Today, we're excited to announce updates to Nova Act SDK. Since our research preview, we've seen incredible adoption from developers building complex, reliable agents. We're adding new enterprise-grade capabilities to our preview through AWS integration, enabling select customers to move from prototype to production with 90%+ end-to-end reliability on early enterprise use cases. Interested in productizing your agent prototype? Learn more: https://amzn.to/40PkFfq
Amazon Science
Research Services
Seattle, Washington 375,675 followers
The latest news and research from Amazon’s science community. #AmazonScience
About us
Amazon Science gives you insight into the company’s approach to customer-obsessed scientific innovation. Amazon fundamentally believes that scientific innovation is essential to being the most customer-centric company in the world. It’s the company’s ability to have an impact at scale that allows us to attract some of the brightest minds in artificial intelligence and related fields. Our scientists continue to publish, teach, and engage with the academic community, in addition to utilizing our working backwards method to enrich the way we live and work. Follow us on LinkedIn and visit our website to get a deep dive on innovation at Amazon, and explore the many ways you can engage with our scientific community. #AmazonScience
- Website
-
https://www.amazon.science
External link for Amazon Science
- Industry
- Research Services
- Company size
- 10,001+ employees
- Headquarters
- Seattle, Washington
- Founded
- 2020
- Specialties
- Artificial Intelligence, Machine Learning, Computer Vision, Cloud, Economics, Sustainability, AI, ML, Conversational AI, Natural Language Processing, NLP, Robotics, Security, Privacy, Information, Knowledge Management, Operations, Scientific Research, Search, Amazon, and Alexa
Updates
-
Announcing the winners of the Amazon Nova AI Challenge! 🏆 After six months of intense competition, we're proud to celebrate Team PurpCorn-PLAN (UIUC) as Defending Champions and Team PurCL (Purdue) as Attacking Champions, along with runners-up Team AlquistCoder (CTU Prague) and Team RedTWIZ (Nova Lisbon). This inaugural challenge showcased new safety techniques, adversarial tools, and multi-turn evaluation methods that push the frontier of secure, trustworthy AI. Learn more: https://amzn.to/4nYITxH
-
-
Introducing Mitra: a foundation model from Amazon researchers that outperforms traditional methods for tabular data by learning from diverse synthetic priors. Available in AutoGluon 1.4 soon, Mitra uses in-context learning to adapt to new tasks without requiring separate models for each dataset: https://amzn.to/4lBeowh
-
-
Amazon researchers developed a new architecture that reduces a foundation model's inference time by 30% while maintaining its accuracy. Like specialized regions in the brain, this new system selects appropriate subsets of neurons depending on the task: https://amzn.to/3IKPA6p
-
-
The Amazon Nova AI Challenge finals featured eight elite university teams competing to advance AI-assisted secure coding. In a tournament-style competition, attacking and defending teams matched up against each other, shaping the future of secure AI development. Watch our judges share their thoughts during the live finals and stay tuned for winner announcements on July 23:
-
That concludes another successful ICML. Thank you to everyone who engaged with our work and connected with our team throughout the conference. Until next time! #ICML2025
-
-
-
-
-
+1
-
-
New independent research from Aymara put 20 leading language models to the test against 10 real-world safety risks. The findings? A clear group of leaders emerged, including Amazon's Nova, demonstrating strong performance across categories like misinformation, impersonation, and privacy. We're proud to see Nova recognized for its safety-first architecture and our ongoing work to build trustworthy AI. Read the full breakdown: https://amzn.to/40mW4yr
-
At ICML, Amazon researchers are presenting SEAD: an anomaly detection model selection method that weights models by their "reluctance" to flag anomalies. Since anomalies are rare, the system rewards conservative models and outperforms 13 baselines across 15 tasks. Explore the research: https://amzn.to/4nOIVrN #ICML2025
-
-
Amazon researchers have accepted publications at ICML spanning generative AI, autonomous agents, time series forecasting, and anomaly detection. Explore the papers: https://amzn.to/450nqgn #ICML2025
-
Amazon researchers are presenting cutting-edge research in machine learning at ICML in Vancouver. We'll be showcasing live demos at our booth (#311), hosting workshops and talks/panels, presenting posters, and engaging with fellow scientists. Explore our accepted publications and discover how we're driving breakthroughs across diverse ML domains: https://amzn.to/4o0Oynd #ICML2025
-