Discount Model Search for Quality Diversity Optimization in High-Dimensional Measure Spaces
Bryon Tjanaka, Henry Chen, Matthew Christopher Fontaine, Stefanos Nikolaidis
We present a method that enhances exploration in quality diversity (QD) optimization and show how this method enables new applications for QD. Project page: https://discount-models.github.io/
Abstract
Quality diversity (QD) optimization searches for a collection of solutions that optimize an objective while attaining diverse outputs of a user-specified, vector-valued measure function. Contemporary QD algorithms are typically limited to low-dimensional measures because high-dimensional measures are prone to distortion, where many solutions found by the QD algorithm map to similar measures. For example, the state-of-the-art CMA-MAE algorithm guides measure space exploration with a histogram in measure space that records so-called discount values. However, CMA-MAE stagnates in domains with high-dimensional measure spaces because solutions with similar measures fall into the same histogram cell and hence receive the same discount value. To address these limitations, we propose Discount Model Search (DMS), which guides exploration with a model that provides a smooth, continuous representation of discount values. In high-dimensional measure spaces, this model enables DMS to distinguish between solutions with similar measures and thus continue exploration. We show that DMS facilitates new capabilities for QD algorithms by introducing two new domains where the measure space is the high-dimensional space of images, which enables users to specify their desired measures by providing a dataset of images rather than hand-designing the measure function. Results in these domains and on high-dimensional benchmarks show that DMS outperforms CMA-MAE and other existing black-box QD algorithms.
Proposes Discount Model Search for quality diversity optimization in high-dimensional measure spaces.
- Uses smooth continuous model for discount values instead of histogram cells to handle high-dimensional measures
- Enables quality diversity algorithms to distinguish solutions with similar measures in high-dimensional spaces
- Introduces image-based measure specification via dataset rather than hand-designed functions
- Quality diversity optimization
- Discount value estimation
- Black-box optimization
- Neural models
Training discount model induces computational overhead
from the paperAlternative targets for discount model training may improve properties like smoothness
from the paperPrimarily considers small MLPs for discount model architecture
from the paper
Explore domains requiring more advanced discount model architectures
from the paper
Author keywords
- quality diversity optimization
- black-box optimization
- derivative-free optimization
- latent space exploration
Related orals
Mastering Sparse CUDA Generation through Pretrained Models and Deep Reinforcement Learning
SparseRL leverages deep RL and pretrained models to generate high-performance CUDA code for sparse matrix operations.
Overthinking Reduction with Decoupled Rewards and Curriculum Data Scheduling
DECS framework reduces reasoning model overthinking by decoupling necessary from redundant tokens via curriculum scheduling.
MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agent
MemAgent uses RL-trained memory modules to enable LLMs to extrapolate from 8K to 3.5M token contexts with minimal performance degradation.
DiffusionNFT: Online Diffusion Reinforcement with Forward Process
DiffusionNFT enables efficient online reinforcement learning for diffusion models via forward process optimization with up to 25x efficiency gains.
Hyperparameter Trajectory Inference with Conditional Lagrangian Optimal Transport
Hyperparameter Trajectory Inference uses conditional Lagrangian optimal transport to reconstruct neural network outputs across hyperparameter spectra without expensive retraining.