VISxAI: 9th Workshop on Visualization for AI Explainability
Organizers:Alex Bäuerle, Angie Boggust, Catherine Yeh, Fred Hohman
Description:
The VISxAI workshop, throughout its past eight iterations, has been a platform for knowledge exchange between researchers with different backgrounds interested in explaining machine learning models through visualization. Its focus on explainables submissions that visually and interactively explain machine learning concepts ranges in complexity from clustering methods to algorithmic biases. These explainables have served as educational resources that have had an impact beyond the academic community. The workshop also always hosted great keynote speakers that connected the domains of visualization and state-of-the-art machine learning and explored the impact visualization can have on explainability. Following the success of VISxAI’25, our goal for this upcoming iteration of VISxAI is to combine strong interactive explainables and great presentations with the interactivity of breakout sessions and live demos. Participants will be encouraged to exchange ideas about the future of visual explainability, interactive articles, and explorable explanations. Furthermore, we will provide a platform for new visualization and interaction ideas that explain machine learning models.
Jump to DiscussionPapers
arxAIv: An AI Generated Research Universe
Byte-Pair Encoding Explainer: An Interactive Visualization for Understanding Subword Tokenization
Embodied XAI Is All You Need: Tracing hierarchical VLM–PPO decisions via 4D spatio-temporal saliency
Exploring The Geographic Assumptions in Text-to-Image Models with Underspecified Prompts
How to Win the Lottery within Neural Networks
Interpreting Bio-Foundation Models with Sparse Autoencoders
emoji_eventsBest Explainable Award
MoEXViz: Interactive Visualizations of Sparse Mixture-of-Experts
More Features Than Neurons
NCDL Explainer: Non-Cartesian Deep Learning Explained
New Francisco or San York? A Visual Explanation of Text Diffusion
emoji_eventsBest Explainable Award
Rectified Flow Explainer: A Visual Introduction to Rectified Flows
Some Vector Algebra for Model Steering
emoji_eventsBest Explainable Award
Superposition Explainer: Visualizing Feature Overcompression Through Guided Experimentation
emoji_eventsBest Explainable Award
The Confident Machine
The Story of Inference-Time Scaling and the training-time loop behind it
TokenTour: Tracing LLM Conversations from Messages to Tokens and KV Cache
emoji_eventsBest Explainable Award