A new method for discovering new molecules with specific properties, developed at MIT, utilizes artificial intelligence and machine learning to provide initial guidance on which compounds to synthesize and test. The approach, called CheMeleon, was built using ChemProp and has already found success in the pharmaceutical industry, outperforming other existing methods in areas where data on a cert...
In this text, the author discusses the challenges of defining primitives in a system where agents take actions instead of just answering questions. The author mentions the repetition of ideas under different names, such as an actor, signal, packaged procedure, and memory. The author introduces a taxonomy of primitives based on the questions "When?", "What?", "Who?", and "How?". Triggers are dis...
When deploying machine learning models on Amazon SageMaker GPU endpoints, the smallest available GPU unit is a whole accelerator. This limitation results in increased costs as each deployment requires a full GPU, limiting the number of models that can be hosted on an instance and potentially leading to underutilization or overpaying for capabilities. Amazon SageMaker Inference now supports GPU ...
At the Conference on Robot Learning in Austin, Texas, researchers are advancing robot capabilities beyond locomotion to manipulation, using simulation to teach finer tasks like aligning gears and assembling components. NVIDIA is introducing open robotics technologies including world action models, simulation tools, benchmarks, and agentic workflows. Two NVIDIA papers, GRAIL and RoboTTT, are fin...
In the context of regression analysis, a simple decision tree model may overfit training data, achieving high accuracy on it but poor performance on new, unseen data. To address this issue, Leo Breiman proposed bagging tree regression, or bootstrap aggregation, in 1994. This method creates an ensemble of multiple decision trees, each trained on a different subset of the training data using rand...
Researchers from MIT, UBC, and other institutions have developed a deep-learning system called WorldParticle that allows scientists and engineers to model various physical phenomena using a single unified artificial intelligence architecture. The system represents every material as tiny particles and predicts where they will move one step at a time, adapting to materials, forces, and environmen...
OpenAI's latest update to its image generation and editing model, GPT-Image-2.6, is now available on ChatGPT, Codex, and the API. The new models, Flare and Sunburst, offer faster image generation, with Flare generating images as quickly as 6 seconds at low quality. The update introduces GIF output, SVG export, markup-guided edits, and a new complexity parameter. ChatGPT users can create and edi...
Chris Alexiuk, a Toronto-based deep learning engineer at NVIDIA, is an advocate for open source AI and openly shares his knowledge through platforms like YouTube and LinkedIn. He believes that transparency is essential for the democratization of intelligence and has a strong affinity for NVIDIA's open source, multimodal models, Nemotron. Alexiuk refers to himself as the "No. 1 Nemotron Fan" and...
Amazon's A+ Content is a rich media widget on product detail pages. Brands use text, images, and video to communicate unique product and brand value, leading to sales boost of 8-20%. However, content quality can vary greatly due to the creative freedom allowed. This post introduces a new approach using Vision-Language Models (VLMs) to evaluate and improve A+ Content holistically, providing acti...
MIT researchers have developed a new memory system for robots that can identify and correct incorrect labels of objects in their memory. The system, called UQ-DAAAM, uses multiple descriptions of the same object from different viewpoints to quantify the robot's uncertainty and improve the accuracy of its descriptions. UQ-DAAAM compares these descriptions to reduce uncertainty and create a more ...
The author of this text expresses their enjoyment of writing code daily and shares their observation that engineering colleagues who transitioned from coding to management roles often lost their coding abilities quickly. They present a novel approach to combine quadratic regression (QR) and gradient boosting regression (GBR) by encapsulating these models in a container class. After independent ...
In the context of security leadership, the use of AI agents raises the question of how to maintain human oversight as tasks are handed over to these agents. Permissions can determine which actions agents can take, but even allowed actions may necessitate human judgment. To address this challenge, Amazon Quick's Smart Approvals evaluate each proposed action in relation to the user's request and ...
Enterprise AI is transforming decision-making for life sciences organizations across commercial, medical, market access, and other business functions. However, the success of every AI model, dashboard, and business insight hinges on one crucial factor: trusted data. Maintaining consistent data quality in today's complex data environments, spread across multiple platforms and business domains, i...
Advanced semiconductor manufacturing is becoming increasingly complex with shrinking error margins. Engineers need to accurately predict process behaviors, material interactions, and root causes of defects to optimize processes and minimize physical trials. AI is transforming semiconductor manufacturing through accelerated computing, physics-based AI, and digital twins. AI shortens simulation c...
In this architecture, an AI agent application interacts with enterprise APIs using Amazon Bedrock AgentCore and Microsoft Entra ID. The goal is to maintain user identity throughout the request chain. AgentCore Gateway, a component of AgentCore, enables on-behalf-of (OBO) token exchange with Microsoft Entra ID. The AI agent authenticates with the gateway using a JSON Web Token (JWT) scoped to it...