Senthilkumar Gopal’s Post

https://lnkd.in/gD5nJ_W5 Amazon Nova Micro, a text-only model that delivers the lowest latency responses at very low cost. Amazon Nova Lite, a very low-cost multimodal model that is lightning fast for processing image, video, and text inputs. Amazon Nova Pro, a highly capable multimodal model with the best combination of accuracy, speed, and cost for a wide range of tasks. Amazon Nova Premier, the most capable of Amazon’s multimodal models for complex reasoning tasks and for use as the best teacher for distilling custom models (available in the Q1 2025 timeframe). Amazon Nova Canvas, a state-of-the-art image generation model. Amazon Nova Reel, a state-of-the-art video generation model.

Much awaited launch announcement - Amazon Nova Family of Foundation models #AWSreInvent2024! Integrated with Amazon Bedrock, these models offer improved speed, cost-effectiveness (up to 75% less expensive), and customization options across 200 languages. The lineup includes: • 𝗔𝗺𝗮𝘇𝗼𝗻 𝗡𝗼𝘃𝗮 𝗠𝗶𝗰𝗿𝗼, a text-only model that delivers the lowest latency responses at very low cost. • 𝗔𝗺𝗮𝘇𝗼𝗻 𝗡𝗼𝘃𝗮 𝗟𝗶𝘁𝗲, a very low-cost multimodal model that is lightning fast for processing image, video, and text inputs. • 𝗔𝗺𝗮𝘇𝗼𝗻 𝗡𝗼𝘃𝗮 𝗣𝗿𝗼, a highly capable multimodal model with the best combination of accuracy, speed, and cost for a wide range of tasks. • 𝗔𝗺𝗮𝘇𝗼𝗻 𝗡𝗼𝘃𝗮 𝗣𝗿𝗲𝗺𝗶𝗲𝗿, the most capable of Amazon’s multimodal models for complex reasoning tasks and for use as the best teacher for distilling custom models (available in the Q1 2025 timeframe). • 𝗔𝗺𝗮𝘇𝗼𝗻 𝗡𝗼𝘃𝗮 𝗖𝗮𝗻𝘃𝗮𝘀, a state-of-the-art image generation model. • 𝗔𝗺𝗮𝘇𝗼𝗻 𝗡𝗼𝘃𝗮 𝗥𝗲𝗲𝗹, a state-of-the-art video generation model. https://lnkd.in/eRnrn2wP

🌟 𝗔𝗺𝗮𝘇𝗼𝗻 𝗡𝗼𝘃𝗮: 𝗔 𝗚𝗮𝗺𝗲-𝗖𝗵𝗮𝗻𝗴𝗲𝗿 𝗶𝗻 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 🌟 Amazon introduces #Nova, a groundbreaking suite of foundation models delivering unmatched 𝘀𝗽𝗲𝗲𝗱, 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆, 𝗮𝗻𝗱 𝗰𝗼𝘀𝘁-𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆—𝗮𝘁 𝟳𝟱% 𝗹𝗼𝘄𝗲𝗿 𝗰𝗼𝘀𝘁𝘀 compared to top-performing alternatives. With multimodal capabilities (text, image, and video), Nova powers innovative tools like 𝗔𝗜 𝗦𝗵𝗼𝗽𝗽𝗶𝗻𝗴 𝗚𝘂𝗶𝗱𝗲𝘀 and creative solutions like 𝗡𝗼𝘃𝗮 𝗖𝗮𝗻𝘃𝗮𝘀 𝗮𝗻𝗱 𝗥𝗲𝗲𝗹 for businesses to scale with ease. Integrated with Amazon bedrock, Nova competes head-to-head with offerings from OpenAI, Google, and Microsoft, setting a new standard in affordability and performance. #ArtificialIntelligence #Technology #Innovation #MachineLearning #AWS

To improve Amazon’s recommendation engine, Amazon researchers are building a knowledge graph that encodes commonsense relationships between products and queries. Next month at SIGMOD/PODS 2024, they'll present COSMO, a framework that uses #LLMs to extract those relationships. Blog post: https://lnkd.in/gac_Hrxr Paper: https://lnkd.in/ggTbr5aM #RecommenderSystems #LargeLanguageModels #ML

Great to have our CEO Andy at #AWSreInvent today to introduce Amazon Nova—a new generation of foundation models that delivers frontier intelligence and industry leading price performance, available today in #AmazonBedrock. These fast & incredibly capable #genai models have impressive results on a wide range of industry standard benchmarks, and give our customers even more choice for their specific needs. The options range from the text-only Amazon Nova Micro to multimodal models like Amazon Nova Lite, Amazon Nova Pro (and Premier coming soon), to Amazon Nova Canvas for incredible images, and Amazon Nova Reel for video generation. https://lnkd.in/g5mkiwJf

I'm delighted to share that my team pikachu secured 7th position in this year's Amazon ML Challenge among 2500+ teams and over 75,000 registered participants! 🎉 The challenge focused on developing machine learning techniques to extract crucial entity values like weight, volume and dimensions, directly from images. This has significant potential for industries like e-commerce, healthcare and content moderation, where complete product information is often lacking. For our solution, we utilized a Vision-Language Model (Qwen2-VL-7B) to perform direct inferencing on images, enabling us to extract values without the need for fine-tuning, ensuring greater generalizability across new and diverse images. This strategic choice also led us to select the 7B model, striking a balance between accuracy and resource efficiency, instead of the larger, resource-intensive 75B model. The top 10 teams from round 1 made it to the 2nd round based on leaderboard scores. In the grand finale, we had the opportunity to present our approach to Amazon scientists and the final rankings were announced based on this presentation. It was an incredible experience to share our work at this level! This experience provided a deep dive into cutting-edge AI techniques and has been a tremendous learning opportunity. A huge shoutout to my incredible teammates Rudransh Agarwal,Akshit Sharma and Advaith Malladi your effort and teamwork were key to our success! 🙌 Thank you Amazon for organizing this amazing challenge. I’m excited to continue exploring and pushing the boundaries of AI and machine learning! #MachineLearning #Amazon #AI #ComputerVision #IIITH #AmazonMLChallenge #Innovation #Teamwork

The latest release of Amazon Q Business enhances visibility into data source sync operations. The new document-level report provides administrators with granular indexing status, metadata, and ACL details for every document processed during a sync job. This feature allows for quick troubleshooting and resolution of ingestion or access issues, ensuring a seamless experience for Amazon Q Business users. How can this new feature improve your experience with Amazon Q Business? --- Hi, 👋🏼 my name is Doug, I love AI, and I post content to keep you up to date with the latest AI news. Follow and ♻️ repost to share the information! #amazonqbusiness #datasync #aiassistant

Introducing Amazon Nova: A New Generation of Foundation Models Amazon Nova Pro performed equal or better on 17 of 20 benchmarks compared to OpenAI’s GPT-4o, equal or better on 16 of 21 benchmarks compared to Google’s Gemini 1.5 Pro, and equal or better on 9 of 20 benchmarks compared to Anthropic Claude Sonnet 3.5v2. In addition to accuracy on text and visual intelligence benchmarks, Amazon Nova Pro excels at instruction-following and multimodal agentic workflows as measured by the Comprehensive RAG Benchmark (CRAG), the Berkeley Function Calling Leaderboard, and Mind2Web. Amazon Nova Micro, Amazon Nova Lite, and Amazon Nova Pro are at least 75 percent less expensive than the best performing models in their respective intelligence classes in Amazon Bedrock. They are also the fastest models in their respective intelligence classes in Amazon Bedrock.

Amazon has introduced Amazon Nova, a new series of foundation models integrated into AWS's Amazon Bedrock platform. The lineup includes models like Nova Micro, Nova Lite, Nova Pro, and the upcoming Nova Premier, each tailored to specific performance and cost needs. Designed for cost efficiency, some models are at least 75% less expensive than comparable alternatives. Supporting multimodal inputs, Amazon Nova can process and generate text, images, and videos, expanding their applicability. They offer fine-tuning capabilities, allowing customization with proprietary data to enhance accuracy. Through integration with Amazon Bedrock Knowledge Bases, the models utilize Retrieval-Augmented Generation (RAG) to ground responses in organizational data, improving precision. The development is supported by Amazon's custom-built Inferentia and Trainium chips, underscoring the company's commitment to advancing generative AI solutions that deliver tangible value. Though Amazon may have entered the AI race later than some, its latest release signals a game-changing move. Backed by a colossal customer base, an almost limitless budget, and cutting-edge models, the retail giant is positioning itself as the dark horse ready to leap to the top of the AI hierarchy. #FoundationModels #GenerativeAI #MultimodalAI #FineTuning #RetrievalAugmentedGeneration #AIInnovation