Spotlight Talks
summary
Shows a simple artificial‑money mechanism that provably approximately implements the core at equilibrium in repeated allocation with packing constraints, proving equilibrium outcomes can be efficient and fair and revealing a formal link to classical monetary mechanism guarantees. Also proves standard score‑based valuations (LOO, Shapley) fail to give constant‑factor subset selection guarantees because of duplicate archetypes and boundary effects, then pins two repair conditions — bounded curvature and coverage — and backs it with a scrape of 1.86M models on Hugging Face showing deep model genealogies, sibling similarity, and collapsing documentation.
AI-written summary · about
chapters (8)
- 0:00Welcome
- 25:20AI ecosystems modeling
- 1:09:33Fixing recommendation algorithms
- 2:00:00Auditing recommendation systems
- 2:13:40Learning preferences from choice data
- 3:01:08Dynamic allocation of goods
- 3:24:25Score-based data valuation fails
- 3:44:29Poster session & closing
related talks
01h 20mGame Theory Applied to Mining Pools, Token Offering, Dao, and Consensus Protocols
24mInformation is the New Money
14mAbstract VCG: Tie-breaking rules in VCG mechanism design
01h 32mNetworks and Economics 3
01h 32mNetworks and Economics 4
01h 07mIndustry Presentations
52mGame-Theoretically Secure Protocols Inspired by Blockchains
39mSketching and Randomization for Distributed Submodular and Coverage Optimization