Research in the AI Era Seminar: Karu Sankaralingam, The Impact Market to Save Conference Peer Review: Decoupling Dissemination and Credentialing

Content

Image

Karu Sankaralingam

Speaker

Karu Sankaralingam (UW Madison / NVIDIA)

Abstract 

Top-tier academic conferences are failing under the strain of two irreconcilable roles: (1) rapid dissemination of all sound research and (2) scarce credentialing for prestige and career advancement. This conflict has created a reviewer roulette and anonymous tribunal model - a zero-cost attack system - characterized by high-stakes subjectivity, turf wars, and the arbitrary rejection of sound research (the equivalence class problem). We propose the Impact Market (IM), a novel three-phase system that decouples publication from prestige. Phase 1 (Publication): All sound and rigorous papers are accepted via a PC review, solving the "equivalence class" problem. Phase 2 (Investment): An immediate, scarce prestige signal is created via a futures market. Senior community members invest tokens into published papers, creating a transparent, crowdsourced Net Invested Score (NIS). Phase 3 (Calibration): A 3-year lookback mechanism validates these investments against a manipulation-resistant Multi-Vector Impact Score (MVIS). This MVIS adjusts each investor's future influence (their Investor Rating), imposing a quantifiable cost on bad actors and rewarding accurate speculation. The IM model replaces a hidden, zero-cost attack system with a transparent, accountable, and data-driven market that aligns immediate credentialing with long-term, validated impact. Agent-based simulations demonstrate that while a passive market matches current protocols in low-skill environments, introducing investor agency and conviction betting increases the retrieval of high-impact papers from 28% to over 85% under identical conditions, confirming that incentivized self-selection is the mechanism required to scale peer review. 

Speaker Bio

Karu Sankaralingam is a Director of Research at NVIDIA Research and Professor at UW-Madison. He is a researcher, inventor and entrepreneur and and works on computer architecture and deep learning. He has graduated 11 PhD students, and co-author over 100 papers with multiple award papers. He is also an IEEE Fellow and holds the Mark D. Hill and David. A. Wood Professorship.