United States Patent and Trademark Office
PulseAugur coverage of United States Patent and Trademark Office — every cluster mentioning United States Patent and Trademark Office across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
AI-driven methods increasingly applied to patent analysis and curation
The recent clusters highlight the growing application of AI in diverse areas of patent work, including subset selection for trademark curation, predicting litigation risk using GATs, and enhancing chemical structure recognition. This indicates a trend towards AI-powered tools becoming integral to patent office operations and legal analysis.
USPTO to explore AI inventor frameworks within 1 year
Given the recent discussion on US patent law struggling to credit AI inventors, it is plausible that the USPTO will begin exploring frameworks or pilot programs for AI-invented patents within the next year. This is driven by the need to incentivize innovation in AI-driven fields like drug discovery.
USPTO to adopt new diversity metrics for trademark curation within 18 months
The development of 'fair multi-view determinant selection' for diverse subset selection, specifically mentioning trademark curation, suggests that the USPTO may investigate or pilot such advanced methods. Adoption of new diversity metrics could be expected within 18 months to improve the efficiency and fairness of trademark curation processes.
-
LLMs revolutionize organic synthesis with automated platforms and reaction prediction
Two new research papers highlight the growing impact of large language models (LLMs) on organic synthesis. The first paper, published on arXiv, details how LLMs are transforming workflows from reaction prediction to aut…
-
New BLANC method identifies patent white space using multi-view clustering
Researchers have developed BLANC, a novel pipeline for identifying unexplored areas in patent landscapes, crucial for strategic R&D. This method employs multi-view neural topic modeling across semantic dimensions like a…
-
Graph Attention Network predicts patent litigation risk
Researchers have developed a Graph Attention Network (GAT) called ClaimGAT to predict patent litigation risk. This model addresses limitations in previous BERT-based approaches by encoding each patent claim independentl…
-
VERDICT system enhances chemical structure recognition accuracy
Researchers have developed a new system called VERDICT that improves the accuracy of Optical Chemical Structure Recognition (OCSR). VERDICT utilizes agreement among multiple recognizers, outperforming pixel-space verifi…
-
US patent law struggles to credit AI inventors in drug discovery
Intellectual property law currently faces a significant challenge as artificial intelligence systems are increasingly capable of generating novel drug designs. Despite AI's role in these discoveries, current patent laws…
-
New method tackles diverse subset selection for trademark curation
Researchers have developed a new method called "fair multi-view determinant selection" to address the challenge of selecting diverse subsets from large candidate pools, particularly in applications like trademark curati…
-
Real-world data crucial for AI chemical structure recognition accuracy
A new research paper explores the challenge of optical chemical structure recognition (OCSR) in real-world documents, highlighting a significant gap between performance on synthetic data and actual patent and journal fi…
-
Mistral AI granted patent for AI tool call implementation
Mistral AI has been granted a patent for a method of implementing tool calls in code, specifically related to AI models. The patent, identified as US12670045, was published by the United States Patent and Trademark Offi…
-
Tennis AI Startup Files Provisional Patent Before Prototype Development
The founder of PureStrokeAI discusses the strategic decision to file a provisional patent before developing a prototype for their tennis analytics technology. This approach was chosen to establish an early filing date i…
-
US Patent System Effectively Protects AI Innovation, Data Shows
The US patent system, particularly Section 101, is effectively safeguarding innovation while preventing monopolies on fundamental concepts. High affirmance rates by the Federal Circuit and unanimous judicial decisions i…
-
Patent database scales to 5.36M records with SQLite/FTS5 optimizations
A patent lawyer and hobby coder shared insights on scaling a patent database from 3.5 million to 5.36 million records using SQLite and FTS5. Key learnings include the importance of running ANALYZE after bulk loading, th…
-
LLMs and graph analysis detect emerging tech convergence
Researchers have developed a new data-driven pipeline to forecast transformative technologies by analyzing patterns of technological convergence. This method utilizes Large Language Models (LLMs) to extract semantic tri…
-
New framework links medical device approvals to patents
Researchers have developed a new framework called Bridge-MedDevKG to link FDA-approved cardiovascular devices with their corresponding patents. This task is challenging due to significant semantic differences between cl…
-
US Patent Office races to define ownership of AI inventions
The United States Patent and Trademark Office is grappling with the ownership of inventions created by agentic artificial intelligence. These AI systems can generate patentable content at a pace that challenges the exis…
-
New AI Framework Simplifies Patent Valuation
Researchers have developed PatentXAI, a novel framework designed to tackle the complex problem of valuing individual patents within large product portfolios. This framework leverages graph-conditioned hierarchical Shapl…
-
DLT-Corpus released: 2.98B tokens for Distributed Ledger Technology NLP
Researchers have introduced DLT-Corpus, a substantial text collection designed for Distributed Ledger Technology (DLT) research, comprising 2.98 billion tokens from over 22 million documents. This corpus includes scient…
-
AI framework cuts patent claim validation costs by 78%
Researchers have developed a new framework called Adaptive Cost-Efficient Evaluation (ACE) to improve the accuracy and reduce the cost of validating patent claims using AI. ACE uses a two-stage process: first, a fine-tu…
-
Author seeks new name amid trademark concerns
The author is facing challenges in finding a unique and legally safe name for their website and music persona. They previously used 'Dawghouse' for their site, but discovered it infringes on existing trademarks, leading…
-
Developer launches unified tool for company intelligence in AI agents
A developer has created a unified tool called Company Intelligence MCP to streamline research on companies for AI agents. This tool consolidates data from various sources like SEC filings, patent databases, and domain i…
-
SpaceXAI files trademark, signaling official launch
SpaceXAI has officially filed for a trademark with the USPTO, indicating a formal launch or public announcement is imminent. The trademark application is currently under review by the United States Patent and Trademark …