Researchers from Northeastern University, MIT, and an independent researcher developed Concept Sliders for text-to-image diffusion models, allowing fine-grained image control and editing. This method enables manipulation of visual concepts that are usually hard to describe in words and offers a practical, disentangling solution for more precise image customization through open-source code and trained sliders.

Artists seeking copyright infringement claims against Stability AI and others have refiled their lawsuit with seven additional plaintiffs. The original case was dismissed, but Judge William Orrick allowed for an amended resubmission. The updated lawsuit uses comments by Stability AI’s CEO and concerns over derivative works and AI’s use of copyrighted data to bolster its…

PGXMAN is a package manager for Postgres extensions, streamlining installation, update, and management processes. It handles dependencies automatically, saving developers time and effort. Installation is easy via pip, and a supportive community further enhances its utility. For more information, visit https://pgxman.com/.

Researchers from Microsoft and Georgia Tech developed TongueTap, a wearable tech interface that uses tongue gestures to control devices without hands or eyes. It combines data from IMUs and PPG sensors in headsets for gesture recognition with 80-94% accuracy, promising improvements for AR interactions.

RAGs, an application by Streamlit, simplifies GPT pipeline creation and deployment with an intuitive interface. The latest version, RAGs v2, enhances user experience with features for building and customizing ChatGPTs, managing RAG pipelines, and supporting multiple large language models. To use it, install with ‘pip,’ create pipelines, deploy, and query via command line. It’s a…

The study by Shanghai Jiao Tong University, Amazon, and Yale explores Chain-of-Thought reasoning in language models, examining its impact on the development and reliability of language agents. It investigates CoT techniques and verification methods, offering insights for both new and seasoned researchers in language intelligence.

UC Berkeley researchers have developed ALIA, an innovative language-guided image augmentation technique that improves dataset variety and classification model performance in fine-grained image tasks without extensive fine-tuning. It uses natural language to generate domain-specific image edits and employs filtering to maintain visual consistency, showing a significant enhancement over traditional methods in experiments.

Porto Alegre’s council passed a law written entirely by ChatGPT on stolen water meter charges, unveiled by Councilman Ramiro Rosário after unanimous approval. His nondisclosure aimed to provoke AI usage debates in legislation, amidst similar AI legislative efforts globally, stirring discussions on transparency and AI’s future role in governance.

Amazon’s AI chatbot, Amazon Q, has allegedly leaked sensitive internal information including AWS data centers and unreleased features. While Amazon denies security breaches, internal Slack communications show employee concerns. This leak is unconfirmed but follows past secrecy about Amazon data center locations and echoes previous AI errors like Google’s Bard.

The Johns Hopkins University team developed an algorithm for matching celestial bodies across different sky surveys. The program accurately compares massive datasets, considering position, brightness, and color, to identify identical astronomical objects, improving data integration for space research.