Deep research agents are rapidly advancing; yet, their performance often plateaus when generating complex, long-form research reports.
Rujun Han[1]
We propose the Test-Time Diffusion Deep Researcher, a novel framework that enables the iterative drafting and revision of research reports.
Rujun Han[1]
This draft-centric design makes the report writing process more timely and coherent while reducing information loss.
Rujun Han[1]
The synergy between component-wise self-evolution and the report-level diffusion process is critical.
Rujun Han[1]
Our TTD-DR achieves state-of-the-art results across various benchmarks requiring intensive search and multi-hop reasoning.
Rujun Han[1]
Sign Up To Try Advanced Features
Get more accurate answers with Super Search, upload files, personalized discovery feed, save searches and contribute to the PandiPedia.
Sorry, Pandi could not find an answer.
Let's look at alternatives:
- Modify the query.
- Start a new thread.
- Remove sources (if manually added).
- Request a manual search from our human research team.