labs · publications
$ Research Output
Six accepted papers, four pre-prints, five open datasets. Peer-review first, productize second.
6p · 6ds
papers
Click any row to expand abstract.
L-009on-chain
Why per-frame on-chain state does not pay
Anderson Gladney · 2026
arXiv
preprint
L-008spatial-state
Sub-tick replication for browser games
Choi · Park · Gladney · 2026
PETS '26
accepted
L-007deterministic
Deterministic UEFN ↔ Roblox bridges
Tanaka · Gladney · 2026
arXiv
preprint
L-006on-chain
Cheap ZK proofs for game receipts
Park · 2025
CCS '25
accepted
L-005spatial-state
Spatial gossip for 10k-agent worlds
Choi · Park · Kim · 2025
SIGCOMM
accepted
L-004multi-agent
Shared-memory multi-agent consensus
Gladney · Tanaka · 2025
NeurIPS
accepted
open_datasets
All freely available. Download links in companion papers.
DS-112.4 GB
spatial-gossip-10k
Position telemetry from 10k-agent Roblox session. 240Hz, 6h runtime. Companion to L-005.
spatial-stateCC-BY-4.0
DS-10840 MB
passport-receipts-anon
Anonymized Passport receipts, 2025 Q1–Q3. Action type, timestamp, cost, ZK witness size. PII stripped.
on-chainCC-BY-4.0
DS-095.1 GB
uefn-roblox-replays
Raw replay captures across matched game sessions. Input-only format. Companion to L-007.
deterministicCC-BY-NC
DS-081.2 GB
agent-disagreement
Disagreement event logs from EXP-26. Agent IDs, memory divergence vectors, resolution outcomes.
multi-agentCC-BY-4.0
DS-07410 MB
edge-latency-112
RTT measurements from 112 edge PoPs to UEFN servers. Used to calibrate tick-rate selection.
spatial-stateCC0
DS-06690 MB
flock-bandwidth
Bandwidth traces from flocking-behavior sims at 500–5000 agent counts. 8 runs per scale.
spatial-stateCC-BY-4.0