Amazon Science

Diffuse-to-Choose is a novel generative AI model from Amazon, which allows users to seamlessly insert any product at any location in any scene. The customer starts with a personal scene image and a product and draws a mask in the scene to tell the model where to insert the object. The model then integrates the item into the scene, with realistic angles, lighting, shadows, and so on. If necessary, the model infers new perspectives on the item, and it preserves the item’s fine-grained visual-identity details. Paper: https://lnkd.in/gKWjxWQv Blog post: https://lnkd.in/gwF5_2Dz #GenerativeAI #ComputerVision #ML

🚨 The deadline for submissions to the Amazon Research Awards spring 2024 call for proposals is today at 11:59pm PT. Learn more about the two research areas and how to apply: https://lnkd.in/ejWpCHDU #InfoSec #Sustainability #ML

We're proud to be a gold sponsor of this year's International Conference on Learning Representations (ICLR 2024) in Vienna, Austria. Find the Amazon team at booth 12 to hear from our scientists and learn about our career opportunities, and visit our website to view accepted publications and workshops. #ICLR2024 #DeepLearning

Scenario Diffusion uses generative AI to navigate safety-critical situations in autonomous vehicles by creating realistic driving scenarios to enhance safety testing, and ensure the reliability of the company’s robotaxis across diverse environments. Learn more about the method from Zoox. Paper: https://lnkd.in/e9n4iVBF Blog post: https://lnkd.in/eNnuQM4e #GenerativeAI #AutonomousVehicles #ML #Zoox

At this year's IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Amazon researchers won the Best Student Paper award for their publication, 'Significant ASR error detection for conversational voice assistants', which proposes a system that can determine, to a high degree of accuracy, whether the semantics of a predicted and reference transcript are significantly different. Read the paper: https://lnkd.in/g5NB7ApX #AmazonScience #ICASSP2024 #ConversationalAI

The AWS-funded projects span large language models, deep learning, and human-algorithm interactions, to help advance safe and responsible AI: https://lnkd.in/guy4ky5U #ResponsibleAI #LLMs #AWS

The Amazon Catalog team uses generative AI to make product information more useful and A/B testing to evaluate enriched data. Two team members explain how causal random forests and Bayesian structural time series help them extrapolate from sparse A/B data. #Economics #GenerativeAI #ABtesting

The recipients, representing 51 universities in 15 countries, will have access to Amazon public datasets, AWS AI/ML services and tools, and more. Congrats to the 99 awardees! #AmazonResearchAwards

🚀 AWS just launched new features that make scaling GenAI applications on Amazon Bedrock even easier: https://lnkd.in/gGsjQChp Here’s what today’s upgrades mean for your business:  🟠  Custom Model Import ➡️ Import your custom pre-trained models onto Amazon Bedrock and get all the same benefits, including seamless scalability and powerful capabilities to safeguard applications, adherence to responsible AI principles – as well as the ability to expand a model’s knowledge base with RAG, easily create agents to complete multi-step tasks, and carry out fine tuning to keep teaching and refining models. All without needing to manage the underlying infrastructure.  🟠 Simplify Model Selection ➡️ Compare models on Amazon Bedrock against specific datasets and evaluation metrics and choose the right model for your project goals.  🟠 Generate Studio-Quality Images ➡️ Produce images for advertising campaigns with Amazon Titan Image Generator using natural language prompts. As part of AWS’s commitment to promoting responsible and ethical AI, every imaged created by Titan contains an invisible watermark to increase transparency about AI-generated content.  🟠 Safeguard Application Development ➡️ Using Guardrails for Amazon Bedrock, customers can block up to 85% more harmful content than protection natively provided by FMs. Customers can define a set of topics or words that need to be removed in their Generative AI application, including harmful words, profanity, and personal and sensitive information. 🟠 Enjoy Expanded Model Choice ➡️ Meta Llama 3 FMs is now available on Amazon Bedrock. Llama 3 is designed for developers, researchers, and businesses to build, experiment, and responsibly scale their generative AI ideas. Cohere’s Command R and Command R+ models are coming soon. These models from Cohere are state-of-the-art FMs customers can use to build enterprise-grade generative AI applications with advanced RAG capabilities, in 10 languages, to support their global business operations.  We’re proud to simplify our customers’ GenAI journey with these latest suite of integrated development tools.   

At AAAI, Amazon scientists introduced a novel approach that significantly advances our ability to capture and model scenes with complex dynamics. Their work not only addresses previous limitations but also opens doors to new applications ranging from virtual reality to digital preservation: https://lnkd.in/e7FeKJnp #GenerativeAI #ComputerVision