AI Paper Atlas

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

DeepFoids: Adaptive Bio-Inspired Fish Simulation with Deep Reinforcement Learning

Yuko Ishiwaka, Xiao Steven Zeng, Shun Ogawa, Donovan Michael Westwater, Tadayuki Tone, Masaki Nakada

NeurIPS 2022 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Vision-language navigation agents · 0 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
two-step training (pretrain then transfer learning) instead of training from scratch per environment (ablation)PPO fish policies · 11 fish cage environmentsconvergence speed and overall performanceImproved
replace PPO with SAC (same reward parameters) (ablation)coho salmon policy · Coho salmon default environmentschool behavior naturalnessHurt
replace PPO with SAC after retuning reward parameters (ablation)coho salmon policy · Coho salmon default environmentcumulative average rewardHurt
SAC instead of PPO (ablation)coho salmon policy · Coho salmon default environment, early trainingaverage rewardMixed
increase population density from sparse (0.59 fish/m3) to dense (14.7 fish/m3)fish school simulation · simulation vs field dataschool state (swarming vs milling): 0.59 -> 14.7 fish/m3Improved
higher fish density and more occlusion in counting videosYOLOv4 counting system · real cage footagecounting accuracy: 22:32Hurt
smaller cage (3 m vs 5 m edge) for 10 red seabreamfish simulation · red seabream cubic cageswimming patternMixed

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