What happens when nanochat meets DiLoCo?

OpenAI has launched a free version of ChatGPT specifically designed for teachers. This initiative aims to provide educators with accessible AI tools to enhance their teaching methods and engage students more effectively. The move reflects OpenAI's commitment to supporting the education sector by leveraging advanced AI technology.

OpenAI has announced the launch of ChatGPT for Teachers, a specialized version of its AI chatbot aimed at K-12 educators and school districts. This service will be available for free to K-12 educators in the United States until June 2027, providing tools to enhance teaching and learning experiences.

OpenAI is expanding its presence in the retail sector as Target prepares to launch a new ChatGPT-powered app for shoppers. The app is set to enter beta testing next week, aiming to enhance the shopping experience through AI technology.

Target is partnering with OpenAI to enhance its sales strategy by integrating ChatGPT into its customer service. This collaboration will allow consumers to tag Target in ChatGPT and specify their shopping needs, aiming to improve user experience and drive sales.

MoHoBench is a newly developed benchmark aimed at assessing the honesty of Multimodal Large Language Models (MLLMs) when confronted with unanswerable visual questions. Despite advancements in vision-language tasks, MLLMs often produce unreliable content. This study systematically evaluates the honesty of 28 popular MLLMs using a dataset of over 12,000 visual questions, revealing that many models struggle to provide honest responses. The findings highlight the need for improved trustworthiness in AI systems.

MVI-Bench is introduced as a comprehensive benchmark aimed at evaluating the robustness of Large Vision-Language Models (LVLMs) against misleading visual inputs. Traditional benchmarks have primarily focused on textual inputs, neglecting the significant impact of visual misrepresentation. MVI-Bench categorizes misleading visual inputs into three hierarchical levels: Visual Concept, Visual Attribute, and Visual Relationship, and includes 1,248 annotated Visual Question Answering (VQA) instances to facilitate detailed robustness assessments.

The paper presents Gaussian Spatial Transport (GST), a new framework that utilizes Gaussian splatting to transfer probability measures from image coordinates to annotation maps. It introduces a method for estimating pixel-annotation correspondence, which is used to create a transport plan based on Bayesian probability. A loss function is derived to integrate this transport plan into standard network optimization for computer vision tasks. Experiments in crowd counting and landmark detection demonstrate the approach's effectiveness, improving efficiency by eliminating iterative transport plan c…

The article presents a novel training strategy called Curriculum-based Relative Policy Optimization (CuRPO) aimed at improving Visual Grounding tasks. It highlights the limitations of Chain-of-Thought (CoT) prompting, particularly when outputs become lengthy or complex, which can degrade performance. The study reveals that simply increasing dataset size does not guarantee better results due to varying complexities. CuRPO utilizes CoT length and generalized Intersection over Union (gIoU) rewards to structure training data progressively from simpler to more challenging examples, demonstrating ef…