Can you visualize what NYC smells like? Yes, turns out, you can (huggingface.co)

Researchers have introduced the "New York Smells" dataset, a groundbreaking multimodal collection that pairs visual imagery with olfactory data captured in various environments across New York City. Comprising 7,000 smell-image pairs from 3,500 distinct objects, the dataset significantly expands the scope of olfactory research, offering approximately 70 times more objects than current datasets. It utilizes an electronic nose (e-nose) equipped with 32 sensors, collecting data in real-world conditions rather than controlled lab settings. This innovative approach addresses a vital gap in machine perception, enabling cross-modal learning between smell and sight. The significance of this dataset for the AI/ML community lies in its ability to facilitate advancements in olfactory representation learning, scene recognition from smell, and environmental classification. By allowing models to learn smell embeddings with visual supervision, researchers can develop systems that mimic animal-like olfactory perception, potentially enhancing applications in environmental monitoring, health diagnostics, and beyond. As olfaction remains largely uncharted in machine learning, the New York Smells dataset paves the way for future research and applications that integrate the sense of smell into artificial intelligence systems, creating new possibilities for multimodal sensor fusion and human-computer interaction.