SciForge Team, Zhangyang Gao, Minghao Fang +9 more
A single piece of research is scattered across papers, code, datasets, odd scientific file formats, model outputs, figures, drafts, and decisions the team made in conversation somewhere. General-purpose AI assistants can help with any one of those, but they rarely hold them together as a coherent state you could later audit.
SciForge is built around that gap. Search, parsing, model routing, workflow execution, plotting, writing and presentation generation all run as modular services an agent can call, while the graphical interface is deliberately reserved for human judgment. The split is the interesting design decision: the machine gets the fetching and assembling, the human keeps the deciding.
The word that carries the most weight in the description is auditable. Plenty of tools will help you produce a result. Far fewer leave behind a research state where you can reconstruct how the result came about, which is what actually matters when somebody asks you to defend it.
Scientific work increasingly spans heterogeneous artifacts -- papers, code, datasets, scientific file formats, model outputs, figures, manuscripts, and team decisions -- yet general-purpose AI assistants rarely preserve these objects as a coherent, auditable research state. We present SciForge, a multimodal research-native AI workbench that reserves the graphical interface for human judgment while search, parsing, model routing, workflow execution, plotting, writing, and…
Adaptive Contrast Enhancement and Optimised Feature Matching for RootSIFT-Based Palm-Vein Recognition
arXiv (cs.AI) · July 17, 2026HCIG: A Hierarchical Cross-Modal Incongruity Graph Network for Multimodal Sarcasm and Cyberbullying Detection
arXiv (cs.AI) · July 17, 2026JoyNexus: Service-Oriented Multi-Tenant Post-Training for VLA Models
arXiv (cs.CL) · July 17, 2026Frontier Language Models Struggle to Copy: Text Can Be Better Viewed in 2D