Multimodal Artificial Intelligence Assistant Developed for Cancer Pathology
A new multimodal artificial intelligence assistant called SlideChat has been introduced to support whole-slide computational pathology processes across various cancer types.
Developed for use in the detection and examination of various cancer types, the multimodal artificial intelligence assistant named SlideChat brings an innovative approach to whole-slide computational pathology processes.
Artificial Intelligence-Powered Pathology Assistant
SlideChat has been introduced to the scientific world as a multimodal and generative artificial intelligence assistant designed to perform whole-slide analyses in the field of cancer pathology.
Datasets and Resources
Various datasets such as TCGA, CPTAC, BCNB, and HISTAI were utilized within the scope of the research; these data were shared with the public through the NIH Genomic Data Commons, Proteomic Data Commons, and Hugging Face platforms.
Open Source Code and Model Weights
While the source codes belonging to the training and inference processes of the model were made accessible on GitHub, the model weights were made available to researchers via the Hugging Face platform.
Institutional Collaborations and Support
The study was conducted with contributions from institutions such as the Shanghai Artificial Intelligence Laboratory, the Shanghai Institute for Innovation, Xiamen University, and Shanghai General Hospital. The project was also supported by the National Natural Science Foundation of China and the Shanghai Municipal Commission of Economy and Informatization.