# foundation-models Published articles for foundation-models. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Improving HCLS AI reasoning with open-source agent skills DevFeed: [Improving HCLS AI reasoning with open-source agent skills]() Original publisher: [Read original article]() Author: Michael Hsieh Published: 2026-09-16T19:00:00Z Content type: article Language: en Sources: [Artificial Intelligence]() Topics: [Agent Skills](), [Artificial Intelligence](), [foundation-models](), [Open Source](), [Bioinformatics](), [Medical imaging]() Tags: [agent-skills](), [agents](), [ai](), [ai-agents](), [amazon-bedrock](), [amazon-quick-suite](), [announcements](), [artificial-intelligence](), [foundation-models](), [generative-ai](), [healthcare](), [healthcare-and-life-sciences](), [intermediate-200](), [kiro](), [life-sciences](), [medical-imaging](), [open-source](), [strands-agents]() ### AI overview This post presents 38 open-source agent skills spanning 11 healthcare and life sciences domains. The skills encode domain decision procedures for AI agents, and the reported evaluation found a 70-86% head-to-head win rate over agents without the skills. ### Source excerpt AI agents on foundation models often misapply healthcare and life sciences decision frameworks, citing the right guideline but applying it incorrectly. This post shares 38 open-source agent skills across 11 HCLS domains that close this gap, with installation steps, three worked use cases, and a 410-prompt evaluation showing a 70-86% win rate. ## Optimizing cost and latency with Amazon Bedrock prompt caching DevFeed: [Optimizing cost and latency with Amazon Bedrock prompt caching]() Original publisher: [Read original article]() Author: Daniel Abib Published: 2026-09-15T16:18:19Z Content type: tutorial Language: en Sources: [Artificial Intelligence]() Topics: [Amazon Bedrock](), [Caching](), [Latency](), [API](), [foundation-models](), [LangChain](), [Multi-tenancy]() Tags: [advanced-300](), [amazon-bedrock](), [api](), [caching](), [foundation-models](), [integration](), [langchain](), [latency](), [pricing](), [technical-how-to](), [tokens]() ### AI overview This AWS post explains how Amazon Bedrock prompt caching can reduce repeated input-token costs by up to 90 percent and lower time to first token when requests reuse the same context. It presents six scenarios using the Converse API, including document, system prompt, tool definition, mixed TTL, tenant-isolated, and LangChain caching. ### Source excerpt Prompt caching in Amazon Bedrock can cut input token costs by up to 90% when you repeatedly send the same context to foundation models. This post walks through six practical prompt caching scenarios using the Converse API: message content, system prompt, tool definition, mixed TTL, tenant isolation, and LangChain integration. ## NASA-IBM Lunar Foundation Model Goes Open Source With a 2M-Tile Dataset and 22% Lower Ice-Mapping Error DevFeed: [NASA-IBM Lunar Foundation Model Goes Open Source With a 2M-Tile Dataset and 22% Lower Ice-Mapping Error]() Original publisher: [Read original article]() Author: Harold Fritts Published: 2026-09-14T16:43:16Z Content type: news Language: en Sources: [StorageReview.com]() Topics: [lunar foundation model](), [foundation-models](), [ibm](), [Open Source](), [hugging face](), [dataset](), [multimodal](), [Benchmark]() Tags: [ai](), [benchmark](), [data](), [enterprise](), [fine-tuning](), [foundation-models](), [hugging-face](), [ibm](), [lunar-foundation-model](), [machine-learning](), [multimodal](), [nasa](), [open-source](), [space]() ### AI overview IBM and NASA have released the NASA-IBM Lunar Foundation Model as open source on Hugging Face, along with its weights, technical report, and training dataset. Built on TerraMind, the model uses multimodal lunar observations for tasks including ice-deposit mapping, volcanic-feature detection, and crater detection. Reported benchmarks show up to 22% lower ice-mapping error than SwinV2-B, while the accompanying dataset contains roughly 2 million image tiles from nine instruments across four lunar missions. ### Source excerpt IBM and NASA have released the NASA-IBM Lunar Foundation Model as open source, one of the first publicly available foundation models built for scientific study of the Moon. The weights, a technical report, and the machine-learning-ready dataset it was trained on are up on Hugging Face under the Prithvi family, which already covers Earth observation, The post NASA-IBM Lunar Foundation Model Goes Open Source With a 2M-Tile Dataset and 22% Lower Ice-Mapping Error appeared first on StorageReview.com. ## The generative AI customization spectrum: From prompt engineering to custom models on AWS DevFeed: [The generative AI customization spectrum: From prompt engineering to custom models on AWS]() Original publisher: [Read original article]() Author: Bhavya Sruthi Sode Published: 2026-09-14T15:47:12Z Content type: tutorial Language: en Sources: [Artificial Intelligence]() Topics: [Generative AI](), [Amazon Web Services](), [Prompt Engineering](), [Fine-tuning](), [Retrieval Augmented Generation (RAG)](), [Amazon Bedrock](), [foundation-models](), [Anthropic Claude](), [Nova](), [llama]() Tags: [advanced-300](), [ai](), [amazon-bedrock](), [anthropic-claude](), [aws](), [best-practices](), [chain-of-thought](), [fine-tuning](), [foundation-models](), [generative-ai](), [llama](), [nova](), [optimization](), [prompt-engineering](), [rag]() ### AI overview This AWS article presents an eight-step decision framework for customizing generative AI workloads. It compares progressively more involved approaches, including prompt engineering, Retrieval Augmented Generation (RAG), fine-tuning, continued pre-training, and custom models such as Amazon Nova Forge, emphasizing that teams should start with the simplest approach and escalate when greater control or domain specificity is required. ### Source excerpt Pick the right generative AI customization approach on AWS with an 8-step decision framework, from prompt engineering and RAG to fine-tuning, continued pre-training, and Amazon Nova Forge. Start simple and escalate only when you must. ## Automate replenishment with MMF, Databricks Genie, and Amazon Quick DevFeed: [Automate replenishment with MMF, Databricks Genie, and Amazon Quick]() Original publisher: [Read original article]() Author: Venkatavaradhan Viswanathan Published: 2026-09-14T15:42:06Z Content type: article Language: en Sources: [Artificial Intelligence]() Topics: [databricks](), [Amazon S3 Tables](), [Amazon S3](), [Automation](), [foundation-models](), [API]() Tags: [advanced-300](), [amazon-quick-suite](), [amazon-s3](), [amazon-s3-tables](), [api](), [automation](), [databricks](), [foundation-models](), [machine-learning](), [retail](), [technical-how-to]() ### AI overview This technical walkthrough presents an unattended replenishment workflow for retail. Databricks Many Model Forecasting uses Chronos-2 to predict seven-day demand for each SKU, Databricks Genie detects demand surges, and Amazon Quick reconciles those surges with supplier availability in Amazon S3 Tables. The workflow places routine purchase orders through a Supplier Order API and escalates cases without a suitable single supplier for human review. ### Source excerpt Foundation models made catalog-wide demand forecasting easy; the hard part is now acting on the forecast. This post builds a closed detect-decide-act loop on Databricks and Amazon Quick that reconciles demand surges against live supplier availability and places replenishment orders unattended, escalating to a human only when no supplier can cover a surge. ## IBM releases SOTA Granite Time Series PatchTST-FM-r2 model with commercial-friendly license DevFeed: [IBM releases SOTA Granite Time Series PatchTST-FM-r2 model with commercial-friendly license]() Original publisher: [Read original article]() Author: Roman Vaculin; Wesley M Gifford; Jiri Navratil; Chandra Reddy; Ayhan Sebin Published: 2026-09-09T15:36:24Z Content type: release Language: en Sources: [Hugging Face - Blog]() Topics: [foundation-models](), [Temporal data](), [releases](), [benchmarking](), [Open Source Models & Datasets](), [Inference]() Tags: [apache](), [benchmarking](), [foundation-models](), [ibm](), [inference](), [model-architecture](), [open-source](), [releases](), [time-series](), [zero-shot]() ### AI overview IBM released Granite Time Series PatchTST-FM-r2, a roughly 385M-parameter time-series foundation model for zero-shot forecasting. The article covers its architecture, probabilistic forecasting, missing-value imputation, benchmark results, licensing, and available reproducibility resources. ### Source excerpt Time-series foundation models are changing the way forecasting systems are built. Instead of training and maintaining a separate model for every dataset, users can use a pretrained model and generate forecasts zero-shot. IBM has released Granite Time Series PatchTST-FM-r2, the latest model in the Granite TSFM family (github, blog). ## K2 Horizon just shipped as six new fully open models -- developers aren't fully convinced DevFeed: [K2 Horizon just shipped as six new fully open models -- developers aren't fully convinced]() Original publisher: [Read original article]() Author: Adrian Bridgwater Published: 2026-09-09T12:00:00Z Content type: news Language: en Sources: [The New Stack]() Topics: [Open Source Models & Datasets](), [foundation-models](), [Training AI Models]() Tags: [ai](), [ai-models](), [developer-tools](), [foundation-models](), [open](), [open-source](), [release](), [training]() ### AI overview The Institute of Foundation Models introduced K2 Horizon, a six-model open-source foundation-model fleet. The article examines its promised release of training artifacts and notes that some data, code, checkpoints, and a final 32B model were not yet available at launch. ### Source excerpt Based in the Emirati capital, Abu Dhabi, the Institute of Foundation Models (IFM) introduced K2 Horizon last week. This group The post K2 Horizon just shipped as six new fully open models -- developers aren't fully convinced appeared first on The New Stack. ## From voice command to robotic arm: how agentic AI on the edge is changing the factory floor DevFeed: [From voice command to robotic arm: how agentic AI on the edge is changing the factory floor]() Original publisher: [Read original article]() Author: Arduino Team Published: 2026-09-01T12:20:24Z Content type: article Language: en Sources: [Arduino Blog]() Topics: [Artificial Intelligence](), [foundation-models](), [Automation](), [Arduino](), [UNO Q](), [AI Agent](), [Physical AI](), [multimodal]() Tags: [agentic-ai](), [ai](), [arduino](), [automation](), [featured](), [foundation-models](), [industrial](), [inference](), [latency](), [manufacturing](), [multimodal](), [robotic-arm](), [robotics](), [smart-factory](), [uno-q](), [usb](), [voice-commands](), [voice-control]() ### AI overview The article describes a demonstration in which Forgis uses a foundation model running on an Arduino UNO Q board to convert voice commands into robotic-arm actions. The system processes multimodal factory data and performs inference locally, enabling real-time control without a cloud round trip. ### Source excerpt For years, bringing real intelligence to industrial automation meant expensive infrastructure, proprietary systems, and steep learning curves. That's changing - fast. Foundation models powerful enough to run at the edge are turning natural language into machine control, and the factory floor is starting to look a lot more like a conversation. AI as the new [...] The post From voice command to robotic arm: how agentic AI on the edge is changing the factory floor appeared first on Arduino Blog. ## TimesFM-3: A zero-shot foundation model for multivariate forecasting DevFeed: [TimesFM-3: A zero-shot foundation model for multivariate forecasting]() Original publisher: [Read original article]() Published: 2026-08-31T17:19:40Z Content type: article Language: en Sources: [The latest research from Google]() Topics: [foundation-models](), [Time Series](), [Google](), [datasets](), [Transformer architecture](), [Fine-tuning](), [generalization in machine learning]() Tags: [data-management](), [features](), [fine-tuning](), [foundation-models](), [google](), [machine-intelligence](), [model](), [product](), [time-series](), [transformer-architecture](), [zero-shot]() ### AI overview Google Research introduces TimesFM-3, a 330-million-parameter time-series foundation model designed for accurate multivariate forecasting in a single forward pass. Pre-trained on more than one trillion real-world and synthetic time points, it jointly models coevolving series and external covariates in zero-shot settings without task-specific fine-tuning. ### Source excerpt Data Management ## GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models DevFeed: [GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models]() Original publisher: [Read original article]() Author: Naoto Usuyama, Jeya Maria Jose Valanarasu, Tristan Naumann Published: 2026-08-31T16:00:00Z Content type: article Language: en Sources: [Microsoft Research]() Topics: [foundation-models](), [datasets](), [data]() Tags: [analysis](), [data](), [datasets](), [foundation-models](), [models](), [performance](), [research](), [research-blog](), [scale](), [validation]() ### AI overview GigaPath-Flash and GigaTIME-Flash are efficient pathology foundation models designed to reduce computational demands while maintaining strong performance. They enable repeated analysis of larger cancer cohorts and support population-scale research into disease biology, biomarkers, and clinical outcomes. ### Source excerpt What if pathology foundation models could do more with less? GigaPath-Flash and GigaTIME-Flash cut computational demands while maintaining strong performance, opening the door to larger studies and broader exploration. The post GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models appeared first on Microsoft Research. ## GlucoFM: Foundation model for continuous glucose monitoring DevFeed: [GlucoFM: Foundation model for continuous glucose monitoring]() Original publisher: [Read original article]() Published: 2026-08-26T18:42:43Z Content type: article Language: en Sources: [The latest research from Google]() Topics: [foundation-models](), [Google](), [data](), [datasets]() Tags: [data](), [foundation-models](), [google](), [health-bioscience](), [machine-intelligence](), [model](), [research]() ### AI overview GlucoFM is a lightweight, self-supervised foundation model for continuous glucose monitoring. Its dual-stream design separates slower glycemic trends from short-term deviations while preserving time-of-day and missingness information. Evaluated across four cohorts and seven clinical prediction tasks, it achieved higher average PR-AUC than the evaluated GluFormer variant. ### Source excerpt Health & Bioscience ## Post-Train NVIDIA Cosmos 3 Edge for On-Device Robot Control DevFeed: [Post-Train NVIDIA Cosmos 3 Edge for On-Device Robot Control]() Original publisher: [Read original article]() Author: Michelle Horton Published: 2026-08-19T16:00:00Z Content type: tutorial Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Cosmos](), [AI, ML & Data Engineering]() Tags: [agentic-ai-generative-ai](), [cosmos](), [deployment](), [edge](), [edge-computing](), [featured](), [foundation-models](), [inference](), [jetson](), [latency](), [nvidia](), [on-device](), [physical-ai](), [post-training](), [robotics](), [robotics-simulation](), [robots](), [simulation-modeling-design](), [thor](), [tutorial]() ### AI overview A tutorial on post-training NVIDIA Cosmos 3 Edge as an on-device robot manipulation policy, serving it on Jetson Thor, running receding-horizon inference, and evaluating it in closed-loop simulation. ### Source excerpt Robots need policies that can adapt to their sensors, environments, and tasks while running on onboard computing hardware. World models offer a foundation for... ## Introducing OlmoEarth embeddings: Custom embedding exports from OlmoEarth Studio for downstream analysis DevFeed: [Introducing OlmoEarth embeddings: Custom embedding exports from OlmoEarth Studio for downstream analysis]() Original publisher: [Read original article]() Author: Kyle Wiggers Published: 2026-08-12T16:14:36Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [Embeddings](), [foundation-models](), [Open Source](), [API](), [Fine-tuning](), [LLM evaluation / benchmarking]() Tags: [api](), [benchmarking](), [data](), [embeddings](), [fine-tuning](), [foundation-models](), [models](), [open-source]() ### AI overview OlmoEarth Studio now supports computing and exporting embedding vectors from open source OlmoEarth foundation models. Users can configure geographic area, time range, encoder, resolution, and imagery sources through the Studio UI or API, then download the results as Cloud-Optimized GeoTIFFs for downstream analysis such as similarity search, segmentation, and exploration. ### Source excerpt Introducing OlmoEarth embeddings: Custom embedding exports from OlmoEarth Studio for downstream analysis OlmoEarth Studio, our platform for building Earth observation models, now lets you compute and export embedding vectors--compact numerical representations of Earth-observation data produced by our open source OlmoEarth foundation models. ## IBM Releases GENCO and the GridFM Development Framework for Electric Grid Analysis DevFeed: [IBM Releases GENCO and the GridFM Development Framework for Electric Grid Analysis]() Original publisher: [Read original article]() Author: Peter Hess Published: 2026-08-11T13:00:40Z Content type: release Language: en Sources: [IBM Research]() Topics: [Artificial Intelligence](), [Algorithms](), [foundation-models](), [Open Source](), [Algorithms, Complexity](), [LLM evaluation / benchmarking](), [linux foundation]() Tags: [ai](), [algorithms](), [benchmarking](), [climate-and-sustainability](), [foundation-models](), [linux-foundation](), [open-source](), [release](), [research]() ### AI overview IBM Research and collaborators announce GENCO, an open-source neural solver for three steady-state electric-grid analysis tasks, alongside the GridFM Development Framework for building and benchmarking neural grid solvers. ### Source excerpt GENCO is a neural solver that, alongside the GridFM Development Framework, unifies three core electrical grid analysis tasks. ## Forking-Sequences -- Part II: Multi-Horizon Forecast Ensembling with Reduced Volatility DevFeed: [Forking-Sequences -- Part II: Multi-Horizon Forecast Ensembling with Reduced Volatility]() Original publisher: [Read original article]() Author: Willa Potosnak Published: 2026-08-10T21:52:23Z Content type: article Language: en Sources: [ML@CMU]() Topics: [Sequences](), [Time Series](), [foundation-models](), [Inference](), [Machine Learning & Artificial Intelligence]() Tags: [ai](), [artificial-intelligence](), [automated-machine-learning](), [big-data](), [computer-science](), [data-science](), [deep-learning](), [ethics-and-ai](), [forecasting](), [foundation-models](), [inference](), [machine-learning](), [ml](), [multi-horizon-forecasting](), [natural-language-processing](), [neural-networks](), [paper](), [research](), [sequences](), [time-series]() ### AI overview The article describes forecast ensembling for multi-horizon time-series models. Exponential smoothing reduces forecast volatility by about 10-13% with less than 0.1% accuracy degradation, including in zero-shot use with pretrained time-series foundation models. ### Source excerpt Based on: Potosnak, W., Wolff, M., Cao, M., Ma, R., Konstantinova, T., Efimov, D., Mahoney, M.W., Oreshkin, B., & Olivares, K.G. "Forking-Sequences: Statistically and Computationally Efficient Multi-Horizon Forecasting with Reduced Volatility." Transactions on Machine Learning Research, 2026. (Disclaimer: Code implementation not used in the paper; not affiliated with Amazon -- provided as a reference for forking-sequences and forecast ensembling) TL;DR Ensembling, nearly for free. Forking-sequences already produces overlapping forecasts for every target date across FCDs in a single forward pass, so ensembling them at inference adds no extra encoder computation compared with window-sampling. Two new forecast volatility metrics. scaled Forecast Percentage Change (sFPC) measures raw revision size in real time (no ground truth needed); Excess Volatility (EV) goes further, rewarding accuracy-improving revisions and only penalizing the ones that move forecasts away from the truth or overshoot it. Reduced volatility without sacrificing accuracy. Exponential-smoothing forecast ensembling (α = 0.9) reduces sEV by 10-13% across all encoder types, with less than 0.1% accuracy degradation. Works zero-shot on models pretrained with window-sampling. Forecast ensembling applied to pretrained Time Series Foundation Models (TSFMs) -- Chronos-2, Toto 2.0, TimesFM, PatchTST, N-BEATS -- cuts volatility by ~10% with negligible accuracy cost (less than 0.1%). In Part [...] ## How Physical Intelligence unified its robotics data stack with Postgres managed by ClickHouse DevFeed: [How Physical Intelligence unified its robotics data stack with Postgres managed by ClickHouse]() Original publisher: [Read original article]() Author: ClickHouse Published: 2026-08-06T00:00:00Z Content type: article Language: en Sources: [ClickHouse Blog]() Topics: [clickhouse](), [Robotics](), [foundation-models](), [data](), [Artificial Intelligence](), [Cloud]() Tags: [ai](), [clickhouse](), [data](), [foundation-models](), [model](), [research](), [robotics](), [robots]() ### AI overview Physical Intelligence describes how it uses ClickHouse Cloud and ClickHouse-managed Postgres to support robotics foundation-model research. The unified data stack combines analytical and transactional workloads, helping the company explore datasets that have grown to roughly 10-100 billion rows. ### Source excerpt Physical Intelligence runs both its OLAP and OLTP workloads on ClickHouse managed Postgres and ClickHouse Cloud ## A new benchmark for evaluating patient-facing health AI agents DevFeed: [A new benchmark for evaluating patient-facing health AI agents]() Original publisher: [Read original article]() Author: Korosh Vatanparvar; Ashutosh Joshi Published: 2026-07-29T15:16:52Z Content type: article Language: en Sources: [Amazon Science homepage]() Topics: [AI Bots]() Tags: [agentic-ai](), [agents](), [ai](), [benchmark](), [conversational-ai](), [evaluation](), [foundation-models](), [health](), [healthcare](), [safety](), [synthetic]() ### AI overview The article introduces PatientAgentBench, a clinician-vetted benchmark for evaluating the safety and task performance of patient-facing healthcare AI agents in realistic, multiturn conversations. ### Source excerpt PatientAgentBench generates a synthetic patient health record, a realistic clinical vignette, and a patient agent that converses with the AI system under evaluation, to capture what a patient-facing agent actually has to do. ## NVIDIA Cosmos-H-Dreams: Bringing Real-Time Generative Simulation to Surgical Robotics DevFeed: [NVIDIA Cosmos-H-Dreams: Bringing Real-Time Generative Simulation to Surgical Robotics]() Original publisher: [Read original article]() Author: Lukas Zbinden; Javier Gamazo; Mostafa Toloui; Sean Huver Published: 2026-07-27T09:32:20Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [Cosmos](), [Robotics](), [foundation-models](), [real-time](), [Nvidia](), [Synthetic Data Generation](), [synthetic-data](), [data](), [datasets](), [GPU](), [Streaming](), [Inference]() Tags: [cosmos](), [data](), [foundation-models](), [generation](), [generative](), [gpu](), [inference](), [nvidia](), [real-time](), [robotics](), [streaming](), [synthetic-data](), [synthetic-data-generation](), [training]() ### AI overview NVIDIA introduces Cosmos-H-Dreams, a real-time, action-conditioned generative simulator for surgical robotics. The model generates future surgical video from an initial RGB frame and live robot kinematics, enabling interactive closed-loop control, offline policy evaluation, and synthetic data generation. ### Source excerpt World foundation models offer a different path. Instead of manually authoring every object and physical interaction, they learn visual dynamics directly from synchronized video and robot kinematics. NVIDIA's Cosmos-H-Surgical-Simulator demonstrated this approach by generating future surgical video from an initial scene and a sequence of robot actions. It enabled faster-than-physical evaluation and synthetic data generation across the Open-H-Embodiment ecosystem. ## Launching Health in ChatGPT DevFeed: [Launching Health in ChatGPT]() Original publisher: [Read original article]() Published: 2026-07-23T00:00:00Z Content type: release Language: en Sources: [OpenAI News]() Topics: [ChatGPT](), [iOS](), [Security](), [Web](), [foundation-models]() Tags: [apple](), [apple-health](), [apps](), [chatgpt](), [foundation-models](), [health](), [ios](), [privacy](), [product](), [security](), [web]() ### AI overview Health in ChatGPT is launching to eligible users in the United States, allowing them to securely connect Apple Health and supported medical records for more personalized health conversations. The feature is available to logged-in users aged 18 and older on web and iOS, with layered privacy and security safeguards. ### Source excerpt Health in ChatGPT now lets eligible U.S. users securely connect medical records and Apple Health to get more personalized insights and better understand their health. ## SceneSmith uses collaborative AI agents to create 3D environments for robot training DevFeed: [SceneSmith uses collaborative AI agents to create 3D environments for robot training]() Original publisher: [Read original article]() Author: Alex Shipps | MIT CSAIL Published: 2026-07-13T18:50:00Z Content type: news Language: en Sources: [MIT AI News]() Topics: [AI research agents](), [robot grasping simulation](), [Simulation](), [vlm](), [Artificial Intelligence](), [Computer Science and Artificial Intelligence Laboratory (CSAIL)]() Tags: [3-d](), [3d](), [adversarial-machine-learning](), [agents](), [ai-agents](), [artificial-intelligence](), [computer-science-and-artificial-intelligence-laboratory-csail](), [computer-science-and-technology](), [electrical-engineering-and-computer-science-eecs](), [foundation-models](), [general-purpose-robotics](), [gpt-5-2](), [machine-learning](), [mit-csail](), [mit-eecs](), [mit-schwarzman-college-of-computing](), [national-science-foundation-nsf](), [nicholas-pfaff](), [research](), [robot-simulations](), [robotics](), [robots](), [russ-tedrake](), [scene-generation](), [scenesmith](), [school-of-engineering](), [simulation](), [simulation-ready-indoor-scenes](), [virtual-playgrounds](), [vision-language-models-vlms](), [zero-shot-policy]() ### AI overview MIT CSAIL and Toyota Research Institute researchers developed SceneSmith, a system that uses three collaborative AI agents to create realistic 3D environments for robot training. The scenes can be loaded into physics simulation software, allowing robots to practice tasks before real-world testing. ### Source excerpt "SceneSmith" system uses collaborative AI agents to create realistic 3D environments of places like kitchens, hotels, and living rooms, where robots can simulate everyday chores. ## How to Evaluate General-Purpose Robot Policies for Real-World Deployment DevFeed: [How to Evaluate General-Purpose Robot Policies for Real-World Deployment]() Original publisher: [Read original article]() Author: Brad Nemire Published: 2026-07-12T01:08:17Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Robotics](), [Benchmark](), [foundation-models](), [Simulation and Design](), [Deployment]() Tags: [benchmark](), [benchmarks](), [deployment](), [evaluation](), [featured](), [foundation-models](), [nvidia-research](), [physics](), [robotics](), [simulation-modeling-design](), [testing]() ### AI overview This developer article examines the challenge of rigorously evaluating general-purpose robot policies for real-world deployment. It discusses simulation as a scalable proxy for expensive real-world testing and identifies limitations in current benchmarks, including shared visual sources between training and evaluation, costly Real2sim reconstruction, static task sets, performance saturation, limited failure diagnostics, and uncertainty in success-rate estimates. ### Source excerpt Robotics foundation models have made remarkable progress. Today's best systems can follow natural language instructions to pick, place, sort, and manipulate a... ## Aurora 1.5: Extending open foundation models for weather and Earth-system applications DevFeed: [Aurora 1.5: Extending open foundation models for weather and Earth-system applications]() Original publisher: [Read original article]() Author: Kenji Takeda, Haiyu Dong, Jonathan Weyn, Amit Misra, Matt Corey, Kevin White, Shannon Monroe, Juan M. Lavista Ferres, Ashley Llorens, Bonnie Kruft Published: 2026-07-09T16:46:22Z Content type: article Language: en Sources: [Microsoft Research]() Topics: [foundation-models](), [Open Source](), [Artificial Intelligence](), [Microsoft](), [GitHub](), [hugging face](), [Fine-tuning]() Tags: [ai](), [fine-tuning](), [foundation-models](), [github](), [hugging-face](), [microsoft](), [models](), [open](), [open-source](), [research](), [research-blog](), [update]() ### AI overview Aurora 1.5 extends Microsoft's open Aurora Earth-system foundation model with 22 additional weather variables, hourly temporal resolution, and probabilistic ensemble forecasting. Released through GitHub with checkpoints on Hugging Face, it is intended for researchers and developers working on weather, climate, energy, agriculture, transport, and related applications. ### Source excerpt Aurora 1.5 adds 22 more variables, hourly temporal resolution, and probabilistic ensemble forecasting to the Aurora foundation model, making it more useful for real-world weather, climate, and energy applications. The post Aurora 1.5: Extending open foundation models for weather and Earth-system applications appeared first on Microsoft Research. ## Why intent prediction needs more than an LLM DevFeed: [Why intent prediction needs more than an LLM]() Original publisher: [Read original article]() Author: Phoebe Sajor Published: 2026-06-30T07:40:00Z Content type: article Language: en Sources: [Stack Overflow Blog]() Topics: [foundation-models](), [personalization](), [Graphs](), [Large Language Model](), [data](), [Deep neural networks]() Tags: [agentic-ai](), [ai](), [data](), [foundation-models](), [llm](), [llms](), [math](), [personalization](), [podcast](), [se-stackoverflow](), [se-tech](), [transformers]() ### AI overview Ryan Donovan interviews Frank Portman, CTO at Yobi, about why large language models are not the right inductive bias for predicting human intent and behavior. The discussion covers Yobi's behavioral foundation models, which use transformers and graph neural networks to support large-scale personalization while keeping consumer data private. ### Source excerpt Ryan sits down with Frank Portman, CTO at Yobi, to talk about why next-token prediction, though great for language, isn't the right inductive bias for forecasting human behavior. They discuss how Yobi builds a "foundation model of behavior" using transformers and graph neural networks instead of chat-style LLMs, and what it takes to run millions of personalization decisions per second while keeping consumer data private. ## Achieving Near-Linear Training Scalability for Pinterest's Foundation Models DevFeed: [Achieving Near-Linear Training Scalability for Pinterest's Foundation Models]() Original publisher: [Read original article]() Author: Pinterest Engineering Published: 2026-06-25T16:01:02Z Content type: article Language: en Sources: [Pinterest Engineering Blog - Medium]() Topics: [Training AI Models](), [NCCL](), [networking]() Tags: [aws](), [distributed-training](), [embedding](), [engineering](), [foundation-models](), [gpu](), [machine-learning](), [nccl](), [pinterest](), [scalability](), [training]() ### AI overview Pinterest describes optimizing multi-node distributed training for an embedding-heavy foundation ranking model. The reported improvements raise scaling from 1.13x to 2.0x at two nodes and from 1.21x to 3.9x at four nodes, with 7.5x scaling at eight nodes. ### Source excerpt Sheng Huang | Software Engineer, AI Platform; Pong Eksombatchai | Machine Learning Engineer, Applied Sciences; Saurabh Vishwas Joshi | Software Engineer, AI Platform; Gaurav Arora | Software Engineer, AI Platform; Karthik Anantha Padmanabhan | Engineering Director, AI Platform At Pinterest, foundation models power recommendations for over 600 million monthly active users. Our latest Foundation Model (ACM RecSys 2025) pre-trains on two years of user activity data and is deployed into Home feed and Related Pins ranking, the platform's two most important recommendation systems. Multi-node distributed training is the key to unlocking the next level of that capacity.¹ But when we first attempted multi-node training, adding a second machine made training 5x slower, producing a scaling factor of roughly 0.2x. Enabling AWS Elastic Fabric Adapter (EFA) for OS-bypass networking fixed the networking layer and recovered a viable baseline, but scaling was still poor: 1.13x at 2 nodes and 1.21x at 4 nodes. Three extra nodes, 3x more GPUs, 3x more cost, yet only 21% more throughput. This post describes how we took 2-node scaling from 1.13x to 2.0x and 4-node scaling from 1.21x to 3.9x (97.5% of ideal), then extended to 8 nodes at 7.5x. The larger models this unlocked have driven significant engagement gains across Pinterest's recommendation surfaces. Figure 1: Training scalability before and after optimization. Left: before EFA and optimization, adding a second node degraded throughput to 0.2x of single-node. Right: after optimization, scaling is near-linear across 2, 4, and 8 nodes, with 8-node reaching 7.5x (93.75% of ideal).Background Training scalability measures whether adding more resources yields proportionally more throughput. Training efficiency measures how much throughput you extract from the same resources. This post focuses on scalability. Our Foundation Ranking Model is embedding-heavy: approximately 99% of parameters reside in embedding tables, with the dense transf [Next page]()