 | During INFORMS 2021, Amazon's Supply Chain Optimization and Technologies (SCOT) team held a virtual panel discussion on innovation and diversity at Amazon. Moderated by Alexandra Jovicic, senior manager of software development, the discussion focused on the meaning of diversity, and how it relates to innovation. Watch the video to hear about their personal experiences, and learn more about diversity, equity, and inclusion at Amazon: https://www.amazon.science/tag/diversity-and-inclusion
Panelists:
• Ilse Debruin, director of global logistics • Juan Huerta, senior manager of research science • Julian Pachon, director or Last Mile science • Stephanie Collett, director of sourcing
#AmazonScience #INFORMS2021 #DEIAlexa & Friends with Pradeep Natarajan | Amazon ScienceAmazon Science2021-10-29 | Pradeep Natarajan, Amazon principal applied scientist, discusses his work with Alexa, and the significance machine learning has had in the field computer vision and deep neural networks. He also covers his career, and his participation in the ICCV 2021 workshop "Instance-level recognition", which focused on artworks, landmarks, and products. Learn more: https://www.amazon.science/videos-webinars/alexa-friends-features-pradeep-natarajan-alexa-ai-principal-applied-scientist
#AmazonScience #ComputerVision #AlexaFrom large pre-trained language models discovering linguistic structure towards foundation modelsAmazon Science2021-10-22 | October 2021, Chris Manning, the inaugural Thomas M. Siebel Professor in Machine Learning in the Departments of Linguistics and Computer Science at Stanford University, and director of Stanford’s Artificial Intelligence Laboratory (SAIL), gave a keynote presentation at Amazon's annual, internal machine learning conference.
His talk considers the following setup: a ML system can interact with an expensive oracle (the “real world”) by iteratively proposing batches of candidate experiments and then obtaining a score for each experiment (“how well did it work?”). The data from all the rounds of queries and results can be used to train a proxy for the oracle, a form of world model. The world model can then be queried (much more cheaply than the world model) in order to train (in-silico) a generative model which proposes experiments, to form the next round of queries.
Systems which can do that well can be applied in interactive recommendations, to discover new drugs, new materials, control plants or learn how to reason and build a causal model. They involve many interesting ML research threads, including active learning, reinforcement learning, representation learning, exploration, meta-learning, Bayesian optimization, black-box optimization.
What should be the training criterion for this generative model? Why not simply use Monte-Carlo Markov chain (MCMC) methods to generate these samples? Is it possible to bypass the mode-mixing limitation of MCMCs? How can the generative model guess where good experiments might be before having tried them? How should the world model construct a representation of its epistemic uncertainty, i.e., where it expects to predict well or not?
On the path to answering these questions, Chris introduces a new and exciting deep learning framework called GFlowNets which can amortize the very expensive work normally done by MCMC to convert an energy function into samples and opens the door to fascinating possibilities for probabilistic modeling, including the ability to quickly estimate marginalized probabilities and efficiently represent distributions over sets and graphs.
#AmazonScience #MachineLearningAutomatic movie analysis and summarization via turning point | Amazon ScienceAmazon Science2021-10-22 | October 2021, Mirella Lapata, professor in the School of Informatics at the University of Edinburgh, whose research focuses on probabilistic learning techniques for natural language understanding and generation, gave a keynote presentation at Amazon's annual machine learning conference.
Mirella's talk centers on Movie analysis as an umbrella term for many tasks aiming to automatically interpret, extract, and summarize the content of a movie. Potential applications include generating shorter versions of scripts to help with the decision-making process in a production company, enhancing movie recommendation engines, and notably generating movie previews.
Mirella introduces the task of turning point identification as a means of analyzing movie content. According to screenwriting theory, turning points (e.g., change of plans, major setback, climax) are crucial narrative moments within a movie: they define its plot structure, determine its progression and segment it into thematic units.
She argues that turning points and the segmentation they provide can facilitate the analysis of long, complex narratives, such as screenplays. Mirella further formalizes the generation of a shorter version of a movie as the problem of identifying scenes with turning points and present a graph neural network model for this task based on linguistic and audiovisual information.
She ends her discussion on why the representation of screenplays as (sparse) graphs offers interpretability and exposes the morphology of different movie genres.
#AmazonScience #MachineLearningOpen problems in machine learning | Amazon ScienceAmazon Science2021-10-22 | October 2021, Rama Chellapa, a Bloomberg Distinguished Professor in the Departments of Electrical and Computer Engineering and Biomedical Engineering at Johns Hopkins University, gave a keynote presentation at Amazon's annual machine learning conference. Learn more: https://www.amazon.science/videos-webinars/amazons-annual-machine-learning-conference-featured-presentations-from-thought-leaders-within-academia
Rama discusses his group's recent works on building operational systems for face recognition and action recognition using deep learning. While reasonable success can be claimed, many open problems still remain to be addressed. These include bias detection and mitigation, domain adaptation and generalization, learning from unlabeled data, handling adversarial attacks, and selecting the best subsets of training data in mini-batch learning. Some of Rama's recent works addressing these challenges will be summarized.
#AmazonScience #MachineLearningAnomaly detection for OOD and novel category detection | Amazon ScienceAmazon Science2021-10-22 | October 2021, Tom Dietterich, Emeritus Professor of Computer Science at Oregon State University, and considered one of the pioneers in the machine learning field, gave a keynote presentation at Amazon's annual machine learning conference.
Thomas discussed how every deployed learning system should be accompanied by a competence model that can detect when new queries fall outside its region of competence.
His presentation explores the application of anomaly detection to provide a competence model for object classification in deep learning. He considers two threats to competence: queries that are out-of-distribution and queries that correspond to novel classes.
Thomas reviews the four main strategies for anomaly detection and then surveys some of the many recently-published methods for anomaly detection in deep learning. The central challenge is to learn a representation that assigns distinct representations to the anomalies.
The talk concludes with a discussion of how to set the anomaly detection threshold to achieve a desired missed-alarm rate without relying on labeled anomaly data.
#AmazonScience #MachineLearningGFlowNets for generative active learning | Amazon ScienceAmazon Science2021-10-22 | October, 2021, Yoshua Bengio, one of the world’s leading experts in artificial intelligence, gave a keynote presentation at Amazon's annual machine learning conference. Yoshua is a professor in the Department of Computer Science and Operations Research at the Université de Montréal and scientific director of the Montreal Institute for Learning Algorithms (MILA).
In Yoshua's presentation, he considers the following setup: a ML system can interact with an expensive oracle (the “real world”) by iteratively proposing batches of candidate experiments and then obtaining a score for each experiment (“how well did it work?”).
The data from all the rounds of queries and results can be used to train a proxy for the oracle, a form of world model. The world model can then be queried (much more cheaply than the world model) in order to train (in-silico) a generative model which proposes experiments, to form the next round of queries.
Systems which can do that well can be applied in interactive recommendations, to discover new drugs, new materials, control plants or learn how to reason and build a causal model. They involve many interesting ML research threads, including active learning, reinforcement learning, representation learning, exploration, meta-learning, Bayesian optimization, black-box optimization.
What should be the training criterion for this generative model? Why not simply use Monte-Carlo Markov chain (MCMC) methods to generate these samples? Is it possible to bypass the mode-mixing limitation of MCMCs? How can the generative model guess where good experiments might be before having tried them? How should the world model construct a representation of its epistemic uncertainty, i.e., where it expects to predict well or not?
On the path to answering these questions, he introduces a new and exciting deep learning framework called GFlowNets which can amortize the very expensive work normally done by MCMC to convert an energy function into samples and opens the door to fascinating possibilities for probabilistic modeling, including the ability to quickly estimate marginalized probabilities and efficiently represent distributions over sets and graphs.
#AmazonScience #MachineLearning40+ years of computer vision with Gérard Medioni, Amazon VP / distinguished scientistAmazon Science2021-10-14 | On Oct. 14, Gerard Medioni, Amazon vice president and distinguished scientist, and emeritus professor at the University of Southern California (USC), presented a talk, “40+ Years of Computer Vision: A Personal Journey”, in which he discusses his work within academia, startups, and Amazon. Learn more: https://www.amazon.science/videos-webinars/gerard-medioni-believes-now-is-a-golden-age-for-computer-vision-research
In this one-hour presentation (view full presentation above), Medioni explains his “non Ivory Tower” approach to research during his USC academic career, five primary lessons learned while working with and for startups, and finally two examples — Amazon Go and Amazon One — of inventing on behalf of customers by starting with a clear understanding of the customer problem, and working backwards to address it.
#AmazonScience #ComputerVisionCareer advice from Amazon Robotics recruitersAmazon Science2021-10-08 | During the International Conference on Intelligent Robots and Systems (IROS 2021), which showcases leading-edge research in robotics, Amazon recruiters Amanda Papp and Jayesh Hirani held a virtual discussion on careers in robotics at Amazon. Watch the video to find out what kind of roles Amazon recruits for, the types of skills that are needed, and what advice they have for job interviews, and more. Learn more: https://www.amazon.science/working-at-amazon/amazon-robotics-ai-leaders-believe-now-is-a-particularly-good-time-to-explore-careers-in-robotics
#AmazonScience #AmazonRobotics #RoboticsQ&A with Sidd Srinivasa, director of Amazon Robotics AIAmazon Science2021-10-08 | On October 6, Siddhartha (Sidd) Srinivasa, director of Amazon Robotics AI, joined Nia Jetter, senior principal technologist, to discuss the field of robotics, Amazon robotics initiatives, where they get their ideas from, and advice on starting a career in robotics. Read the full article: https://www.amazon.science/working-at-amazon/amazon-robotics-ai-leaders-believe-now-is-a-particularly-good-time-to-explore-careers-in-robotics
Srinivasa joined Amazon as director of Robotics AI in 2018, and since 2017 has been the Boeing Endowed Professor at the School of Computer Science and Engineering at the University of Washington. Prior to that, he was the Finmeccanica Associate Professor at the Robotics Institute at Carnegie Mellon University where he founded the Personal Robotics Lab in December 2005. Srinivasa, who describes himself as “a full-stack roboticist, with a focus on robotic manipulation”, has worked in the robotics field since 1999.
Jetter, who earned a bachelor of science degree from MIT in math with computer science, and a master's degree in aeronautical and astronautical engineering from Stanford University, joined Amazon earlier this year. Previously, she spent 20 years in the aerospace Industry, including more than 18 years at Boeing where she rose to become a technical fellow in autonomy and AI.
#AmazonScience #AmazonRobotics #RoboticsAlexa & Friends with Jasha Droppo | Amazon ScienceAmazon Science2021-09-24 | Jasha Droppo, senior principal applied scientist in Amazon's Alexa AI team, discusses his work with Alexa and the significance neural networks and deep learning have had on the field of speech recognition. Learn more: https://www.amazon.science/videos-webinars/alexa-friends-features-jasha-droppo-amazon-alexa-ai-senior-principal-applied-scientist
#AmazonScience #ArtificialIntelligence #MachineLearningAdvice for scientists considering a role in Alexa AI | Amazon ScienceAmazon Science2021-09-14 | In this video, four Amazon scientists share their career advice, experiences, and tips for finding a job in the Alexa AI team, whose goal is to make voice interfaces ubiquitous and as natural as speaking to a human. Learn more: https://www.amazon.science/tag/alexa
#AmazonScience #AmazonAlexa #ArtificialIntelligenceAlexa & Friends with Yang Liu | Amazon ScienceAmazon Science2021-08-26 | Yang Liu, Amazon principal applied scientist, discusses her work in speech recognition and understanding, prosody modeling, summarization, and natural language processing. She also covers her career, 2021 ISCA Fellow and 2021 IEEE Fellow awards, how commonsense inference for smooth and effective communication is impacting human-like language responses, the growing popularity of virtual assistants, and the challenges that presents for entity resolution. Learn more: https://www.amazon.science/videos-webinars/alexa-friends-features-yang-liu-principal-applied-scientist-alexa-ai
#AmazonScience #ArtificialIntelligence #MachineLearningAmazon Scholars at KDD 2021Amazon Science2021-08-25 | During the Knowledge Discovery and Data Mining Conference (KDD 2021), three Amazon Scholars held a panel discussion about their projects, best practices, and how they leverage their experience at Amazon. Moderated by Academic Talent Partner, Larissa Kmiotek, this panel provides insight into ways academics can partner with Amazon in this domain and beyond.
Academics at Amazon: https://www.amazon.science/academics-at-amazon KDD 2021 event page: https://www.amazon.science/conferences-and-events/kdd-2021
#AmazonScience #KDD2021The science behind Alexa Conversations | Amazon ScienceAmazon Science2021-08-19 | A demo on the science behind Alexa Conversations, a new approach for building goal-oriented dialogue systems that is scalable, extensible as well as data efficient. This approach provides out-of-the-box support for natural conversational phenomena like entity sharing across turns or users changing their mind during conversation without requiring developers to provide any such dialogue flows. The demo is based on the NAACL-HLT 2021 Systems Demonstrations paper, "Alexa Conversations: An extensible data-driven approach for building task-oriented dialogue systems." Learn amore about the science behind Alexa here: https://www.amazon.science/tag/alexa
Speakers: Chien-wei Lin and Shuyang Gao
Authors: Anish Acharya, Suranjit Adhikari, Sanchit Agarwal, Vincent Auvray, Nehal Belgamwar, Arijit Biswas, Shubhra Chandra, Tagyoung Chung, Maryam Fazel-Zarandi, Raefer Gabriel, Shuyang Gao, Rahul Goel, Dilek Hakkani-Tür, Jan Jezabek, Abhay Jha, Jiun-yu Kao, Prakash Krishnan, Peter Ku, Anuj Goyal, Chien-Wei Lin, Qing Liu, Arindam Mandal, Angeliki Metallinou, Vishal Ishwar Naik, Yi Pan, Shachi Paul, Vittorio Perera, Abhishek Sethi, Minmin Shen, Nikko Ström, and Eddie Wang.
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterResearch and engineering of robust machine learning systems | Amazon ScienceAmazon Science2021-08-19 | Tom Diethe, applied science manager in Amazon AWS, discusses solutions for "technical debt" when deploying ML systems at the Bristol Data Science Seminar series. He covers engineering-based solutions available through AWS SageMaker cloud machine learning services that ensures ML systems are efficient in long-term usage, robust to changes in the environment, and potential errors are discoverable.
Learn more about Tom's work at Amazon here: https://www.amazon.science/author/tom-diethe
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterAlexa Prize Socialbot Grand Challenge 4 Finals | Amazon ScienceAmazon Science2021-08-16 | The Alexa Prize, launched by Amazon in 2016, is a competition for university students dedicated to advancing the field of conversational AI. Teams are challenged to design socialbots that Alexa customers can interact with via Alexa-enabled devices. Their ultimate goal is to meet the Grand Challenge: earn a composite score of 4.0 or higher (out of 5) from the judges, and have the judges find that at least two-thirds of their conversations with the socialbot in the final round of judging remain coherent and engaging for 20 minutes. Learn more: https://www.amazon.science/academic-engagements/czech-technical-university-team-wins-alexa-prize-socialbot-grand-challenge-4
#AmazonScience #AlexaPrizeA live chat with Prime Video scientists | Amazon ScienceAmazon Science2021-08-13 | Join four of scientists and engineers from Prime Video, and find out how they build the future of digital entertainment for millions of Amazon customers. Learn how they conduct customer-obsessed science in their team, and how the company's leadership principals guide their decisions every day. If you're interested in joining Amazon, check out our available career opportunities in science here: https://www.amazon.science/careers
Speakers: • Ali Roshan Ghias, applied science manager • Caren Chen, senior applied scientist • Raffay Hamid, principal scientist • Yongjun Wu, senior principal software development engineer
#AmazonScience #PrimeVideoThe history of Amazon’s forecasting algorithmAmazon Science2021-08-09 | Today, after a decade-long journey, Amazon’s Supply Chain Optimization Technologies forecasting team has drawn on advances in fields like deep learning, image recognition and natural language processing to develop a forecasting model that makes accurate decisions across diverse product categories. Read the full story: https://www.amazon.science/latest-news/the-history-of-amazons-forecasting-algorithm
Ping Xu, forecasting science director; Kari Torkkola, senior principal research scientist; Dhruv Madeka, principal applied scientist; and Ruofeng Wen, senior applied scientist, were among those who worked to unify Amazon's forecasting model.
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterSyntiant CEO on the future of ambient computing | Amazon Alexa Startups ShowcaseAmazon Science2021-08-08 | During the Amazon Alexa Startups Showcase, Kurt Busch, CEO of Syntiant, an Alexa Fund company, explained how they're using the latest in voice technology to invent the future of ambient computing, and why he thinks voice will be the next user interface. Learn more: https://www.amazon.science/latest-news/3-questions-with-jeremy-holleman-how-to-design-and-develop-ultra-low-power-ai-processors
The #AlexaFund provides up to $200 million in venture capital funding to fuel voice technology innovation. We believe experiences designed around the human voice will fundamentally improve the way people use technology. Learn more: https://www.amazon.science/tag/alexa-fund
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletter #AmazonScienceAlexa & Friends with Julia Hirschberg | Amazon ScienceAmazon Science2021-07-23 | On July 22, Julia Hirschberg, an Amazon Scholar and professor of computer science at Columbia University, joined Jeff Blankenburg, principal Alexa evangelist, on Alexa & Friends to discuss her work in speech and natural language processing, and her ongoing research into one of the biggest unresolved challenges for dialogue systems: detecting what people are feeling and addressing it appropriately. Learn more: https://www.amazon.science/videos-webinars/alexa-friends-features-amazon-scholar-julia-hirschberg
In August 2020, Hirschberg joined the Alexa AI Natural Understanding organization, which is working on multimodal enhancements for Alexa and developing models to identify dialogue acts using speech as well as text.
Hirschberg is the Percy K. and Vida L. W. Hudson Professor of Computer Science at Columbia University and was Chair of the Computer Science Department at Columbia from 2012 to 2018.
She earned her second PhD in computer science from the University of Pennsylvania. She worked at Bell Laboratories and AT&T Laboratories for 19 years as a member of technical staff and a department head, creating the Human-Computer Interface Research Department.
Hirschberg has spent most of her career studying speech and the intonation in text-to-speech synthesis and how to improve it. Since 2016 she has been working on diversity issues with the International Speech Communication Association, which runs the Interspeech conference.
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterComputer vision at Amazon | CVPR 2021 | Amazon ScienceAmazon Science2021-07-15 | During the annual conference on Computer Vision and Pattern Recognition (CVPR 2021), Amazon Scholars and Amazon Visiting Academics hosted a panel discussion on computer vision at Amazon. In the video, they discuss how they leverage their CV expertise and apply their research at Amazon to innovate on behalf of our customers. Moderated by Academic Program Manager, Lindsey Weil, this panel provides insight into ways academics can partner with Amazon in this domain.
Panelists: • Srinath Sridhar, Amazon Visiting Academic & Assistant Professor at Brown University • Thomas Brox, Amazon Scholar & Professor at University of Freiburg • Zhuowen Tu, Amazon Scholar & Professor at UC San Diego
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterMeet our Alexa AI scientists | Amazon ScienceAmazon Science2021-07-01 | What’s it like to be a scientist in Amazon’s Alexa AI team? In this video, we interview five scientists from across the Alexa organization to find out what they love about working here, and what customer-obsessed science they’re doing to create a better experience for Alexa’s bilingual customers. If you're interested in joining Amazon, check out our available career opportunities in science here: https://www.amazon.science/careers
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterMeet our Consumer Science team | CVPR 2021 | Amazon ScienceAmazon Science2021-06-30 | During the annual conference on Computer Vision and Pattern Recognition (CVPR 2021), three Amazon scientists from the company's Consumer Science team held a virtual discussion about their work in computer vision. Learn more about Amazon's presence at the conference on our event page: https://www.amazon.science/conferences-and-events/cvpr-2021
Panelists include:
• Frederick Devernay, senior applied scientist, Imaging Tech • Michael Lou, senior applied science manager, Visual Search • Aleix Martinez, senior applied science manager, Home Innovation
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterAlexa & Friends with Ariya Rastrow | Amazon ScienceAmazon Science2021-06-25 | On June 24, senior principal scientist Ariya Rastrow joined principal Alexa evangelist Jeff Blankenburg on Alexa & Friends to discuss his expertise in speech recognition technologies and career with Alexa, from the early days of bringing the conversational AI technology to customers. The two discussed Rastrow’s career, some of the most pressing challenges in the field of automatic speech recognition and machine learning, and the interesting research and themes that came out of the ICASSP 2021 conference earlier in June.
Ariya Rastrow joined Amazon in 2012 as a member of the initial team that built Echo and Alexa and has served as a technical lead building ASR and NLP technologies from ground up, and building low-latency ASR runtime decoder (both on edge and for cloud) and scalable language modeling technologies for Alexa. Rastrow earned his PhD at the Center for Language and Speech Processing (CLSP) at the Johns Hopkins University and has more than 15 years of hands-on experience working on automatic speech recognition and machine learning.
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterWomen in conversational AI panel discussion | NACCL 2021 | Amazon ScienceAmazon Science2021-06-16 | Amazon was a Platinum sponsor of the annual conference of the North American Chapter of the Association for Computational Linguistics (NAACL 2021), a conference focused on research and professional networking for scientists in computational linguistics.
Following the close of the conference, on June 15th, Professor and Amazon Scholar Heng Ji, AWS applied scientist Anna Currey, AWS senior research scientist Rashmi Gangadharaiah, Alexa AI senior applied scientist Maryam Fazel-zarandi and Alexa AI principal applied scientist Yang Liu hosted a discussion on the importance of diversity in STEM.
Watch the replay of the 45-minute discussion where you’ll hear directly from the panelists on topics including their career paths, the interesting and unique challenges they are tackling in their respective roles, and the importance of mentorship and opening doors for women in science. Learn more: https://www.amazon.science/videos-webinars/women-in-conversational-ai-virtual-panel-discussion
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterAlexa & Friends with Nikko Ström | Amazon ScienceAmazon Science2021-05-28 | On May 27, Nikko Ström, Alexa AI vice president and distinguished scientist, joined principal Alexa evangelist Jeff Blankenburg on Alexa & Friends to discuss his broad portfolio of work within conversational AI, including speech recognition, machine translation, natural language understanding, and dialogue management.
Watch the replay of the live discussion where the two discuss Ström’s career, some of his past research, and where he sees the field of conversational AI headed next.
Ström is a technologist and scientist with a deep background in speech technologies. He was a founding member of the team that built Amazon Echo and Alexa. Ström has more than 20 years of experience in the field of automatic speech recognition, working for some of the most prominent research laboratories and companies in the world, and has published extensively.
Learn more here: https://www.amazon.science/videos-webinars/alexa-friends-features-nikko-strom-alexa-ai-vice-president-and-distinguished-scientist
Follow us: Website: https://www.amazon.science/ Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterMeet our scientists | Amazon ScienceAmazon Science2021-05-27 | Amazon employs scientists from around the world to work on large-scale technical challenges in artificial intelligence (AI) and machine learning (ML). Meet four scientists from our retail, ML Solutions Lab, recommendation systems, and Amazon Go teams to learn more about what they're working on, and what gets them excited to drive customer-obsessed scientific innovation at Amazon. If you're interested in joining the company, check out our latest career opportunities in science here: https://www.amazon.science/careers
• Sneha Rajana, applied scientist • Ravi Shanker, senior data scientist • Antonia Schulze, data scientist • Carl Morris, research scientist
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterAlexa & Friends with Spyros Matsoukas | Amazon ScienceAmazon Science2021-04-23 | On April 22, senior principal applied scientist Spyros Matsoukas joined principal Alexa evangelist Jeff Blankenburg on Alexa & Friends to discuss his broad portfolio of work on Alexa, including speech recognition, machine translation, natural language understanding, and dialogue management.
Watch the replay of the live discussion where the two discuss Matsoukas’s career, what inspired him to pursue a career in conversational AI, and what he sees as the future for the field.
Matsoukas has been with the Alexa AI organization for more than seven years, helping to launch Alexa, expand voice technology into new languages, and enable Alexa to learn from customers and respond more accurately to their requests.
Before joining Amazon, he worked at BBN Technologies, conducting research in acoustic modeling for ASR, speaker diarization, statistical machine translation, speaker identification, and language identification.
He has contributed to more than 60 publications in peer reviewed conferences and journals, with three best paper awards.
Matsoukas earned a master’s degree in computer science from Northeastern University, as well as a master’s degree in computer engineering from the University of Patras, Greece.
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterThe innovation behind Amazon’s delivery servicesAmazon Science2021-03-31 | Interested in learning how Amazon uses science and engineering for its delivery services? Watch the replay from our LinkedIn Live discussion featuring four leaders from the company’s Last Mile, Fulfillment Technology, and Robotics teams. The speakers discuss the ways each team works to develop new solutions, keep up with increased consumer demand, maintain a focus on sustainability, and more.
Speakers include: • Josephine Bolotski, Principal Engineer, Robotics AI • Beryl Tomay, Vice President, Last Mile • John Darrow, Sr. Principal Engineer, Amazon Fulfillment Technology • Aaron Parness, Principal, Research Science, Robotics AI
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterAlexa & Friends with Yoelle Maarek | Amazon ScienceAmazon Science2021-03-26 | On March 25, Yoelle Maarek, vice president of research and science for Alexa Shopping, joined Alexa & Friends to talk more about the new Alexa Prize TaskBot Challenge, where university teams compete to develop agents that assist customers in completing tasks requiring multiple steps and decisions. Learn more: amzn.to/yoelle
Maarek earned her PhD in computer science at the Technion in Israel and an engineering degree from Ecole Nationale des Ponts et Chaussées in Paris, France. She has played a pioneering role in researching the fields of information retrieval — the computer science discipline behind search — web search, and data mining. She is an ACM Fellow, and a member of the Technion Board of Governors and management council, and was recently elected member of the National Academy of Engineering.
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterWhats it like to be a scientist at Amazon?Amazon Science2021-03-23 | What's it like to be a scientist at Amazon? In this video, we interview five scientists from across the company to find out what they love about working at Amazon, and the type of customer-obsessed science and innovation they're responsible for in their role. If you're interested in joining at Amazon, check out our available career opportunities in science here: https://www.amazon.science/careers
• Kaylin Lee, Data Science Manager • Constantinos Papayiannis, Applied Scientist • Jin Ding, Applied Scientist • Patrick Taylor, Data Scientist • Julian Pachon, Science Director
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterPanelists discuss the Alexa Prize Challenge | WSDM 2021 | Amazon ScienceAmazon Science2021-03-11 | Amazon was a platinum sponsor of the 14th ACM International Conference on Web Search and Data Mining (WSDM 2021), a conference promoting best practices and advances in operations research, management science, and analytics to improve operational processes, decision-making and outcomes.
During the conference, seven Amazon scientists gathered for a roundtable event where Amazon Scholar Eugene Agichtein talked about the Alexa Prize Socialbot Grand Challenge from his perspective, as a faculty advisor to the Emory University team in the first two years of the competition. He also introduced the newly announced Alexa Prize TaskBot Challenge, in which university teams will compete to develop agents that assist customers in completing tasks requiring multiple steps and decisions.
Following Agichtein's presentation, the panelists participated in a Q&A panel discussion about the Alexa Prize, Amazon's collaborations with academia, their careers and scientific focus areas, and more.
Amazon panelists:
• Yoelle Maarek, VP Research and Science, Alexa Shopping • Eugene Agichtein, Amazon Scholar, Alexa Shopping • Liane Lewin-Eytan, Senior Manager, Applied Science, Alexa Shopping • Andrew Borthwick, Principal Applied Scientist, Amazon Selection and Catalog Systems • Karthik Subbian, Principal Applied Scientist, Search Science and Artificial Intelligence • Hugo Zaragoza, Senior Manager, Applied Science, Amazon Books • Nikhita Vedula, Applied Scientist, Alexa Shopping
Follow us: Website: https://www.amazon.science Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterAlexa & Friends with Kayoko Yanagisawa | Amazon ScienceAmazon Science2021-02-26 | Recently, the Alexa Text-to-Speech (TTS) team announced a new neural TTS technology that enables a multilingual model to use the same voice for Spanish and English responses. On Feb. 25, Alexa evangelist Jeff Blankenburg hosted Kayoko Yanagisawa, Amazon senior speech scientist, on the 'Alexa & Friends' Twitch show to discuss this recent development and other speech research topics. Watch the recorded interview to hear more about Yanagisawa’s career history, the TTS challenges that come with languages other than English, and the science behind Alexa’s Polyglot system. Yanagisawa has more than 15 years of experience in speech research, with a doctorate in experimental phonetics and a master’s in phonetics from University College London.
Follow us: Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterAmazon Science Live StreamAmazon Science2021-02-15 | ...Swami Sivasubramanian machine learning keynote | AWS re:Invent 2020Amazon Science2021-01-14 | In December, at AWS re:Invent 2020, Swami Sivasubramanian, vice president of machine learning, Amazon Web Services, delivered the first-ever machine learning keynote at the ninth annual conference, and the first to be held virtually.
During the nearly two-hour keynote, Sivasubramanian discussed the latest developments in AWS machine learning, invited guest speakers from the National Football League and Philips to discuss how their organizations are using AWS machine learning technologies, and had other AWS scientists join him to announce, demo, and dive deeper into some of the new technologies and services introduced.
Since his talk was posted to YouTube on Dec. 18, 2020, more than 250,000 people have viewed the video. This video is a short outtake of his talk; the full keynote is available for viewing here: youtube.com/watch?v=PjDysgCvRqY
Follow us: Twitter: twitter.com/AmazonScience Facebook: facebook.com/AmazonScience Instagram: instagram.com/AmazonScience LinkedIn: linkedin.com/showcase/AmazonScience Newsletter: https://www.amazon.science/newsletterFairness in AI panel discussion | NeurIPS 2020 | Amazon ScienceAmazon Science2020-12-11 | One important topic for the field of machine learning is fairness in AI, which has become a table-stake for ML platforms and services, driven by customer / business needs, regulatory / legal requirements and societal expectations. Researchers have been actively studying how to address disparate treatment caused by bias in the data and the resulting amplification of such bias by ML models, and how to ensure that the learned model does not treat subgroups in the population unfairly.
During NeurIPS 2020, five Amazon scientists working on these types of challenges gathered for a 45-minute virtual session to address the topic. Watch the recorded panel discussion here, where the scientists discuss how fairness applies to their areas of AI / ML research, the interesting studies and advancements happening in the space, and the collaborations they’re most excited to see occurring across the industry in an effort to advance fairness in AI.