Interactive 3D timeline reveals vision language models sense of time
TimeLineXR: An Interactive Exploration of Time in Vision-Language Models
Computer Vision and Pattern Recognition
Summary
Some vision-language AI models can figure out when objects first appeared in history by understanding time in their internal math spaces. The authors made TimeLineXR, a web tool that shows these insights visually in 3D. Users can upload images and see where they fit on a color-coded timeline representing years. They can also interact with this timeline in virtual reality, exploring how different images relate to time. This helps people better understand the AI’s sense of time by letting them actually see and play with it.
What this means in practice
- •For interactive exhibit designers: Create immersive museum or educational installations where visitors can explore historical timelines visualized through AI-embedded image analysis.
- •For virtual reality developers: Build VR experiences that visualize and let users manipulate AI-derived time relationships of images for educational or storytelling purposes.
Authors
Nidham Tekaya, Markus Passecker, Matthias Zeppelzauer
Abstract
Recently, it has been shown that vision-language models (VLMs) have a certain intrinsic understanding of time. They can estimate the temporal origin of objects and encode time along a low-dimensional, non-linear manifold inside their embedding space~\cite{tekaya2025matter}. That result is quantitative and reproducible, but the structure itself stays out of reach for end users: they cannot see it, query it with their own images, or inspect where a query lands relative to other years on the time axis. We present TimeLineXR, an interactive demonstration that turns this time axis into an explorable 3D scene. A web demo lets users upload or select images and see each placed on a color-coded 3D time axis; a Three.js WebXR scene renders the same time axis as a room-scale representation that can be grabbed, scaled, inspected, and queried by throwing image cards into it. Both clients run the same prediction pipeline against the same cached temporal index: encode the query with CLIP, project it through cosine-kernel KPCA, and snap it to the Bézier time axis. The system turns each image into a point on the time axis, so users can compare where different objects land and read off their estimated first year of appearance.Code source and reproduction steps can be found in: https://github.com/sterngefeuert/timeline-xr.