MIT researchers developed a method boosting federated learning efficiency by 81%, enabling secure AI training on resource-constrained edge devices. This breakthrough could expand AI applications in healthcare and finance, bringing powerful models to small devices.
PwC's AI-driven annotation (AIDA) solution, built on AWS, streamlines contract analysis, reducing manual review time by up to 90%. AIDA combines large language models with automated extraction workflows to extract structured insights and provide context-specific answers, revolutionizing contract management.
AI agents utilizing the Model Context Protocol (MCP) gain diverse capabilities. Amazon Bedrock AgentCore Gateway offers centralized governance for agent-tool integration, while a serverless MCP proxy on AgentCore Runtime allows customizable controls for MCP traffic.
Poolside AI introduces Laguna M. 1 and Laguna XS. 2, MoE models with impressive performance metrics. Laguna XS. 2 showcases innovative efficiency decisions in architecture, offering unique features for practitioners.
AI bias in medical AI models can lead to misdiagnoses. New debiasing approach WRING aims to address bias in VLMs like OpenCLIP, avoiding the Whac-A-Mole dilemma.
NVIDIA Nemotron 3 Nano Omni on Amazon SageMaker JumpStart offers a unified multimodal model for intelligent applications. It simplifies agent workflows by processing video, audio, images, and text in a single inference pass, enhancing efficiency and reducing latency.
Migrating text agents to voice assistants with Amazon Nova 2 Sonic for natural, real-time interactions in various industries. Key differences in user input, response style, and latency budget must be considered for successful migration.
Machine learning regression models predict numeric values like credit scores. Various techniques like linear regression and neural networks can be used for training. Demo in C# language showcases different techniques for training linear regression models.
MOSS-Audio by OpenMOSS, MOSI. AI, and Shanghai Innovation Institute is an open-source model that unifies speech, sound, music understanding, and more. It consists of four variants optimized for different tasks, all powered by a modular architecture with an audio encoder, modality adapter, and large language model.
Popsa uses AI and design automation to create personalized Photo Books in minutes, enhancing user experience and satisfaction. By implementing Amazon Bedrock and Amazon Nova models, over 5.5 million personalized titles were generated in 2025, leading to increased engagement and purchase rates.
Amazon SageMaker AI endpoints provide organizations with control over compute resources and infrastructure placement, while leveraging the managed operational layer of AWS. Strands Agents SDK simplifies building AI agents, integrating with SageMaker AI models, and implementing A/B testing for continuous improvement.
AI growth will increase U.S. data center electricity use; MIT & IBM develop rapid power prediction tool for sustainable AI efficiency. Tool allows quick estimates for energy consumption, aiding data center operators and algorithm developers.
Refactoring matrix pseudo-inverse via normal equations simplifies machine learning code. Cholesky decomposition reduces complexity for training data matrices in ML scenarios.
LoRA struggles with capturing complex factual knowledge due to its low-rank updates. RS-LoRA stabilizes learning by adjusting the scaling formula, improving model retention of high-dimensional information.
Deloitte used Amazon EKS and vCluster to transform their testing infrastructure. Automated solution syncs S3 data with Amazon Bedrock Knowledge Bases, respecting service quotas and rate limits.