PulseAugur / Brief
EN
LIVE 10:59:21

Brief

last 24h
[2/2] 222 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

  1. TOOL · arXiv cs.NE (Neural & Evolutionary) English(EN) ·

    Simultaneous Model-Based Evolution of Constants and Expression Structure in GP-GOMEA for Symbolic Regression

    Researchers have developed a new approach to symbolic regression using genetic programming, a method for constructing symbolic expressions that fit data. Their novel technique simultaneously optimizes both the structure of expressions and their contained real-valued constants. This integrated approach, merging the real-valued variant of GOMEA with GP-GOMEA, demonstrated superior performance compared to other methods of handling constants in GP-GOMEA. AI

    IMPACT Introduces a more accurate method for symbolic regression, potentially improving AI's ability to derive mathematical models from data.

  2. TOOL · arXiv cs.NE (Neural & Evolutionary) English(EN) ·

    GP-GOMEA with GPU-Based Fitness Evaluations: Design and Performance Analysis

    Researchers have developed a GPU-accelerated version of GP-GOMEA, an evolutionary algorithm for symbolic regression. This new approach significantly increases the speed of fitness evaluations, allowing for more complex problems and larger datasets to be tackled. The enhanced performance enables GP-GOMEA to discover smaller, more interpretable models and provides new insights into how expression structure affects search difficulty. AI

    IMPACT Accelerates symbolic regression capabilities, potentially enabling more efficient discovery of interpretable models in scientific research.