Latest AI patents, Page 213
How the biggest tech companies are building AI: language models, assistants and agents, vision, voice, and the systems that train them. The artificial intelligence and machine learning patent applications we have broken down from the companies we track, newest first.
- in the Nov 12 release
- 96
- in November so far
- 188
- filings tracked
- 2,637
- sub-topics
- 10
Recent filings
Newest first · 2,637 total-
Microsoft · Published May 14
Microsoft Patents a Dense Context Engine That Primes AI With Company-Specific Knowledge
Before your AI assistant answers a question, what if it could automatically load up a compressed briefing of everything relevant about your company and industry? That's exactly what Microsoft is patenting here.
-
Microsoft · Published May 14
Microsoft Patents a Neural Prior Model for Photorealistic 3D Avatars
Microsoft is filing patents on a system that can generate a photorealistic, animatable 3D avatar of you from a single photo — no studio, no body scan, no lengthy setup required. The trick is a neural network trained on thousands of 3D head models that already 'knows' what human faces look like before it ever sees yours.
-
Microsoft · Published May 14
Microsoft Patents an AI System That Sorts Your Screenshots by What You Were Actually Doing
Microsoft is patenting a way to automatically slice a long stream of screenshots into discrete activity chunks — so instead of a wall of captures, you'd see 'email session,' then 'coding session,' then 'meeting prep' as clean segments. It's the organizational layer that could make AI-powered screen history actually usable.
-
Microsoft · Published May 14
Microsoft Patents an AI System That Scores URLs Before Indexing Them
Every second, search engines discover millions of new URLs — and most of them are junk. Microsoft's new patent describes an AI system that can score a URL's likely value before a crawler even bothers to visit it.
-
Microsoft · Published May 14
Microsoft Patents a System That Translates Classic ML Models Into Neural Networks
Most companies have years of traditional machine learning models — decision trees, gradient boosters, scikit-learn pipelines — that can't easily tap into modern deep learning tricks. Microsoft's patent describes a system that automatically converts those classic models into neural networks, without a full rewrite.
-
Microsoft · Published May 14
Microsoft Patents an AI Pipeline That Generates New Molecules From Chemical Neighbors
Instead of asking an AI to invent molecules from scratch, Microsoft's new patent describes a smarter approach: find the chemical neighbors of a known compound first, then fine-tune a generative model on that richer neighborhood — nudging the AI toward territory that's already proven interesting.
-
Microsoft · Published May 14
Microsoft Patents an AI System That Figures Out What's Missing From Your Photos
Most AI image tools wait for you to tell them what to add. Microsoft is patenting a system that looks at your image and decides what's missing — and where it should go — on its own.
-
Meta · Published May 14
Meta Patents a Web Platform for Building Custom AI Chatbots
Meta is filing patents for a drag-and-drop-style AI chatbot creation system that uses data already sitting on its platforms to train your bot. This looks a lot like the foundation for turning every business page — or personal profile — into an always-on AI assistant.
-
Nvidia · Published May 14
Nvidia Patents a Training Method for Turning Text Into 3D Objects
Nvidia is building a pipeline that takes a plain-text description — say, 'a flowing medieval tunic' — and spits out a fully simulation-ready 3D object. The interesting part isn't just the generation; it's that the output is designed to work inside physics simulators, not just look good in a render.
-
Nvidia · Published May 14
Nvidia Patents a Configurable Neural Network Document Transcription System
Nvidia is patenting a smarter way to turn document images into structured text — one where you can dial up exactly what kinds of annotations you want the AI to produce, rather than getting a one-size-fits-all output.
-
Nvidia · Published May 14
Nvidia Patents a Multi-Teacher Knowledge Distillation System Using LoRA Towers
Training a small AI model to match the performance of a large one is notoriously hard — especially when you want it to learn from multiple expert models at once. Nvidia's new patent describes a clever system that uses dynamically-sized "LoRA towers" to absorb knowledge from multiple teachers simultaneously, adjusting their own complexity on the fly.
-
Google · Published May 14
Google Patents a Diffusion-Based System for Labeling Every Pixel in Images and Video
Google is applying the same generative AI technology behind image synthesis tools to one of computer vision's hardest problems — identifying and tracking every single object in a scene, pixel by pixel, across entire videos.
