AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

The Neural Testbed: Evaluating Joint Predictions

Ian Osband, Zheng Wen, Seyed Mohammad Asghari, Vikranth Dwaracherla, Xiuyuan Lu, Morteza Ibrahimi, Dieterich Lawson, Botao Hao, Brendan O'Donoghue, Benjamin Van Roy

NeurIPS 2022 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Interactive multi-agent trajectory prediction · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace MLP with ensemble+ (adds randomized prior functions) (ablation)ensemble · Neural TestbedKL^10: 1.356 -> 1.015Improved
replace MLP with sgmcmcMLP · Neural TestbedKL^10: 1.367 -> 0.947Improved
replace MLP with hypermodelMLP · Neural TestbedKL^10: 1.367 -> 1.107Improved
replace MLP with ensembleMLP · Neural TestbedKL^10: 1.367 -> 1.356No clear effect
replace MLP with dropoutMLP · Neural TestbedKL^10: 1.367 -> 1.347No clear effect
replace MLP with bbb (variational inference)MLP · Neural TestbedKL^10: 1.367 -> 1.375No clear effect
replace MLP with sgmcmcMLP · Neural TestbedKL^1 (marginal): 0.129 -> 0.122No clear effect
replace MLP with ensemble+MLP · Neural TestbedECE: 0.078 -> 0.085Hurt
add randomized prior function to ensemble (ensemble+) (ablation)ensemble · Neural Testbed, low data, medium SNRKL^10Improved
add randomized prior function to ensemble (ensemble+) (ablation)ensemble · Neural Testbed, medium SNRKL^1 (marginal)No clear effect
increase prediction order tau (ablation)ensemble+ vs ensemble · Neural TestbedKL^tauImproved
use agents with better joint prediction (KL^10) in Thompson samplingThompson sampling · Neural bandit, N=1000, d=50regret at T=50,000Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy